DCIGInc.com

Down Economy is Prompting Upper Management to get more Involved with Disaster Recovery

2009-07-02T10:00:00Z

There were a lot of interesting statistics that came out of the just released June 2009 study that was done by Applied Research on behalf of Symantec Corporation. However the one stat that caught my attention was the increasing involvement that upper management is taking in disaster recovery within enterprise organizations. Executives in North America increased their participation on DR committees by almost 50% in the last year (67% in 2009 versus 46% a year ago) while globally executives more than doubled their participation on these DR committees from 33% in 2008 to 70% in 2009.

During the month of June, Applied Research contacted 1650 corporations worldwide that have at least 5000 employees with 350 of these enterprise organizations from North America. While the study did not cite any exact reasons why executive participation increased so dramatically in 2009, I suspect the down economy that all enterprise companies are working their way through right now probably had as much to do with increased executive involvement as anything.

Since credit is tight, new growth opportunities are hard to come by and many bills are working their way through Congress that will likely inhibit external growth, executive management appears to have more incentive and time to focus on internal processes and get their houses in order. As they do so, they are focusing on some long standing problems within their organizations with disaster recovery apparently showing up for many on their list of priorities.

Their increased participation and focus on DR is now having an impact on how these organizations are budgeting for disaster recovery initiatives in 2009 and 2010. Budgets for disaster recovery initiatives generally include money for such items as backup and recovery software, clustering, archiving, spare servers, replication, tape costs, services, DR plan development and offsite costs.

Despite the fact that many of these companies are planning to decrease their budget over the next 12 months (Globally - 42%, North America - 29%), many more are seeing their DR budgets hold steady (52% globally and 66% in North America) while a small percentage are even seeing an increase in their budgets. This is notable considering that most organizations are seeing an overall downturn in their total revenues so just staying flat is an accomplishment in this environment.

However I differ slightly from Symantec as to why this is occurring. In a prepared statement, Symantec said, "Symantec believes that some of this increase from previous years and attention on DR may be due to DR becoming a competitive differentiator, as well also a number of other factors including the size of DR budgets and the impact on customers of downtime that leads executives to focus more on keeping existing processes running smoothly."

I agree in part with this conclusion but I think it goes beyond DR just becoming a competitive differentiator and the impact on customers of downtime. These are both true statements but I think as executive management gets more involved with their DR committees, they are seeing the true state of DR within their organizations and they are getting a reality check. They are discovering that after spending millions of dollars on hardware and software over the years, they still can often only recover a fraction of their applications in the time that they need to recover them. As a result, their entire business is at risk if they do not fix this.

The good news is that it appears that as executive managers engage in these DR committees, they are finding out exactly what their company's internal DR requirements are and what additional software and hardware they need to procure to fix their DR situation once and for all.

This I see as the more likely reason as to why so many DR budgets are staying flat or even increasing in 2009 despite the down economy. DR now has a voice and a champion in the ranks of upper management and the need for DR is being articulated in terms that other executives on the board can understand, appreciate and justify the allocation of additional funds.

This bodes both good and bad for IT providers. Short term, it looks to be good news as this survey indicates. Enterprise organizations are going to fill their gaps in DR so they have a DR plan that works reliably. Conversely, as this survey also indicates, once these companies make these investments and fix their current DR environment, they plan to decrease spending in 2010 because they have brought their DR environment and supporting processes back up to snuff so they do not need to budget as much money going forward into 2010 on DR initiatives.

The cost of downtime and more stringent recovery time objectives (RTOs) may be the reasons that enterprise organizations are citing as to why they are giving new priority to DR requirements. Maybe, but these reasons for doing DR have been around for at least the last decade. A more likely reason is that as organizations cut back and refocus inwardly, they are seeing that DR is one area that needs attention. So by making some strategic investments in DR now, they can plug some existing holes, ensure their DR processes run more efficiently and potentially spend less money on DR in 2010 if they spend wisely now.

Cloud Storage Architecture Gives MSP New Found Flexibility to Respond to Spikes in Application Requirements

2009-07-01T10:00:00Z

One time occurrences that are accompanied by spikes in capacity and performance requirements are the bane of data centers. While many organizations can excuse IT for their inability to respond to unexpected one-time or occasional demands, perceptions and attitudes change when organizations know a heavy load is coming and IT cannot adeptly respond. It is this type of challenge that Carrenza Hosting, a managed hosting company based in London, England, intended to solve when it began to start down the path of adopting cloud-based computing and storage services. Part of its intent was to create a more flexible infrastructure so it could meet a client's forecasted spike in the need for computing and storage resources with minimal or no disruption to the client's application service levels.

The impetus behind my conversation with Carrenza's Director of Information Services, Jason Reid, was sparked by a recent joint press release from 3PAR and Carrenza. 3PAR recently made a technology donation to Carrenza in support of Comic Relief's annual charity event, Red Nose Day. On the surface, 3PAR's contribution seemed well-meaning and innocuous enough but taking advantage of a contribution of storage hardware is not as always as easy as it sounds.

Having previously working in an enterprise data center, installing storage, migrating the application to and from it and then decommissioning the storage once an event is over to return it can be a laborious and time-consuming process - so much so that the effort involved can outweigh whatever benefits the technology was supposed to provide. So to better understand why Carrenza needed this donation and how (and if) it could took advantage of 3PAR's contribution without undue labor I spoke to Jason Reid, Carrenza's Director of Information Services.

Reid explained that Red Nose Day is one of the largest one day charity events in the UK that takes in about $80 million US dollars. Comic Relief, the organization that runs Red Nose Day, normally has a platform that is quite small as there are normally not that many transactions occurring on a day-to-day basis.

But when the Red Nose Day occurs, it goes from needing a very small platform to one that is very large. To support this once-a-year event, Carrenza works with its technology partners (3PAR, Cisco, and HP) to provide extra capacity for the Red Nose application platform just for the duration of the event. Reid says, "Comic Relief does not need this huge platform sitting there for the rest of the year consuming power, costing money and generally being completely unused."

To enable the movement of the application from a small, less-powerful hardware platform to one that can handle these extra transactions, the entire infrastructure behind the Red Nose application is virtualized and highly redundant. So to move it, Carrenza used VMware's vMotion feature to bring up another instance of the application on a more robust hardware platform, migrate the data and then shut down the old instance.

Carrenza's Reid leveraged 3PAR's storage contribution in a slightly different way as 3PAR's contribution came in two forms. 3PAR first contributed extra disk that went into one of Carrenza's 3PAR storage arrays at Carrenza's primary site. This storage was used for the day of the event as it was needed to support the Oracle database back end of the Red Nose application.

3PAR's other technology contribution was made in the form of a second array that was installed at Carrenza's secondary site. This was needed to support an Oracle RAC stretch cluster that was put in place to support the application on the day of the event. Since this cluster configuration has severe distance limitations and Carrenza did not have sufficient capacity at its secondary site to handle the requirements of the Red Nose application, it needed the secondary array from 3PAR to support these requirements.

But most impressive was Reid's testimony as to the ease of implementing the new 3PAR storage and then decommissioning it when the event was complete. Reid estimated that it took at most a day and a half for one of his engineers to configure the storage and get the Red Nose application up and running on the new server and storage hardware, including the syncing of the data between the primary and secondary sites. He says, "There was no impact on the night of the event, it ran perfectly. There is not much more to say than that. We did not have any problems nor did we encounter any performance bottlenecks. Then once the event was over, the storage was just as easy to decommission as it was to setup."

Storage management and data migrations are two of the biggest obstacles that companies now face as they look to respond to occasional or one-time spikes in demand. However as Carrenza's experience with the 3PAR InServ Storage Systems during the Red Nose Day event shows, these obstacles can now be overcome. By using 3PAR as its back end storage platform, Carrenza was able to dynamically and quickly scale out compute, capacity and performance for a client application and, once the event was over, just as easily scale it back to its previous levels with minimal or no disruption to its client and without requiring heroic efforts on the part of its IT staff to accomplish this task.

But what makes this especially interesting is how Carrenza is leveraging its virtualized infrastructure to grow its business and become more competitive. In a forthcoming blog, I'll get into how Carrenza is leveraging 3PAR to deliver this more robust, virtualized infrastructure to lower its costs, set the foundation for the foundation for its future and meet the current and future needs of its clients.

A "Destroy All Data" Policy will not absolve You of Your Data Retention Responsibilities

2009-06-30T18:00:00Z

"If it really costs millions to do that [e-discovery], then you're going to drive out of the litigation system a lot of people who ought to be there." This quote by Supreme Court Justice Stephen Breyer cuts to the heart of current issues surrounding eDiscovery. A recent DCIG blog highlighted how out of control litigation costs have become and have left companies with hard decisions on whether it is best to settle cases based solely on the cost of eDiscoveryattempt to litigate or . But as companies face unprecedented economic pressure, a key question comes to mind, "Are these costs driving risky data retention strategies such as destroying all of your data?"

A December 2008 poll at Law.com showed immature processes is the rule across corporate America when it comes to eDiscovery. The survey found 30% of companies in the survey lacked even basic policies for preserving evidence for litigation discovery. So based on these statistics, it is reasonable to assume this lack of knowledge in eDiscovery coupled with immature processes could lead to higher risks being taken by companies.

But a question I regularly hear is, "Why not set a policy that mandates the quick destruction of data and delete everything quickly?" The thought process behind this is simple. If an eDiscovery event occurs, simply point to the policy and attempt to show a routine and good faith destruction of data and avoid the associated costs. While it is a tempting to adopt this policy in order to try to avoid the costs of eDiscovery, it is a flawed approach and could result in more harm than good for your company.

The "Safe Harbor" eDiscovery provision, otherwise known as Rule 37(f), provides a means for companies to limits sanctions if (and I quote): "Absent exceptional circumstances, sanctions cannot be imposed for loss of ESI resulting from a routine, good faith operation of an electronic information system." Under this rule, a court may not impose sanctions on a party for failing to provide electronically stored information lost as a result of the routine, good faith operation of an electronic information system.

Based on this wording it would appear reasonable why companies might take a risk and attempt to limit their legal risks by quickly and routinely deleting documents such as email. But there are several areas of concern for businesses that rely on routine data destruction processes when it comes to their eDiscovery strategy, such as;

eDiscovery is still evolving and the rules can be a moving target. Even when safe harbor would appear to extend to your company, the courts can bring a new wrinkle as it pertains to eDiscovery and suddenly your company could be facing a huge sanction. For example, a court ruling in the recent case titled Phillip M. Adams and Associates, LLC v. Dell, Inc., provided sanctions against ASUS for not preserving e-mails dating back to 1999, even though the plaintiff didn't bring a claim against ASUS until 2005. This has cast serious questions on the future of rule 37(e).

What you view as routine destruction could in fact be spoilation. The above cited case is another good example of a company thinking they would be covered by safe harbor, and instead their idea of reasonable destruction of data through routine maintenance of their information system, was instead viewed by the court as spoilation of data that should have been held for litigation.
Legal hold of data is open to interpretation. Legal hold of data is a process for holding all relevant information pertaining to a case when litigation is reasonably anticipated. The term reasonable is open to interpretation by the courts, and court interpretation is rarely predictable.

The sheer volume of e-mail and its impact on eDiscovery continues to be a pain point for companies searching for answers to costs. Products such as Estorian's LookingGlass provide an answer for companies looking to control the costs and complexities of e-mail in litigation. LookingGlass provides structure to historically unstructured data as well as providing search functionality for answering eDiscovery requests which becomes a valuable resource in controlling e-mail review costs.

In today's economic climate it is understandable why companies are tempted to try and avert costs through risky data retention strategies. But, this high risk strategy will fail and the costs and consequences could financially ruin your company. Proper preparation and deploying technologies such as LookingGlass provide a vastly lower risk point than attempting to rely on policy and pray for safe harbor.

Deduplication can be a Quick Backup Fix but Permanent Data Management Solutions Exist

2009-06-30T10:00:00Z

Most organizations simply do not like to think about their backup problems. To many their backup problems feel so overwhelming and the steps to fix them are so painful and complicated that they are desperately looking for a quick fix. So when a technology like deduplication comes along that appears to do exactly that, their initial reaction is to buy it. But organizations should not fail to consider other products that include deduplication technology as part of their solution. These solutions can provide the quick backup fix that organizations crave while addressing persistent, underlying data management issues that also need attention.

Deduplication is one of the more revolutionary technologies to emerge in the data protection space in quite some time with interest in the technology and its potential currently reaching a fever pitch. The interest in deduplication is most likely driven by two factors that are currently on the minds of many:

The pending acquisition of Data Domain by either EMC or NetApp. Data Domain's acquisition price of nearly $2 billion dollars raised eyebrows from Main Street to Wall Street. Not only did it re-ignite the conversation around whether deduplication is a feature or a technology, but it raises the specter of what potential that deduplication may hold for other applications besides backup.
Staffing cutbacks coupled with continuing data growth rates. Job loss estimates for 2008 and 2009 vary but an estimated 3 million jobs were lost in 2008 and more than 2 million jobs have already been lost in 2009. In spite of fewer people working, many attendees at a recent TechTarget Backup School event in Waltham, MA, cited data growth as their number one challenge. Fewer staff and more data are forcing organizations to identify better, more effective ways to protect their data and attendees saw deduplication as one way to accomplish this goal.

But here is the trap that organizations must be careful not to fall into when looking to deploy deduplication when it is part of a target-based deduplication appliance. The primary promises that target-based deduplication appliance providers confidently make regarding deduplication are reducing backup times, improving backup success rates and reducing data stores.

However they will rarely if ever make any promises about improving the manageability of your data stores. In fact, when I pressed one provider on the topic, the provider absolved itself of responsibility in this matter and say it is not their problem nor was it their intent to solve this issue.

This is where organizations need to recognize deduplicating target-based appliances for what they really are. They provide a 3, 6 or even a 12 month respite from an organization's current backup problems by suppressing the symptoms of out-of-control data growth. However they do little or nothing to stem an organization's underlying problems around continual data growth, managing data retention and searches of archived, backup or production data.

While in good times organizations may be able to sweep this issue under the rug, a recent blog by CommVault's David West makes the point that a perfect storm is descending upon organizations. Inefficient IT processes, immature backup software, shrinking budgets, flat or declining head counts and growing legal requirements to discover and produce data is forcing organizations to spend their money more prudently
.
As they do so, they need to think more strategically about what solutions they buy. They can no longer expect to just throw money at a problem in the hopes that it will magically go away and, if it does not, then receive more money next quarter or next year to try to deal with the issue then. Those days are coming to an end if not already over.

Instead organizations should take a step back and see what other problems they can solve beyond just deduplicating their backup data. Archiving email and file data, performing eDiscoveries and reporting on data growth and backups are other issues that persist in many organizations. While these features do not have the same glamour that deduplication currently possesses, by bringing them in-house organizations can begin address the real issues that are causing the backup problems in the first place.

Data management software such as CommVault® Simpana® supports all of these features - archiving, deduplication, replication, reporting, search and many more. Further, in the last few years CommVault has taken great strides to tailor Simpana to meet the requirements of any size organization. CommVault can deliver a solution that meets an organization's immediate deduplication needs while putting in place a platform that can meet their future data management needs.

The hype around deduplication has not been all bad as it has served to provide greater attention on the issue of backup with which many organizations struggle. But organizations that view deduplication as some type of silver bullet for their backup problems are setting themselves up for disappointment down the road. There are no silver bullets to solving long term backup problems but data management software like CommVault Simpana does a remarkable job of providing a short term fix while delivering a platform as a long term solution that fits the needs of most any size organization.

Small Organizations not Exempt from Need for Redesign of Backup Infrastructures

2009-06-29T18:00:00Z

Data protection is a ubiquitous need that cuts across all size organizations and has resulted in dozens of products with specific features to address these needs. In fact, one can easily wonder why any vendor even thinks it stands a chance to compete by coming to market with new backup software. But still they do and part of the reason is that backup problems still persist; so much so that backup redesign has topped the list among end-users for three (3) years running as they struggle to meet new backup requirements. No where is this need for a better backup option more acute than in small organizations which is creating new opportunities for emerging data protection vendors like ROBOBAK.

The ROBOBAK data protection suite resembles other traditional enterprise-class backup software products as it ships with the types of enterprise features that small businesses are coming to want and expect. For instance, ROBOBAK offers:

Thin Backup. This backup option does a full backup of a server's data on the first backup and then backs up daily changes forever. Known also as "incrementals forever" in some enterprise backup products, Thin Backup has the same net effect. It contributes to dramatically shortening backup times and reducing the amount of backup traffic that is sent over corporate networks.
Heterogeneous operating system support. It is safe to say that most small organizations run the majority of their applications on Linux and Windows operating system. However, small shops always have a least a few UNIX servers lurking in the shadows that require backup and recovery so ROBOBAK must - and does - meet this small business prerequisite.
Integrates with Microsoft Exchange. Exchange is mission critical in many small organizations today which makes Exchange data protection critical for any new backup software to merit consideration. ROBOBAK smartly chose to integrate with and support MS Exchange and, in its latest v9 release, it even provides a new Thin Backup Option for Exchange (TBO-E) that further expedites Exchange backups and recoveries.

While ROBOBAK has numerous other features that are representative of enterprise backup products, I only list the aforementioned ones to give prospective customers a sense that the ROBOBAK data protection suite is a competitive offering for small organizations as well as large.

But to justify making a switch from known backup software to a relatively unknown suite such as one from ROBOBAK requires it to provide features that make it both compelling and justifiable to switch. Based upon an initial evaluation of ROBOBAK's data protection suite, there are three features that stand out in my mind that should pique the interest of small organizations evaluating this product and provide the needed motivation to make such a change.

Agentless deployment. "Agentless" means that an organization does not need to spend time deploying backup software on each server (physical or virtual) in their organization. Instead, all an organization has to do is introduce ROBOBAK's software into its environment and it will discover the servers installed there. The discovery occurs quickly, eliminates the time normally required to install and maintain the agents that traditional backup software requires and facilitates these size organizations quickly and easily configuring ROBOBAK to protect their environment.
Reporting features. I received a phone call just the other day from a consultant in North Carolina who was trying to help one of his clients get a handle on what they have in their environment. They were trying to make some decisions about how to best protect and recovery their data without spending gobs of money and time doing so. ROBOBAK can help in this scenario. Using its Smart Crawl Technology in conjunction with its agentless technology, ROBOBAK can auto-discover files across corporate computers and provides results in seconds - not days or weeks. Using these reports, organizations can assess how much data they have and will be required from a cost and infrastructure perspective to appropriately protect their environment.
It deduplicates and replicates. Small organizations may be small in terms of people but their data growth rates and offsite recovery requirements are just as great as larger companies. In response to these needs, ROBOBAK backs up data to disk and then deduplicates it to keep data stores minimal. Once deduplicated, it includes replication so these organizations can copy their deduplicated data offsite, often using existing WAN connections.

Small organizations need to redesign their backup infrastructures just like large organizations but they must do so with a fraction of the time, expertise, staff and resources that large organizations may have at their disposal. This is where ROBOBAK stands a reasonable chance at succeeding as it fills this gap that exists in these size organizations.

While ROBOBAK may not yet be as well-recognized or known and its product suite still lacks some needed application support (its lack of support for Microsoft SharePoint is noticeably absent), its data protection suite delivers the features that small and mid-size organizations need now which gives ROBOBAK a fighting chance in a saturated market segment but one that is a ripe for change.

Recovery Should Not Displace Backup as the Next IT Headache

2009-06-29T10:00:00Z

Most organizations recognize that the introduction of disk into the data protection process is fundamentally changing the landscape of how data is protected. But what organizations are failing to entirely grasp is how disk fundamentally alters how applications can be protected and recovered. Disk can minimize the impact of data protection on production applications while providing shorter recovery times and improving recovery reliability. It is as organizations come to this realization that they also begin to grasp how recovery can displace backup as the next IT headache.

Backup has been a problem within organizations for so long that most IT managers are just relieved that troubleshooting last night's backup issues is no longer at the top of their daily "To Do" list after they deploy disk. But now that their data is protected, the bigger issue of managing the timely recovery of their applications takes on a whole new life.

Organizations may use multiple means - backup software, replication software and clustering software - to protect and recover mission critical applications while primarily using backup software to protect applications not deemed "mission critical".

Yet where they fall down is in providing a single, common mechanism to recover key enterprise applications such as Microsoft Exchange, SQL Server, SharePoint and Oracle. Organizations may now rely on backup software to recover some of these applications, replication software to perform recoveries for other applications and clustering software for yet another set of applications.

So what emerges over time is that the money and manpower that organizations once spent troubleshooting backup problems is now redirected towards managing the recovery of their applications using these different software tools. This adds both cost and complexity back into the infrastructure.

A better approach for organizations to take as they look to redesign their backup infrastructure is to adopt a broader mindset that is based on "Consolidated Recovery" and then select a solution that supports that approach. While I have previously defined some of the characteristics that a consolidated recovery solution should possess, the most important aspect to keep in mind is that regardless of the solution deployed, it should NOT result in recovery displacing backup as the next IT headache.

It is for this reason that organizations looking to solve their backup issues without creating new recovery issues should take a hard look at InMage Systems' DR-Scout software. It provides organizations with a common software platform that they can use to protect and recover these different applications. For example, using DR-Scout organizations can:

Centrally protect and recover these different applications using one software-based tool. This feature of DR-Scout eliminates the need for organizations to configure backup software to recover one type of application, replication software to recover another and clustering software to recover yet another. Because DR-Scout continuously replicates data and constantly creates application recovery check points as part of its replication process, organizations can use just one product instead of many to protect and recover all of their applications.
Cost effective recovery of "Tier 1" and "Tier 2" applications. User expectations for the recovery of their applications are rising - and not just for mission critical applications. Users do not want to wait hours for IT to recover an application server just because IT classified a server as "Tier 2". From the end-users' perspective, if they are using an application and it goes off-line for whatever reason, they stop working. So to those users the application is mission critical even if IT may not classify it that way.

Using DR-Scout, IT can extend the same level of near-real time recoveries to multiple different applications with minimal increases in costs since only one software platform, not multiple, are needed. Further, DR-Scout may actually lower IT's support costs while improving service levels. IT staff no longer need to monitor each application server to determine if and when they need to reclassify a "Tier 1" as "Mission Critical" since all applications protected by DR-Scout are afforded the same high levels of protection and recoverability.

Improve your data protection and recovery methods while showing an ROI. Using multiple data protection products create real, ongoing hard dollar OPEX costs in the form of licensing fees. Backup agents can be removed from production servers since DR-Scout is now effectively collecting data necessary to support any recovery operation. If you still want to dump data to tape, you can back up disk-based images created by DR-Scout, effectively off-loading production servers. By consolidating on DR-Scout, organizations can lower software agent costs and minimize the impact of agent maintenance operations. As they protect and recover more applications with DR-Scout, the savings should continue to increase.

The growing adoption of disk for use in the data protection process is fundamentally changing how organizations think about backing up their data. However organizations are failing to fully realize how eliminating their backup problems is creating new recovery challenges that current data protection technologies do not address, address only in part or are so expensive that organizations can only offer these high levels of recovery to a small subset of their application servers. Only by implementing software such as InMage Systems' DR-Scout that takes all of these factors into account can organizations begin to solve their backup and recoveries once and for all by consolidating backup and recovery under one roof.

Backup Software - Tape Library Combo Runs Counter to Current Disk-Based Backup Craze

2009-06-26T10:00:00Z

One might think the data protection world has gone mad. After all of the coverage over the last few years about the "goodness" of disk and the "evils" of tape, a recent announcement from Spectra Logic that it had entered into a new OEM agreement with Symantec ran counter to this disk-based backup craze. What specifically caught my eye in this announcement was that Spectra Logic was now bundling solutions that would integrate Symantec's NetBackup and Backup Exec software solutions with either its disk or its tape library products.

One of the emerging trends in the data protection market is the growing availability of appliances that bundle backup software with disk. In that respect, a component of this press release reflects this broader trend as Spectra Logic can bundle Symantec's Backup Exec software with its nTier family of Classic and Deduplication VTL appliances.

The nTier Classic series of appliances are based on Microsoft Windows Storage Server, so Spectra Logic can install Backup Exec on an nTier appliance and provide customers with a single bundled solution that consists of backup software and disk. This is an obvious synergy between Spectra Logic and Symantec but it is similar to other announcements from Symantec where it has bundled its backup software with a hardware vendor's product.

But what caught me somewhat by surprise is that Spectra Logic is also bundling its tape libraries with Symantec's backup software. This I found a bit confusing. After all, if organizations already have backup software and are using tape, why would they find a tape library that includes backup software appealing?

To try to get some answers to that question, I spoke to Spectra Logic's Senior Product Manager, Kevin Dudak. He explained that among Spectra Logic's customer base there is still a lot of confusion around deduplication. Because deduplication comes in so many forms (target-based deduplication in appliances, source-based deduplication on clients and even deduplication that occurs on backup software media servers), many of its customers are still not ready to move forward with deduplication and bring it into their environment so they are sticking with tape for now.

This is where Spectra Logic encounters some difficulties. Many of its customers that continue to use tape are using older versions of Symantec's backup software with a larger percentage using older versions of Backup Exec. This can create problems when Spectra Logic goes to deploy a new tape library in the customer's environment. Since the customer has not upgraded the backup software or kept its support up-to-date, Backup Exec can not take advantage of some of the newer technology available on the new Spectra Logic tape library.

This is where the impetus behind bundling Symantec's Backup Exec backup software with Spectra Logic's tape library comes in. Rather than requiring the customer to make two separate purchases, do a Backup Exec software upgrade and then install the tape library, by purchasing both the backup software and the tape library from Spectra Logic (or from a Spectra Logic reseller), everything can be configured at once.

Combining these two purchases solves problems on a number of fronts. First, it simplifies life for both the customer and Spectra Logic. By buying everything from Spectra Logic, Spectra Logic's engineers can handle both the upgrade and/or installation of the backup software and the setup and configuration of the tape library.

Second, doing it this way helps to ensure that when the Spectra Logic engineer leaves the customer site, everything is operational and working. Dudak has already witnessed a circumstance where an engineer discovered that the customer had insufficient licenses for everything to work together as designed. Because the engineer was onsite and configuring Backup Exec and the tape library, he was able to pinpoint the problem and advise the client to obtain more licenses so the whole system ran more efficiently.

Backup to disk and deduplicating backup data are generating a lot of interest among end-users. But interest does not always translate into a decision to move ahead with deduplication and many are still keep both disk and tape as part of their backup infrastructure for now.

So as organizations familiarize themselves with how to use disk in general and deduplication specifically in their backup infrastructure, they are still opting to buy technologies that they know and trust. That's why new alliances such as this one between Spectra Logic and Symantec are so valuable. They provide organizations with an intermediate step to addresses their immediate needs around backup software and tape, it allows them to use tape to store their data for long term retention and offsite DR and it keeps their options open in regards to introducing deduplication into their environment at some point in the future.

Herbalife's Organic Data Growth Creates Unexpected Backup Challenges

2009-06-25T10:00:00Z

Exposed. That was the position that Herbalife's Principal IT Engineer, Andy Hansen, found himself in more frequently in mid-2007 as he watched Herbalife's data growth explode and the backup software that he was using struggle to keep up. Much of Herbalife's new data growth was driven by its new corporate-wide enterprise resource planning (ERP) software initiative that increased its production data stores from 32 TBs to 240 TBs of data. This growth plus new backup demands left Hansen uncertain as to if Herbalife could recover from data loss or application disruption should any type of outage occur - minor or major.

The rapid explosion of data that Herbalife experienced only exacerbated the issues that Hansen was already having with his current backup software. So in his role as Principal IT Engineer, he was tasked with identifying a solution that addressed the problems Herbalife was encountering and could scale into a new environment. The specific problems that Hansen was encountering included:

No backup reporting. Hansen had little or no insight into the causes of why his backup jobs were failing or what steps he needed to take to fix them. Since his current backup software stored its indexes and backup job information in a SQL Server database, he had to write his own SQL queries to pull needed information out of the database to try to understand why backup jobs were failing.
Backups failing in the middle of bundled backup jobs. His backup software allowed him to bundle multiple backups for individual servers into one queue such that when the backup of one application server was complete, the backup of the next application server in the queue would begin. The problem that Hansen was encountering with this approach was two-fold. First, if the backup of one application server in the backup queue failed, all of the other backups after that application server in the queue would also fail. Second, it was very difficult to construct queries to find out exactly why the backup of a particular application server in the queue failed. As a result, Hansen was often left to guess exactly why his backups were failing since he did not have the time to research and diagnose the cause of each backup failure.
No Linux agents. Herbalife's new ERP application used an Oracle database that ran on a Linux platform. However Hansen's backup software did not offer Linux agents for Oracle backup at that time.
No integration between different instances of the same backup software. As part of Herbalife's ERP initiative, it was consolidating the management of its data in its central and remote sites so all of its resources could be centrally tracked and managed. When Hansen checked on how well his current backup software would support this new configuration, he discovered that there was no way to combine the different instances of the backup software catalogs at the central and remote sites so they functioned as one.

The scope of problems with his current backup software coupled with the growth of data in Herbalife's environment led him to conclude that Herbalife needed to upgrade to enterprise backup software to properly protect its new ERP environment to minimize its current level of exposure. So with these requirements in mind, Hansen brought in three enterprise backup software products for evaluation: CommVault® Simpana®, EMC Networker and Symantec NetBackup. In the next blog entry in this 3-part series, I'll take a look at what Hansen discovered during his evaluation of these three different products.

Knowing What Tweaks You Can Make to the Microsoft Exchange Spreadsheet Calculator

2009-06-24T10:00:00Z

Plug-n-play - that's part of the idea behind the Microsoft Exchange Storage Calculator spreadsheet which provides organizations the general guidelines that they need when planning and configuring the storage that will support an instance of Microsoft Exchange. However one should not assume that this spreadsheet takes into account every possible variable regarding storage systems - it most certainly does not and says as much within the spreadsheet. So before making any tweaks to the spreadsheet, one must know something about the capabilities of the storage system that one will use to host Exchange and what changes one can confidently make to the spreadsheet without botching an Exchange implementation.

The "Storage Design" page within the Exchange Calculator spreadsheet is the one that organizations will specifically want to take a hard look at when they start to plan for and configure the storage system that will host Exchange. At the top of this page, Microsoft provides some guidelines as to how much overhead a disk drive failure within a RAID group will introduce during rebuild times (35% for RAID 1/0 and 100% for RAID 5 or 6).

The motivation behind Microsoft providing these percentages is to assist organizations in appropriately sizing their storage system should a disk drive failure occur. These percentages ensure the storage system maintains optimal performance for Exchange even in the event of a disk drive failure. So as an organization interprets these percentages and applies it to their environment, this is the amount of extra capacity they must add to their storage system to maintain production levels of IOPs and throughput for Exchange.

In the case of RAID 1/0, an organization needs an additional one third (1/3) more capacity (35%) to ensure forecasted production levels of I/O are met during the rebuild of a failed disk drive. In cases where RAID 5 or RAID 6 are used on the storage system to support Exchange, organizations need to plan to implement twice the disk drives if they want to offset the 100% overhead that rebuilds of disk drives in these RAID configurations incur.

While these overhead percentages associated with these different RAID configurations appear logical on the surface (and they are), organizations should verify that these percentages are applicable to the storage system that they are looking to implement behind Exchange. Since Microsoft makes a number of assumptions in its spreadsheet about RAID technology, this is one area that organizations can look to tweak without negatively impacting their Exchange implementation.

Bill Plein, a storage architect with 3PAR's Strategic Accounts Engineering Team, explained that he works very closely with 3PAR customers to help them understand that they can satisfy all of Microsoft's best practices without needing to follow this particular detail on Microsoft's storage calculator spreadsheet. While 3PAR supports and follows all of Microsoft's best practices, he helps customers understand some recommendations around storage design are important for availability reasons and others for performance.

In the case of the 3PAR InServ Storage Server, it uses a wide striping technology so it does not have the same performance problems regarding the RAID rebuild overhead that Microsoft's recommendations are trying to address. Instead 3PAR spreads out its rebuilds across every drive in the system, not just to a single replacement drive.

So if a 3PAR InServ Storage Server has hundreds of drives in it and it looses one drive, almost every drive in the system get a little bit warmer because all of the drives start rebuilding the parity or the mirror chunklets that were lost when that one drive went down. Then when that drive is repaired, all of the data is leaked back to the new drive in a background process.

Since the drives participate one by one and restore the data in the background, 3PAR's rebuild time is fast and the performance penalty is extremely low - less than 5% if measurable at all. So even though Microsoft says allow for a 35% overhead on RAID 1/0 systems and 100% on RAID 5 and 6 systems, 3PAR can deliver the same reliability and performance using its wide striping technology as these other approaches with less than a 5% overhead on its system.

Plein concedes that 3PAR customers conservatively plan for a percentage higher than 5%. He says, "Even if the customer goes with an ultra-conservative ratio of 10 - 20% of overhead for RAID 5, 3PAR is basically selling its customers less disk as they do not need all of the Mirosoft-recommended spindles. Here is a place where 3PAR does it better and they can back off on the Microsoft recommendation."

In most organizations, Microsoft Exchange is too critical of an application both from a business and technical perspective to make any tweaks to Microsoft's recommended configurations without solid evidence that they will work as designed. However this is one example where Microsoft is over-compensating in its recommendation because it has to design its specifications to the lowest common denominator which, in this case, is traditional RAID arrays. Using 3PAR's wide striping technology in lieu of traditional arrays, the customer can be pretty confident that when the solution is rolled out there will be ample head room on the system without needing to purchase a lot of extra disk to accomplish that objective.

The Consequences of NOT Clustering Your SAP Implementation

2009-06-23T10:00:00Z

Attempting to make your entire SAP environment highly available can be a gargantuan challenge, especially considering the number of moving parts contained within an SAP landscape. Most of the time when one looks to ensure that any application is protected and made highly available, it's common practice to ask the application vendor for a set of best practices and guidelines to do so. However, SAP's typical response is, "Work with our partners and/or 3rd party consultants to help you achieve the level of availability you are looking for." In some cases this may be the right answer but these partners and consultants then have to spend considerable amounts of time and effort learning your R/3 or NetWeaver environment - and on your dime, of course.

Most of the high availability situations that I have run into involve the customers asking for help to make the database layer of their SAP implementation highly available. While making their database highly available does protect a critical component of an SAP implementation, it does nothing for other equally important components of the SAP install such as:

Central Instance
SAP Central Services (SCS) Instance
Enqueue Replication Service Instance
Dialog (Application Server) Instance
NFS (Network File System) Service

In other cases I have seen customers with SAP HA configurations that are a culmination of one-off's, such as Oracle RAC for the database, NFS Clustering (Multiple NAS Heads or file servers), Windows clustering for application servers, and other techniques for the portions of SAP that run on VMware (HA and DRS). These configurations become especially problematic when it comes to synchronizing transactions across all of these different platforms.

The real concern is the amount of effort that goes into recovering from and performing failovers that actually work across this complex menagerie that SAP infrastructures can become. Recovering just one component of an SAP implementation, such as the database, is of little or marginal value as it introduces some new dangers and problems which include:

The SAP Central Services Instance is the driving engine of the SAP application. Without this component a critical portion of the environment is neglected. Operating as portals, CRM and many other areas integrate directly into the CI that would in many cases deprive your end users of access to the overall system
SAP strongly recommends combining the Standalone Enqueue Server included in the SAP Central Services Instance with an Enqueue Replication Instance, which maintains a replica of the critical enqueue lock table in memory. In case of a SCS (SAP Central Services Instance) instance failure, the cluster will restart the SCS on the node running the replication instance. In case of a successful failover no SAP transaction lock data is lost and therefore no open user transactions need to be rolled back.
NFS (Network File Service) provides shared access to the central file systems which are required in most situations to keep the SAP environment up and running. NFS provides the linkages to all the application servers, web servers and other portions of SAP. If NFS is out for any reasons the overall SAP implementation will be greatly impacted.
Application Server Instances, although not considered a SPOF (Single Point of Failure) in the SAP landscape, can be clustered together with other more important services to improve uptime and the overall user experience. The application servers maintain specific aspects of the SAP delivery to the end user so while losing one may not be a concern seeing as most of the time the application servers are load-balanced, if multiple instances become unavailable the SAP environment will not function properly.

The real goal should be to put in place a solution that works across all of these different components of an SAP solution so that all of them can be recovered and not just the database component which is how HA solutions for SAP are too often deployed. To do this requires a protection scheme that works across all applications, databases, file systems, and operating systems that exist in your SAP environment. Ideally all of these HA processes should be controlled thru a single management framework that is simple and straightforward to operate even for junior level system administrators with the appropriate training.

Veritas Cluster Server (VCS) and Cluster File System (CFS) provide these layers of functionality, which, when properly configured, take all the vital SAP services and makes them fully redundant. Adding VCS and CFS to all the appropriate tiers inside the SAP environment enable organizations to deliver speedy recoveries from hardware outages but also ensure uptime during patching and normal maintenance activities. In a forthcoming blog entry, I'll take a look at how VCS delivers this functionality, what specific problems it solves and how organizations are showing an ROI using VCS on SAP.

Customers, Channel Now Rely on Appliances to Deliver Latest Software Functionality

2009-06-22T10:00:00Z

Independent software vendors (ISVs) that sell software based on x86 hardware platforms face a new type of challenge in today's economic environment. While their software can run on any vendor's hardware platform, the time it takes for them to install, configure and support their software on each platform gives resellers pause and is prompting resellers and customers alike to look for the ISV's software in the form of appliance-based solutions. In this first of a 3-part series, DCIG meets with Tom Baylark, an independent consultant to Bell Micro, to discuss why ISVs should consider offering their software on an appliance and how appliances can improve their relationship with the channel.

DCIG: Why should ISVs consider making their software part of an appliance? How will this help them in the channel?

Tom: If you take a look around, customers are starting to rely more on appliances in all aspects of life whether it is the MRI or CAT Scan machine at the hospital or the remote data terminals that are popping up in healthcare. These are all appliances meaning they are running a specific application on a specific piece of dedicated computer hardware.

People who have digital video recorders (DVRs) today in the form of TiVO or whatever brand they have built into their satellite literally have a little Linux server inside of their house with a specific type of application that is running on top of it. In the data center, customers have firewalls and networked attached storage which again are dedicated servers running a specific type of application.

Can you imagine having to purchase the operating system and the software functionality to implement a TiVO? Can you imagine having to purchase that stack to implement your network attached storage?

No, customers want it all-in-one; it's simpler. This is always a real plus that an ISV has when they can say, "Put this system or this solution to a specific problem in your environment, turn it on and in a very small period of time, it will rapidly configure, you can install it and you can see all of the benefits and none of the hassles of trying to make sure it integrates into your business."

In regards to the channel, the appliance fits well with the traditional channel because the channel likes to sell their experiences and promote their success in "fixing" entities. The channel likes to say, "I have a server", "I have a piece of storage", "I have a switch", "I have a NAS box", "I have this piece of functionality" that it can sell you and it doesn't have to be an expert on the software functionality.

The channel doesn't like selling Oracle databases because they have to be Oracle database savvy. Due to the amount of products that traditional channel partners are trying to contend with, having to develop expertise on any particular piece of software is not really practical plus then they have to supply support for it. They know they can't just sell software and run away as the customer will want to call them when it is not working and if they don't have the expertise to manage it, that doesn't look good in the eyes of the customers. The channel is very good at selling fixed entities such as appliances. It's here, it has a SKU and I can sell it. It is an easy sell.

DCIG: Software vendors, even those that work through the channel, don't always know that. What is your experience in this regards?

Tom: With very horizontal applications like backup and recovery, over time the reseller may embrace the idea of putting together a practice around being able to go and execute a holistic sales cycle around those types of solutions. But for the most part, if the customer wants a specific software solution, the reseller will get the box of software and leave it on the customer to manage the implementation and all of the life cycle of the software stack.

I don't see that changing any time soon. From a perception standpoint, software is many times seen by the channel as a real headache.

DCIG: So resellers see software as taking too long to sell?

Tom: Absolutely. The sales cycles goes through the roof because, from a customer perspective, hardware is pretty simply to buy as customers have been buying hardware for a long time. When it gets down to the software functionality, customers have to do significantly more expensive testing, the decision cycle takes longer to happen and, since software is more expensive, it has to go through more levels of review during the sales cycle. Resellers just do not want to deal with that.

Further, hardware does add some tangibility to what it is that the reseller is selling. The customer can see it, the brand name is sitting there and the blinking lights are happening on the reseller's behalf.

DCIG: When you are talking to ISVs about the Bell Micro OEM business, what is their reaction? Does it convince them or are they still skeptical?

Tom: ISVs are very interesting folks. The only value they see is in their piece of software. From the ISV perspective, isn't hardware just free? Heck, it is an X86, just plug it in and go. That's the model that has been successful for them.

But there are challenges that are going on today that are going to change how ISVs have to act in order to be successful and grow effectively. There are some companies that have been very happy with the status quo. Well, the status quo is going downhill if you have looked at your 401(k) lately.

There are resellers who are just marking time and trying to hang on as they have no growth strategy for 2009. The goal of two of the resellers I have talked to is to just be around in 2010. At the end of 2010, they just want to be here. They are marking time and saying they plan to keep on doing things just the way they have been doing them.

As you get into the public sector or larger enterprises, these sectors give to their technology providers a list of preferred vendors and anything they buy has to have a specific vendor's logo on it. It is pretty cool when you put together a solution and it is all HP running this vendor's software.

That is an immediate pull. If a reseller goes into a data center now and the reseller wants to implement an email recovery system, the data center may say, "Great!" But when the data center looks at the solution and sees a little white box, it may say, "No way! We are not putting a white box in because we only buy HP servers." In that situation, it behooves the ISV to align with a specific hardware vendor like HP and manufacture some sort of appliance.

In part 2 of this 3-part series, Tom discusses why some standardized hardware platforms are better solutions than a heterogeneous, white box solution and how branded solutions differ.

In part 3 of this 3-part series, Tom discusses how offering software on an appliance can broaden software's appeal without increasing and even possibly lowering ISV costs.

Quantum DXi2500-D Cuts to the Chase as to What Mid-size Organizations Need in a Deduplication Appliance

2009-06-19T10:00:00Z

"We need cheaper and simpler backups and recoveries for our remote and branch offices." That statement is repeated more often by mid-size companies as they seek solutions that take the pain and management overhead associated with backup and recovery out of their remote offices without breaking the budget or requiring heroic efforts to implement. This is exactly the type of scenario that the recently announced DXi2500-D appliance and version 3.0 of Quantum Vision™ Software from Quantum is designed to address.

Mid-size organizations are exactly that - mid-size - so the types of backup and recovery problems that they encounter are symptomatic of both small and large organizations. On the "small" end of the spectrum, mid-size companies only have a few IT staff that must support the entire data protection infrastructure in all offices from deploying the backup software to managing the recoveries. On the other end of the spectrum, the amount of data that they need to protect requires the skill sets, hardware and software that are more typically found in large organizations.

This is the gap that the DXi2500-D was designed to bridge. The DXi2500-D is sized with the capacity to easily handle sites with as little as a couple of hundred GBs of primary data to sites that have as much as 1 TB of data. It offers the right core technology features that are most likely to meet the needs of short-handed, cost conscious mid-size organizations. Using the DXi2500-D in conjunction with the Quantum Vision software, the can meet these needs in the following ways:

A price point and feature set that makes disk-based data protection for remote and branch offices almost a no-brainer. Talk about features all you want, the first thing IT managers in midsize organizations want to know is, "What does it cost?" To cut the chase, the list price for the DXi2500-D is $12,500. This includes 1800 GB of disk, a NAS interface, deduplication, replication and Symantec's OpenStorage (OST) API software. That works out to about $7/GB raw or just under 35 cents/GB when the backup data is deduplicated assuming a 20:1 ratio. Further, because deduplication and replication are included with the DXi2500-D, organizations can frequently continue to use their existing WAN connections without needing to upgrade to send deduplicated data back to the home office.
Simple to deploy. IT staff at midsize organizations understand and know how to manage Ethernet networks. So by making the DXi2500-D a NAS-only appliance, Quantum makes the upfront installation and setup of these appliances simple and easy for organizations to accomplish as it integrates with their existing networks.
Near 100% backup and recovery success rates. No one (including me) can guarantee that just by installing the DXi2500-D that an organization's backup and recovery will be 100% successful. But that said, I have never spoken to an organization where its success rates have dropped after it has implemented disk as a backup target. The almost universal response is that their success rates always improve and whatever problems remain after they implement disk are now much more manageable.
Consolidated management across remote offices. Deploying DXi2500-Ds across branch offices brings these benefits to an organization, but they still require management that must be done with existing IT staff. Quantum's Vision software enables these organizations to centrally manage all their Quantum systems while at the same time providing detailed reports across time and sites on items like capacity utilization, deduplication ratios, system performance, system status and replication throughput.

Mid-size organizations are acutely aware of the pain associated with managing backups in their remote and branch offices but are also just as aware of the costs to fix their backup problems. The Quantum DXi2500-D now provides them a viable, cost-effective means to address their tactical backup and recovery problems while Quantum's Vision software addresses their ongoing management concerns by centralizing and consolidating the management of DXi-2500-D's and the data that they contain.

IT managers need fewer heroes and more solutions that work as reliable processes across their enterprise. The DXi2500-D coupled with Quantum Vision software solves these issues while laying a foundation that centrally and more easily manages their backup data and the appliances that host it going forward.

NetApp Mum on Data Domain; CEO's History of Integrity Lends Credence to "Not for Sale" Comments; 2009 NetApp Analyst Days Recap

2009-06-18T16:45:00Z

In the face of its public tussle with EMC over Data Domain, NetApp forged ahead with its annual Analyst Days at its Sunnyvale, CA, headquarters. Well attended by many NetApp executives and 80+ analysts from the US and around the world, it was both informative and well run. However it was the kickoff keynote by NetApp's CEO Dave Warmenhoven that I first wanted to summarize and comment on as he did a pretty good job of encapsulating the major themes of the presentations and one-on-one meetings that were to follow.

During his opening keynote, Warmenhoven commented on a number of topics ranging from the economy to how NetApp's new branding was affecting its ability to get into enterprise accounts to NetApp's culture and how it fosters innovations from within. However, these were the four main items I gleaned from his opening keynote presentation:

NetApp's positioning and growth in the current economic environment
The current state of its outstanding offer for Data Domain
His thoughts on why NetApp will not be acquired
NetApp's enterprise cloud storage offering

In regards to NetApp's growth in this economic environment, it is being affected like everyone else. Overall storage growth is flat after many years of double digit growth but not all is as bad as it sounds for the following reasons.

He pointed to a number of significant wins (many of which appear to be coming at the expense of EMC) as well as an increased presence by NetApp in Fortune 5000 companies.
Organizations want more flexibility and lower costs in their storage infrastructures and they are finding that their current infrastructures do not meet these new needs. This is creating new sales opportunities for NetApp as these organizations are willing to look at alternative ways to manage their storage infrastructure.
He so far has not seen this downturn hitting NetApp as hard as the 2000-2001 tech crash. It is still maintaining margins of 9 - 12% whereas during the tech crash profits were at or close to 0%.

In other side conversations I had with NetApp executives after Warmenhoven's presentation, apparently there is also some debate occurring internally as to how exactly the economy is impacting NetApp. It was notable that while there was a leveling off from FY2008 to FY2009 in terms of storage sales, there was no drop-off. Some NetApp insiders speculate that NetApp's customers are starting to make better use of new technologies like deduplication (which it now gives away for free) which is contributing to the flattening in storage sales while others are attributing it to the economy.

He then took a few minutes to talk about NetApp's bid for Data Domain. There was very little that he (or anyone) at the conference would publicly or privately disclose about the acquisition that was new information. One of the tidbits that I did pick up from his comments is that approximately 1/3 of Data Domain's employees are former NetApp employees. This certainly lends support to those who allege that the cultures of NetApp and Data Domain should mesh nicely should the two companies eventually merge.

Warmenhoven then stated emphatically why NetApp is not for sale and went on to explain why no other major technology company wants NetApp nor does it not make sense for any of them to acquire NetApp. The five companies that he mentioned that most often come up as possible acquirers of NetApp include Cisco, Dell, EMC, HP and IBM and he gave specific reasons as to why none of these companies are likely to act.

Cisco benefits from an alliance with both EMC and NetApp and if Cisco were to buy EMC or NetApp, he surmised it would be EMC. EMC would provide Cisco a greater footprint in enterprise accounts and EMC has a larger market share in storage than NetApp.
Dell already owns EqualLogic and is doing well with it. It also has an existing relationship with EMC for FC storage and Dell's cash position is not as strong as it was a few years ago.
EMC is almost out of the question as the anti-trust issues are too great.
HP's Mark Hurd has stated in previous earnings calls that he is happy with HP's current storage offerings and prefers to grow organically.
IBM is already selling NetApp and has seen its sales of the N-Series grow dramatically. The relationship is good so why buy NetApp?

These comments sparked some follow-on discussion after the presentation. Another analyst that I spoke to wondered if Warmenhoven was just making these comments so as to set the stage to fetch a higher price for NetApp. As for myself, I'm not so sure. Warmenhoven and the entire NetApp team strike me as pretty straight-up and, from my perspective, the synergies don't really feel right for NetApp to be acquired by any of these companies.

I can't point to any specifics but I get the sense that everyone (except EMC) seems to enjoy (too strong a word in this day and age?) working with NetApp. While I view Cisco and IBM as the most likely acquirers, my general feeling is that if an acquisition was in the works or even in the discussion stages, Warmenhoven simply would have avoided this subject entirely.

I came to this conclusion after reading NetApp's founder and EVP Dave Hitz's book, How to Castrate a Bull, on my flight home last night. In it, there is a section where Hitz talks about Warmenhoven's prior involvement in Network Equipment Technologies that was involved in all sorts of unethical behavior. Once it was uncovered, Warmenhoven was made the CEO and tasked with the job of cleaning up the mess. That lack of integrity on the parts of his predecessors at N.E.T. left such a bad taste in his mouth that I just can't see him standing in front of all these analysts and saying what he did if he didn't mean it.

Call me naïve, but based on the culture I observed while in attendance at NetApp's Analyst Days and the people I met, if one of these aforementioned five companies (or some other company) is getting ready to acquire NetApp, I don't think he knows anything about it.

Warmenhoven then concluded his presentation with some thoughts around NetApp's plans for future growth which include a growing emphasis on data protection and recovery, storage efficiency and simplified data management, of which delivering an enterprise storage cloud is a big part of realizing these ambitions. To its credit, during the two-day event NetApp provided a substantial amount of evidence to support its enterprise cloud storage story and is one of the better stories I have heard in this vein. In a subsequent blog, I plan to cover more about how NetApp is executing on delivering a storage cloud that meets the specific needs of the enterprise.

Satisfy Microsoft's Best Practices for Exchange Storage Configurations While Using Next Generation Storage Systems

2009-06-17T10:00:00Z

Best practices for configuring storage systems in enterprise Microsoft Exchange environments requires that they look beyond Microsoft's basic guidelines for how to configure their storage systems. While these guidelines may be sufficient when deploying Exchange in smaller shops where direct attached storage (DAS) or traditional array-based storage systems are the norm, enterprise organizations need to know when to look beyond these guidelines and tweak them when deploying Exchange on a next generation storage system like the 3PAR InServ Storage Server. The key to making these tweaks, however, is to satisfy Microsoft's best practices without negating the inherent benefits that storage systems like 3PAR provide.

As previously explained, one of 3PAR's key benefits is that it can disperse data throughout its entire storage system using its wide striping technology. In most cases, organizations will achieve even better performance metrics than what they saw when following Microsoft's recommendations for configuring storage. However because 3PAR Storage Servers use this wide striping technique (Marc Farley nicely illustrates how wide striping works on this blog post), organizations also should take advantage of 3PAR's template feature so they can concurrently use wide striping and still satisfy Microsoft's best practices for Exchange.

Bill Plein, a storage architect on 3PAR's Strategic Accounts Engineering Team, explained that there are different kinds of templates on a 3PAR storage system. One type of template is the volume template that can describe in great detail the layout of the proposed volume and even allows administrators to potentially lock down certain variables so they can't modify the template or change certain values. The value of these templates is that they only need to be created once and can then be re-used multiple times.

In the case of Exchange, organizations need to segregate the volumes on which they put their Exchange data and logs. So for example if an organization needs a 9:1 ratio between its Exchange data and logs, administrators can simply choose magazines or disks 0 - 8 from every tray or shelf of disks and assign those disks to the "Exchange Data" template and then choose magazine or disk 9 and put that disk into the "Exchange Log" template.

Configuring the 3PAR system this way enables organizations to then follow and implement best practices for Exchange storage configurations such as aligning volumes using the Windows 'diskpart' command or segregating data and log files on the storage system. Equally important, templates enables organizations to keep 3PAR's wide striping feature in effect since it stripes data across all of the disks in that template pool.

3PAR's autonomic provisioning feature also remains in play when templates are used. When Exchange needs more storage, administrators only have to instruct the 3PAR system to create a log or a data volume. When that request is submitted, 3PAR assigns the volume to the appropriate template and then the data is appropriately striped over these resources. Plein says, "At this point, organizations are automatically following Microsoft's best practices for Exchange since the policies are embedded into the templates that are part of the 3PAR storage system."

Just because enterprise organizations are placing their Exchange application on next generation storage systems does not mean that they have to sacrifice best practices, ease of administration or performance - they can still achieve all three of these objectives. Granted, to satisfy all of these requirements you can't just back up the truck, unload the 3PAR InServ Storage System, connect it to Exchange and expect it to be in compliance but neither is it an exceptionally arduous, ongoing task that it can become on other systems. A little upfront planning and work should enable enterprise organizations to satisfy Microsoft's best practices for Exchange storage configurations while still keeping their enterprise storage system supremely easy to configure and manage.

Veritas Storage Foundation for Windows Delivers a Big Step Forward in Automating Hyper-V Storage Management

2009-06-16T10:00:00Z

Symantec's recent announcement that it will support its Veritas Storage Foundation for Windows (SFW) 5.1 in the Microsoft Hyper-V parent tremendously increases the breadth of functionality available to the child virtual machines (VMs) of Microsoft Hyper-V environments. Immediate new benefits that child VMs will realize include the restoration of full path and storage management capabilities that often were severely handicapped once physical servers were virtualized. But other benefits that also come along with moving SFW 5.1 into the Hyper-V parent include better utilization of thinly provisioned storage volumes assigned to the Hyper-V parent along with an attractive SFW licensing option for Hyper-V servers.

Fully realizing the benefits of thinly provisioned volumes assigned to Hyper-V servers is a major value that organizations gain from SFW's move into the Hyper-V parent. One of the historical problems of Windows servers (physical or virtual) is the storage over-allocation that occurs on Windows servers with storage utilization rates in the 20% range fairly common.

Organizations can become excited by the prospect of virtualizing their Windows servers since they know this storage over allocation exists and, once virtualized, they only need to buy a fraction of the storage that their Windows servers are using now. Further, by putting thinly provisioned volumes behind their virtualized child Windows servers, the Windows Hyper-V servers still think they are getting their full allocation of storage even though beneath the covers the only storage they are using on the array is what is actually utilized.

However managing storage assigned to a child VM of a Hyper-V server is a much more complicated process than managing it on a physical server. In the case of "fat" volumes (non-thinly provisioned volumes), all fat volumes are first provisioned on the Hyper-V parent before the Hyper-V provisions VHDs on top of those volumes and presents those VHDs to the child. The problem with this approach is that there is no easy way to expand that volume in the child VM's storage pool when storage growth occurs.

This is partly the motivation behind using dynamically expanding VHDs and array based thin provisioned volumes in virtualized server environments. Both of these, in theory, should solve this problem because as a child VM's needs for storage increase, the dynamically expanding VHD would grow and the thinly provisioned volume assigned to it should also increase in size. While it does do that, the flaw with this approach is two-fold.

First, the use of dynamically expanding VHDs could have unforeseen performance impacts for those applications running in the child that have both a higher workload and data change rate.
Second, regardless of the type of VHD located on a volume in the parent, the Windows NTFS file system has a tendency to "wander". As it wanders, it writes bits of data to unused components of the thinly provisioned volume. As a result the storage consumed on the thinly provisioned volume increases disproportionately to the actual amount of storage that the child VM actually needs.

To account for this behavior of NTFS, best practices for the allocation of thinly provisioned storage volumes containing the VHDs assigned to child VMs call for provisioning volumes that are minimal in size. But by doing this, it recreates the storage management problem that thin provisioning was supposed to alleviate since using small thinly provisioned volumes precludes the ability for the thinly provisioned volume to provide on-demand, dynamic storage growth.

Moving SFW into the Hyper-V parent addresses these storage management issues and allows for the more aggressive use of thinly provisioned volumes in Hyper-V environments. Because SFW now manages the storage associated with each Hyper-V child VM through the parent, organizations can leverage SFW's Autogrow feature to automatically add more space to the volumes that contain VHDs for one to multiple child VM's as specific utilization thresholds are crossed on individual VMs.

For example, if storage utilization on volume containing VHDs for a child VM reaches 80%, an alert can be sent to the administrator and, if the storage utilization reaches 95%, SFW can automatically grow what essentially can be referred to as the child VM's storage pool- the volume that contains that child VM's VHDs. This is possible because, again, SFW now resides at the Hyper-V parent level so it can grab more storage from the Hyper-V parent's storage pool and add more space to that volume that contains one to many VHDs for a single or multiple child VMs.

What organizations will find equally encouraging is that Symantec is licensing SFW for the Hyper-V Server using the same licensing and pricing model that it uses for other Microsoft Windows Server products. This will allow organizations to more affordably offer baseline storage management functionality to all child VMs hosted by a Hyper-V server and potentially offer automated storage management all the way up into each Hyper-V's child VM.

As I have said in a previous blog and which I want to reiterate here, moving SFW into the Hyper-V parent opens up all sorts of new server virtualization possibilities that were simply not possible before now on file-based server virtualization platforms. Not only does it make it possible to offer high availability and path management capabilities to all child VMs but organizations can now take advantage of other technologies like thin provisioning while mitigating the adverse behavior of NTFS without breaking the bank. Granted, there are still some automation features that Symantec needs to build into SFW for organizations to fully leverage it but no doubt this represents a big step forward in the automation of management of virtualized server environments.

Part 1 of this three part series examines how file based server virtualization operating systems have failed to address high availability and performance requirements of certain applications and how Symantec Storage Foundation for Windows is addressing that in Microsoft Windows Hyper-V.

Part 2 of this three part series examines how moving Storage Foundation for Windows into the Hyper-V parent provides enhanced path and storage management functionality for Hyper-V child VMs.

The Cost of eDiscovery is Bringing the American System of Justice to the Brink of Destruction

2009-06-15T10:00:00Z

"There is no truth if you cannot find relevant evidence and, unless companies get their eDiscovery act together, eDiscovery is about to destroy the American System of Justice as we know it." That statement summarizes the opening remarks that Ralph Losey, the noted eDiscovery attorney of FloridaLawFirm.com, made during a recent presentation. From there, he went on to explain why he believes most organizations - public or private, large or small - have no viable strategy for eDiscovery and why a reactive approach to eDiscovery is putting the viability of the American System of Justice as we know it at risk.

In early June, I had the opportunity to hear Losey speak at an archiving and content management writer's conference hosted by EMC Documentum at the Hotel Gansevoort in New York City, NY. Losey began his presentation with some insightful observations and supporting statistics as to how much eDiscovery is already costing American organizations and why these costs pose such a threat to the current American System of Justice:

The cost of an eDiscovery associated for Microsoft is between $10 and $20 million dollars for each and every lawsuit.
Losey recently wrote in his blog about a case where the Washington D.C. Appeals Court affirmed an order requiring that the Office of Federal Housing Enterprise Oversight (OFHEO) spend $6 million, or 9% of its annual budget, to comply with an eDiscovery subpoena request.
Litigation is becoming too expensive so organizations are opting not to go to court and instead just settle.
The American system of justice is very different than Europe's. Europe permits the voluntary disclosure of information so European companies may only choose to disclose information that helps them in a court case. The American System of Justice requires organizations to turn over all relevant information whether it hurts or helps them in a case.
The written word has evolved over the centuries to become considered the best form of evidence. Since lawsuits involve events that occur in the past (often years ago), organizations need to be prepared to go back years to produce written documentation. The written word today now almost exclusively exists in the form of electronic communication.
In 2006, Networkworld cited a study conducted by the Butler Group that employees now spend as much as 25% of their day searching for the right information to complete a given task. Losey now believes that the percentage of time employees spend looking for stuff is closer to 40%.
The best estimates available are that during an eDiscovery organizations can only retrieve about 22% of the writings that are relevant to a case.

It is this last statistic that specifically gives Losey concern about the future of the American System of Justice. Technology has evolved tremendously over the last 20 to 30 years and it has become a real struggle for the law to keep up with this level of change. In 30 years, organizations have essentially switched from storing all of their written communications on paper to storing all of them electronically. Unfortunately, they have not adequately changed their internally processes to manage this information.

To respond to this change in information management, Losey recommends that organizations take two steps now to prevent the costs of litigation and eDiscovery from crippling or even bankrupting them in the future:

Be proactive, not reactive, about information management. The first step in the Electronic Discovery Reference Model (EDRM) is "Information Management" and yet most organizations start managing their information only after they receive an eDiscovery request and then are forced to start with some step in the middle of the process - such as the collection stage.

That is what gets companies into trouble. Starting with the collection of data as a means to do eDiscovery in a chaotic environment becomes very expensive. As a result, much of the money that organizations spend on an eDiscovery is wasted since so little relevant information is retrieved during the process.

A better way for organizations to start is by managing and archiving their email using software like Estorian LookingGlass. Software like this gives them a means to capture and manage the flow of information in and out of their organization while making it accessible and searchable if and when an eDiscovery occurs.

Prepare for a future where a random sampling of electronically stored information (ESI) becomes the norm. While it is unlikely organizations ever retrieved 100% of relevant information when it was stored in paper, achieving that level becomes even more unlikely that it is stored electronically. So to adapt to this new environments, organizations need to prepare to employ methods of random statistical sampling of their ESI and then prepare to defend this method of eDiscovery in court to keep eDiscovery costs from spiraling out of control. Losey says, "Common sense dictates that sampling and other quality assurance techniques must be employed to meet requirements of completeness."

If an organization is not already employing some means to archive and manage its ESI, it is going to lose in some way - the only question is how much are they willing to lose? Even winning a court battle may only be a Pyrrhic victory as it may be cheaper for an organization to settle out of court than defend itself.

Yet the greater danger that the American population faces is more than the destruction of the current American System of Justice. It begins to change who we are and what we are as a nation as it creates an environment where only the rich can afford to litigate and ultimately exonerate themselves of any wrong doing. As for the rest of us, we may be looking a future of forcing to settle outside of court for we can not afford to uncover the truth and justice is never fully measured out.

Fact-Based, not Faith-Based, Confidence Needed When Implementing eDiscovery Processes

2009-06-12T10:00:00Z

Regardless of how one approaches an eDiscovery project, having processes and tools in place to help guide you through the EDRM (electronic discovery reference model) are critical elements that influence how effective an eDiscovery project will be. Data retention policies, access to outside resources, and technology are all critical components to have in place in order to successfully complete an eDiscovery.

Yet an equally important question that organizations need to answer is how reliable is the information they discovered in their eDiscovery? Or, better put, how do they move from a faith-based approach of eDiscovery where they assume they have all of the information that they need to a fact-based approach where they have confidence that all of is the information found during the eDiscovery is accurate and defensible in court?

An eDiscovery strategy that does not provide accurate results and/or is not defensible in court could bring serious sanctions such as a negative inference ruling against an organization. Therefore when framing an accurate and defensible eDiscovery strategy, organizations should test their eDiscovery strategy as well as ensure it is applicable in the following areas:

How is the integrity of ESI (electronically stored information) maintained?
Is critical metadata maintained so as to provide a complete accounting of requested data?
Is chain of custody maintained throughout the process?
How is searching of ESI performed?

Using technology such as Kazeon's Information Server provides the ability to meet the accurate and defensible mandate. An independent study entitled "Search Accuracy and Relevance: Critical for Defensible eDiscovery" released by Reason-eD, LLC, shared its results after putting the Kazeon Information through a battery of tests. The study found:

Integrity, metadata and chain of custody were retained while Kazeon Information Server scanned and collected ESI.
Kazeon Information Server found unique search terms in over 40 file types during searches of large, complex ESI collections.

Reason-eD concluded that using the Kazeon Information Server organizations could put in place a mechanism to deliver an auditable, understandable, documented, and automated process for answering an eDiscovery request. But more importantly, organizations could use its results from the Kazeon Information Server as a means to defend in court the actions they had taken in responding to an eDiscovery request.

However implementing an eDiscovery solution is not enough - organizations need to test the solution as well. In a recent webinar, George Socha cautioned against just believing that an eDiscovery solution will work but that they actually need to test the solution as well. Companies that believe these processes will work without testing it are adopting a high risk strategy that is sure to show vulnerability if and when closely scrutinized.

Again, this is where solutions like the Kazeon Information Sever can effectively kill two birds with one stone. Reason-eD's testing showed that the Kazeon Information Server provides an eDiscovery solution that gives organizations a sense of confidence based on fact not faith that they are presenting accurate and defensible information before it comes under scrutiny in court.

Being able to answer the simple "How do you know and know that you know?" question is pivotal in building an accurate and defensible eDiscovery strategy. As Reason-eD's study showed, using an eDiscovery solution such as Kazeon's Information Server ensures that the integrity, metadata, and chain of custody are preserved during the eDiscovery process while providing an auditable process that can help organizations justify their position in court. In so doing, organizations can provide information requested during an eDiscovery process and know that their results are based on fact, not faith.

Branded Solutions or White Boxes for Geeks: Bell Micro Supplies and Supports Both

2009-06-11T10:00:00Z

Build your own (white box) or purchase an industry-branded solution for implementation of appliances - that is the question. Even today, this question continues to plague many customers as well as divide OEM businesses. Depending on the equipment (components or solutions) OEMs are quick to argue their case as to why white boxes are better than a branded solution or vice versa. However each solution has its place in the market. And, for the smart OEM, maybe the choice doesn't have to be one or the other.

White Box. Companies looking to build a solution that solves a problem in a very unique way - usually faster application performance - typically includes a specific hardware element in their solution. As a result, they are at the forefront of technology and innovation and require newer and faster components that deliver a superior, combined software and hardware solution.

Choosing an industry-standard solution is almost out of the question since they have designed their own computing or processing architecture while leveraging some kind of real-time operating system. Time is of the essence and building a very specific purpose-built piece of equipment, such as a specialized card that they put in a system to leverage some aspect of the x86 architecture, is the only way to speed their time to market.

Branded Solution. It's no secret that many software vendors are moving to a branded appliance platform as it can more easily solve a set of problems for those vendors that in the past have deployed solutions on a white box platform. Customers deploying branded equipment typically don't need the state of the art or bleeding edge technology to offer a great solution. All that they require is a solid technology with robust options that they can take advantage of.

Industry-branded solutions enable companies to quickly go to market on brand name hardware. The hardware provider includes its own engineering and global field support and has gone through a number of certification and qualification processes that meet the requirements of individual customers. But the downside that OEMs can face is if they for some reason need to tweak the standard configuration of the hardware vendor's branded solution, they can potentially break - and void - the hardware vendor's warranty or support contracts.

The challenge that some suppliers like Bell Micro face is how to compete against branded solutions when their legacy is delivering white box solutions. Though Bell Micro continues to have success with customers who use white box solutions, it finds that it is competing more frequently against players like Dell. As a result, only offering white boxes is a huge obstacle when a customer needs or wants an industry branded solution.

But the face of Bell Micro is changing. What DCIG sees Bell Micro doing is redefining itself through its OEM relationships with partners such as HP in a way that is truly unique in the market place. Bell Micro is expanding its set of solutions to include both white box and branded solutions so it can provide customers a real choice that makes sense for their particular solution while eliminating some of the downsides that using branded solutions typically present.

If a customer wants the benefits of a branded solution so they can take advantage of global service, standardization, and certification, Bell Micro can help provide an industry branded solution from HP without compromising on application requirements. What makes obtaining an HP-branded offering from Bell Micro unique is that OEMs can work with Bell Micro to develop the branded HP solution to meet their specific application requirement without voiding the standard HP warranty or support. Conversely, if the customer wants complete control over the configuration and a stable configuration with a really long life, Bell Micro can help get that customer into an appropriate white box solution.

The point is that Bell Micro supports both solutions in a unique way. Bell enables OEM customers to integrate third-party components into HP servers and still retain the warranties and support agreements. This now gives OEMs a single supplier to call whether they want to pursue the white box option where Bell Micro is the industry leader in these solutions or a branded solution that meets their specific needs without voiding the hardware warranty or both.

To educate OEMs about this new offering, Bell Micro's strategy is to talk to customers, let them know they now offer a complete range of solutions and then help them step through the business drivers that will help them make a decision between a white box or appliance solution.

Acting as a trusted advisor, Bell Micro understands that there are customers with different needs in different product lines. Now due to its new and enhanced relationship with HP, Bell Micro can help deliver solutions that meet either of their customers' needs. In this role, Bell Micro can leverage either approach to address their customers' needs if and when your needs change (such as when the next version of a product is released) so that they can still leverage Bell Micro without having to change their supply chain just because their choice of architecture changes.

It's Time to Contemplate Consolidated Recovery

2009-06-10T10:00:00Z

Bounce the phrase "consolidated recovery" off of most individuals in IT and you are just as likely to get a blank stare as a good answer as to what it means or how to accomplish it. Most IT staff keeps so busy on a day to day basis just managing their assortment of backup, clustering and replication products that they never get much beyond focusing on the protection and recovery of each application. So for them to contemplate the consolidation of protection and recovery using just one methodology has more than likely not even crossed their minds.

Most organizations have different tiers of protection and recovery simply because that it is the conventional wisdom of how they should approach data protection and recovery. For example, organizations use:

Backup software to protect applications that can typically withstand longer outages (two (2) hours or longer) and/or are tolerant of recoveries of application data that is hours, days or even weeks old.
Storage system-based replication software to replicate data from one storage system to another (usually located at another site) for applications that they need to recover in minutes to hours
Clustering software that is either operating system based such as Windows 2003 Microsoft Clustering Server (MSCS) or Windows Server Failover Clustering or replication software that is part of an application such as Microsoft Exchange's Standby Continuous Replication (SCR) or Cluster Continuous Replication (CCR). These clustering and/or replication technologies provide recovery of the application and/or application in minutes or tens of minutes.

So if such a product did exist, it would have to satisfy all three of these different approaches to data protection and recovery. Each one addresses a specific business need that would not disappear if they were consolidated. So some key attributes that such a product would need to possess include:

Be cost effective. This is probably the biggest point in the minds of end-users. The primary reason that organizations continue to use tape instead of disk is that backup is viewed as an expense and right now the last thing any organization wants to do is spend more money on backup. Since this product would need to protect everything from lowest tier to the highest tier application server, it needs to be sensitive to the cost constraints of these lower tier application servers which make up the bulk of servers in most organizations.
Simple to install, configure and manage. IT needs less complexity, not more, and while a consolidated recovery product sounds great on the surface, if they have to baby sit it all day long, what does IT gain? Whatever the new product looks like, it cannot require more time to manage and ideally should require a lot less while delivering the benefits of these other solutions.
Create a point in time snapshot of their application data. At a high level, this is what backup software does on a daily basis even though rarely does anyone refer to it this way. Organizations will continue to need a static point-in-time copy of their data so they can recreate what their environment looked like at that point in time should a data corruption occur or they are subjected to an eDiscovery search when they need to substantiate who knew what and when.
Replicate data locally or remotely irrespective of the application or operating system. As organizations continue to move up the stack in their application server tiers, some applications will require near-immediate recovery (within 30 minutes to an hour) either locally or remotely. So the product must provide asynchronous replication and support data recovery anywhere they need it when they need it.
Provide near-real time failover. Clustering software is often put in place based on the assumption that it provides real-time failover. While this is true in some cases, if more organizations look closely at how the clustering software on their mission critical servers is configured, it is likely configured in Active-Passive mode where if one node fails, the other takes over the application processing within minutes at best and can take more than an hour in many environments.Any new solution should therefore meet these requirements of recovering an application in minutes or tens of minutes locally or remotely.

There are obviously other requirements that such a "consolidated recovery" product would need to have. But from these points, one can begin to get a picture of what a product that delivers "consolidated recovery" would need to look like. In short, it has to deliver on the best elements of backup, replication and clustering software without requiring all of the associated costs and management overhead that managing them individually entails now. In a forthcoming blog, I will take a look at how InMage Systems' DR-Scout is already positioned to deliver on this emerging concept of consolidated recovery.

The $20 Billion Dollar Mid-Market Deduplication Opportunity; Interview with ExaGrid Systems' CEO Bill Andrews

2009-06-09T18:00:00Z

It's time to start thinking ahead. Over the next 60 - 120 days someone is going to acquire Data Domain - be it EMC, NetApp or some other suitor that may yet emerge. That means Data Domain, today's leading mid-market deduplicating disk-based appliance player, will be changing its colors. While that occurs, it presents a unique opportunity for ExaGrid Systems, the #2 player in the mid-market deduplication space, to exploit and grow its presence. To discuss how and if ExaGrid can take advantage of this opportunity, I recently met with ExaGrid's CEO, Bill Andrews, to discuss this development, his perspective on the deduplication market as a whole and how ExaGrid stands to benefit (or lose) from Data Domain's acquisition.

Jerome: How do you see Data Domain's pending acquisition by EMC, NetApp or whoever affecting the industry?

Bill: The biggest thing that this acquisition signals to me is that the deduplication technology that EMC and NetApp are using right now for the mid-market to small enterprise isn't competitive. Why else would they be willing to pay $2 billion for Data Domain's deduplication technology? But as this situation over who acquires Data Domain plays out, there is a lot going on in the background.

This acquisition is going to make HP and HDS very nervous. HDS is exposed for letting Diligent Technologies slip through its hands and get purchased by IBM. Then, once it let that one get away, HDS retrained its entire staff on Data Domain and started selling that. Now Data Domain is off the market. I predict that the likelihood exists that SEPATON will be acquired by HDS within 90 days as former SEPATON employees such as CTO, Miki Sandorfi, have already found a new home at HDS.

HP is harder to predict. They are using SEPATON at the high end but do not have a really good answer for the mid-market. HP has its own deduplication solution but we rarely if ever encounter it and it seems best suited for the low end of the market (1 TB or less of data). Another negative is that it only offers a VTL interface which we find that no one in the mid-market, much less in the low end of the market, wants.

Jerome: What defines the mid-market deduplication space and makes it so special?

Bill: Environments with less than 1 TB of data are better suited for a service play, low cost disk or backup applications with built in deduplication as the amount of data and retention is low. Conversely environments with 60 TBs or more of production data generally already have FC SANs and will want VTL.

Mid-market to small enterprise disk-based deduplication appliances are specifically targeted for environments with 1 - 60 TBs of primary data which is where ExaGrid plays and wants to remain. Further, the mid-market is anti-VTL.

Any time a customer inquires about a VTL and ExaGrid starts to explain the labor associated with setting up virtual tape drives and tape cartridges and then managing them, they don't want that much complexity in their backup environment. So ExaGrid only offers a NAS interface because that is what the mid-market to small enterprises wants because they understand NAS and it is plug-and-play to their existing backup application.

Jerome: Assuming only EMC and NetApp remain in the bidding for Data Domain, who would you prefer wins?

Bill: On previous earnings conference calls, Data Domain had already intimated it was starting to leave the mid-market and go upstream to the enterprise which ExaGrid has detected in the field. So when I heard that NetApp was acquiring Data Domain, I popped the cork on a bottle of champagne and started cheering.

NetApp lives in the Fortune 2000 zone and that is where it primarily focuses. So if NetApp acquires Data Domain, ExaGrid expects NetApp will take about 6 months to figure out how to best position Data Domain within the organization, re-arrange Data Domain accordingly and then make the appropriate changes to sales and product management internally within Data Domain.

Once those changes occur, I expect NetApp to be much more focused on the Fortune 2000 accounts which leaves the door wide open for ExaGrid in the mid-market to small enterprise as ExaGrid focuses mostly on the Fortune 2000 down. ExaGrid now goes toe-to-toe with Data Domain about 50% of the time that ExaGrid is involved in a competitive situation which largely goes away after NetApp acquires Data Domain. Bottom line, I would love to compete with NetApp.

However once EMC entered the fray, I put the cork back in the champagne bottle. EMC has a much broader reach than NetApp by as much as 4-5x and ExaGrid sees EMC in many more deals now. While we almost always beat EMC, I can see that changing if EMC acquires Data Domain. However, ExaGrid's win rate against Data Domain is over 70% so I suspect that ExaGrid can continue to win against EMC.

What further concerns me is EMC's long standing relationship with Dell. I don't know that Dell would resell an EMC-branded Data Domain appliance but it only makes sense that Dell would.

On the upside, ExaGrid has been anticipating this battle for a few years and we already knew the final battleground would involve Dell and HP. ExaGrid is preparing for that battle in 2010 with a very differentiated product and price leadership.

Jerome: It is DCIG's opinion that EMC was preparing to acquire Data Domain but that the NetApp offer forced EMC to move more quickly than it anticipated. What have you heard?

Bill: Rumor has it that NetApp started talking to Data Domain last year about acquiring Data Domain and the more NetApp talked to Data Domain, the more it liked of what it saw. I believe NetApp especially saw a culture that was compatible with NetApp. So eventually the first deal you saw announced was the one crafted by NetApp and Data Domain.

The minute it was announced, however, EMC had an 'Oh my God!' moment as I believe it may have caught EMC completely flatfooted. EMC could not afford to be left without a viable offering in the mid-market deduplication space as it stands to be a $20 billion market and here is how I arrived at that number.

The average sale size for a deduplication appliance is $50,000. Assuming there are 400,000 opportunities out there that converts to $20 billion in sales over the next 6 - 7 years. So let's just consider this year a wash in terms of sales if EMC acquires Data Domain and it incorporates its product line into its sales pipe line. I can easily see it attaining sales in year one (2010) of $1 billion; year two could result in $2 billion dollars in sales; and every year thereafter $3 billion in sales or more. That's over $20 billion in sales in the next 7 years.

Bottom line, if EMC acquires Data Domain, the mid-market deduplication space will turn overnight. Paying $2 billion for Data Domain is a drop in the bucket considering the market opportunity that is in front of EMC right now.

Jerome: So where does ExaGrid go from here?

Bill: Midsize to small enterprise organizations are under tremendous pressure to switch from tape to disk. The problem is they do not have a lot of answers from which to choose.

When you look at enterprises with over 60 TBs of data and small organizations that have less than1 TB of data, they have lots of answers. But the mid-size market has very few answers since its problems are symptomatic of both large and small organizations.Midsize companies have relatively large amounts of data (typically ranging from 5TBs to 30 TBs and beyond) and insufficient staff to manage tape which is why they are looking to use disk.

ExaGrid already has over 350 customers and even if ExaGrid had only 3% market share, by 2011 it would have $100 million in revenue and be well positioned for a public offering. ExaGrid plans to keep its head down and stay entirely focused on the mid-market to grow its presence there.

Veritas Storage Foundation for Windows Achieves Path and Storage Management Breakthroughs in Server Virtualization

2009-06-09T10:00:00Z

Consolidating multiple physical machines onto one physical server that hosts multiple virtual machines (VMs) requires that organizations consider much more than just if the new server has sufficient processing, network bandwidth and memory to support the applications. Specifically, path and storage management issues can result when physical servers are virtualized which can preclude organizations from virtualizing some of their mission critical servers. However the recent announcement that Symantec now supports the movement of its Veritas Storage Foundation for Windows 5.1 from the child to the parent partition on Microsoft Hyper-V Server breaks through these issues and opens the door for organizations to virtualize all of their physical servers.

Symantec's decision to move Storage Foundation from Hyper-V guests down to the Hyper-V parent was driven, in part, by calls that its support team was receiving from users in EMEA (Europe, Middle East and Asia) that were deploying Storage Foundation at the Hyper-V parent layer. These users needed software storage management features on their Windows platforms for campus clustering which, as part of the deployment, had a requirement for mirroring. These customers wanted to use software mirroring to abstract the storage team out of the solution so their system admin teams could manage the disaster recovery (DR) solutions without directly involving the storage admin team.

At the time that these customers moved Storage Foundation from the Hyper-V child to the parent, it was done unbeknownst to Symantec since Symantec did not officially support this configuration. Yet Symantec and its customers discovered some unexpected benefits when Storage Foundation was moved into the Hyper-V parent. Aside from the fact that it worked, it addressed a persistent problem that file system based server virtualization solutions like Microsoft Hyper-V currently present - the inability to provide advanced path and storage management features to guest VMs.

When storage volumes are presented by the storage system (either direct attached or via a storage network) and detected by the Microsoft Hyper-V parent, the Hyper-V parent virtualizes all of the storage. From this storage pool, Hyper-V parent administrators create multiple Microsoft Virtual Hard Drive (VHD) files. Each of these VHD files contains a partition of the Hyper-V parent's storage pool is assigned to and then owned by that VM. The difficulty with this approach is that once a VHD is assigned to a VM, it becomes very difficult to perform any further advanced path and storage management features on that VM and the VM may lose some or all of these features that it previously possessed when it was a physical server.

Running Storage Foundation in the Hyper-V parent changes this. Storage Foundation turns the devices presented by the array into dynamic disk, creates dynamic volumes on the dynamic disk and formats these volumes with NTFS all within the Hyper-V parent. Hyper-V then places VHDs on these dynamic volumes which are presented to individual child VMs.

The new variable in this equation is that each and every VHD now sits atop of a Storage Foundation dynamic volume so the path and storage management features normally found in and managed on the guest VM are now located in and managed at the Hyper-V parent level.

This represents a major breakthrough in path and storage management for file system based server virtualization that Microsoft Hyper-V and Storage Foundation are the first to deliver to the best of my knowledge. A Microsoft Hyper-V server with Storage Foundation now gives each of its child VMs access to path and storage management features that they normally would have lost when they were virtualized (there are some exceptions when using iSCSI that I won't get into here).

Storage Foundation can now communicate directly with the actual volumes of the underlying storage system as well as the physical server's Fibre Channel (FC) host bus adaptors (HBAs). This keeps legacy path and storage management features remain intact since the VHD assigned to the VM is using a dynamic volume created by Storage Foundation that can deliver this functionality.

So what does this practically mean for end-users in the near term? Two things:

Advanced path management for applications running in FC environments is back. Storage Foundation can now detect and communicate with the physical FC HBAs on the Hyper-V Server and send traffic from each VM down separate physical paths. This delivers load balancing and high availability for each VM on the Hyper-V server should a FC HBA fail or if a FC network path goes down. Further, if a FC HBA on the Hyper-V server starts to become congested, Storage Foundation can detect this congestion and route more of the I/O traffic down a less congested FC path.
Administrators can now grow storage volumes containing a specific VM's VHD or set of VHDs or even move that volume containing one or more VHDs to other storage without taking the VMs offline. Again, because each VHD is using Storage Foundation dynamic volumes beneath the surface, Storage Foundation's functions are now extended and available to every VM on that Hyper-V server.

File system based server virtualization solutions such as Microsoft Hyper-V and VMware ESX server had, up to this point, faced a significant obstacle to overcome in terms of delivering the path and storage management features that organizations needed to justify virtualizing their mission-critical servers.

In this particular case, Symantec with the assistance of its customers in EMEA, accidently discovered the breakthrough that will make Microsoft Hyper-V a much more viable alternative to the VMware ESX server in the near term while laying the groundwork for some exciting new possibilities going forward. In a final blog entry in this series evaluating this new application for Storage Foundation for Windows 5.1, I'll get into what new opportunities moving Storage Foundation for Windows in the Hyper-V creates for organizations.

Part 1 of this three part series examines how file based server virtualization operating systems have failed to address high availability and performance requirements of certain applications and how Symantec Storage Foundation for Windows is addressing that in Microsoft Windows Hyper-V.

Part 3 in this three part series examines how moving Storage Foundation for Windows into the Hyper-V parent addresses some current issues around thinly provisioned volumes as well as how Symantec keeps the licensing for Storage Foundation for Windows attractive.

The Coming New World of Healthcare Data Management Demands Healthcare Savvy Data Management Providers

2009-06-08T18:00:00Z

Obama's administration allocated $17B of the recent (Feb 09) stimulus spending package to healthcare, for the purpose of building better healthcare infrastructure. The goal of the new infrastructure is to move patient records online and enable a ubiquitous Electronic Health Record (EHR) to be shared universally among hospital systems. Obama himself promised a total of $50B in spending for this purpose during his campaign. Some experts believe that even more is to come. But, now that some of the money is allocated, how are healthcare institutions getting access to it and what are they doing with it?

Recently, I sat down with the data management team at BridgeHead Software. BridgeHead offers data management software exclusively for hospital systems, leads in protecting Meditech system data, and serves over 1000 hospitals. According to BridgeHead, getting access to the stimulus funding has been a lot of "hurry up and wait" for hospital IT teams. This is despite the advice of HIMSS (the Healthcare Information and Management Systems Society) to healthcare institutions, which says "don't be a procrastinator" to apply for funding and begin infrastructure enhancements.

Part of the problem is that the criteria for "meaningful use" of the stimulus funds has not yet been defined. Also, most hospitals have some type of EHR system already in place including systems from Siemens, Cerner and McKesson. From a certain perspective, these teams may believe that "there is nothing for me to do." After all, they already have at least some of their patient data online where they are accessible to support patient care.

One key question, however, is whether those systems encompass everything required in the EHR - including scanned paper documents, email, insurance records, patient histories, prescription records, DICOM images, and more - and most do not. All of that EHR data must also be accessible and shareable with other hospital networks and systems. Putting all of that data online also brings up questions like "who really owns the patient record?"

One outcome of this confusion, as experienced by the BridgeHead team, is that some hospitals are becoming very entrepreneurial. These hospitals have seen the need, and they plan to fill it. They have gone beyond their own institution to establish data systems infrastructures that serve entire regions of hospitals and point-of-care facilities. Known as Regional Centers, these facilities have the potential to grow into Regional Health Information Organizations, or RHIOs (pronounced Ree-Os).

RHIOs are the key building blocks necessary for establishing the nation-wide health network envisioned by Obama and could also provide local physician offices with access to online patient records, eliminating the paper records which have become such a problem to manage and eventually, dispose.

Those at BridgeHead believe that the key to making EHR practical is, in part, making sure that the underlying data management system is designed to support all EHR requirements. This includes managing all of the types of data that make up the EHR, as well as enabling fast search and retrieve of that data from anywhere in the managed repository. Without EHR-aware data management systems, the associated cost of storage can spiral out of control, data records may be unsearchable or searchable only with multi-step, time-consuming processes, and patient records will be next-to-impossible to share across systems.

With EHR-aware data management solutions, however, hospitals have the foundation in place to meet current and future requirements. EHR-awareness includes integration with the applications producing all of the types of data that contribute to the EHR. Applications producing EHR data include digital imaging (DICOM) systems, but also include Microsoft Exchange for email data, SQL Server databases, and scanned document files. EHR-aware data management helps to tie all of these types of data together into a single logical repository, from which the records can be more efficiently searched and more quickly retrieved.

In short, EHR-aware data management provides lower-cost storage retention of the images, faster restore, along with better search and retrieve capabilities. But EHR-awareness is more critical than these benefits. With the coming new world of healthcare patient records going online, EHR-aware data management enables a bridge from old to new world data management. BridgeHead Software is aptly named, providing EHR-aware data management that solves data management challenges for their traditional hospital customers while also preparing those hospitals to offer data services within their regions.

For hospitals that want to take the step of becoming a RHIO, healthcare-savvy data management vendors like BridgeHead Software can be expected to be an essential part of the EHR infrastructure.

Dell and CommVault Bring Intelligent Plug-n-Scale Deduplication to the Midsize Market

2009-06-08T13:00:00Z

If one didn't know any better, one would think that deduplicating backup data is going to solve all of IT's backup pains. The current train of thought goes something along the lines of "Plug in a deduplicating appliance, point the backup software at the new appliance and, Voila!, the backup problems are solved." The only problem with that viewpoint is that deduplicating appliances alone do not solve equally pressing corporate data management problems and may even create new backup and data management challenges along the way.

The two potential problems that organizations that select a deduplicating disk-based backup appliance may encounter include:

Inability to scale the solution. As the amount of backup data grows, so do the requirements of the backup appliance, either from a capacity perspective, a performance perspective or both. Once one or the other of these two features on the deduplicating appliance is maxed out, organizations must purchase another appliance - either a secondary one or a larger one.
Silos of data. The problem of data silos emerges if a company elects to purchase a secondary appliance to complement their existing appliance and then do not migrate their existing deduplicated data to a larger appliance. This is likely what many companies will do for one very simple reason: economics. But when they do so, not only do they have to deduplicate the backup data from scratch again, they now create another repository where their backup data resides.

This does not mean some organizations should avoid deduplicating disk-based backup appliances. As long as it is a small organization with minimal amounts of backup data (under 1 terabyte of data), it will likely not encounter any issues since these are the types of loads these appliances are designed to handle. But if it is a midsize organization that has tens or even hundreds of TBs of data under management that require backup, it needs to take a moment and think about how best to proceed with not just deduplication but data management as well.

Deploying one of these deduplicating disk-based appliances will likely solve the immediate backup pain in a midsize organization. But in 3, 6, 9 or 12 months when the appliance is at capacity or maxed out on throughput, then what? Spending a huge amount of money to upgrade and migrate to a larger appliance is rarely seen as desirable by anyone in this economic climate so introducing a second less expensive appliance becomes the natural choice.

By doing so the organization starts going backwards again in regards to solving its backup problems. Now its IT staff is forced to start balancing the backup jobs between two (or more) appliances as they lose the benefits of deduplication from the first appliance since the second one does not have the deduplication index from the first and two silos of data are created that they need to manage.

So what's the alternative for midsize enterprises to just "throwing" deduplicating appliances at their backp problem? Well, you can still "throw" a deduplicating appliance at the problem, but just throw one at the problem that deduplicates and manages the data in a different manner. That's what today's announcement of the Dell PowerVault DL2000 bundled with the CommVault® Simpana® 8 with block deduplication is designed to solve. Because rather than just deduplicating the data after it is sent to the disk-based target, Simpana deduplicates the backup data before it is stored on the disk target.

The DL2000 offers a number of distinct advantages for midsize organizations.

First, it means that the DL2000, not a deduplicating backup appliance, manages the deduplication index. Because it manages the index, the DL2000 can store data on any PowerVault system - be it a Dell MD1000 disk array or an ML6000 tape library. So as an MD1000 disk array fills up an organization can leverage Simpana to move deduplicated data from disk to tape while losing none of the deduplication benefits that Simpana provides.
Second, using Simpana goes to the heart of what organizations should ideally be trying to accomplish. Deduplicating backup data on an appliance solves the immediate backup pain but does little or nothing to solve data management problems that are brewing beneath the surface in every organization. Because the DL2000 delivers a full-featured version of Simpana 8, if organizations want to take advantage of its archiving, content indexing or search capabilities later on, the foundation is in place for them to do so.
Third, if an organization uses advanced features like archiving on the DL2000, it solves another problem that deduplicating appliances do not - a means to stem the long term growth of data. By archiving aging and stagnant data on the front end, it reduces the amount of data that organizations need to backup which further shortens backup windows and recovery times and reduces the amount of data that the DL2000 needs to manage as part of its deduplicated data store.
Finally, midsize organizations are also looking to both consolidate and simplify the management of their backup data in remote offices. By deploying DL2000s both locally and remotely, all of their sites gain the benefits of deduplication plus they can use Simpana's replication feature to replicate deduplicated data from remote sites to a central site. Organizations thereby create a standard interface that they can use to manage backups, deduplication, recoveries and replication.

Deduplication is becoming much more than a backup feature - it is becoming an integral component of how backup is done. But organizations should not assume that there is only one way to do deduplication - the DL2000 proves that. So midsize enterprises especially need to look before they leap with deduplication to make sure that in the process of solving today's problems associated with backup and recovery they do not create larger problems tomorrow. The new DL2000 with CommVault Simpana 8 is one way that midsize organizations can avoid this scenario as it delivers deduplication while putting in the foundation to solve the longer term problems that deduplication can potentially create.

The Hidden Costs of Unreliable Disk Drives; The Dirty Little Secrets of Managing RAID Storage

2009-06-05T10:00:00Z

It is only after organizations start to deploy storage systems with SATA HDDs that they really start to focus on the reliability of SATA HDDs. While it is unlikely they will lose any data regardless of whose system they deploy, the amount of time that their IT staff spends and resulting operational expense in managing the replacement of failed SATA HDDs can and will vary widely according to which storage system they select. It is only by selecting SATA storage systems that specifically account for the idiosyncrasies of SATA HDDs that organizations can both protect their data and not have to deal with headaches of constantly replacing failed SATA HDDs.

Organizations are storing more data than ever on disk. Archives, backups, DR and video surveillance data along with unstructured data are largely contributing to the explosive growth of data in organizations. Yet one of the dirty little secrets of managing the large disk farms needed to store all of this data is managing the replacement of failed SATA hard disk drives (HDDs) in these disk farms. While current RAID technologies do an adequate job of protecting from data loss in most of these environments, when a SATA HDD fails, it still does require someone to replace it.

Replacing failed SATA HDDs may be no big deal in smaller environments. But when you start to consider how potentially unreliable some SATA HDDs are and the time involved with managing their replacement in large disk farms, the process becomes much more complicated. Here are just some of the steps that I had to follow when I worked at a Fortune 500 data center and had to replace a failed HDD (SATA or otherwise):

Open a trouble ticket in my organization's change control system
Open a trouble ticket with the vendor to replace the disk drive
Determine the urgency of replacing the failed disk drive. (i.e. - was it part of a RAID 1 array on a storage system with no hot spare used by a mission critical application or was it in a RAID 5 group on a storage system with a hot spare used for backup?)
Schedule a time for the HDD replacement. Depending on what applications the failed HDD supported, I might schedule the failed HDD to be replaced as soon as a new one arrives or I might wait until the middle of the night so it can be replaced during normal maintenance hours.
Notify the affected application, server, change control and security teams. They need to be notified so they can monitor for any performance impact to their applications or servers as well as be on the lookout for the technician so he can access the data center floor to replace the failed hard drive and install the new one.
Verify the new drive was successfully installed and close out the open trouble tickets.

While not every organization has to go through all of these steps to replace failed SATA HDDs, regularly replacing failed HDDs becomes a cost and a risk to any company. And let's face it, organizations are using SATA HDDs for more applications than ever before and SATA HDDs are the ones increasingly tapped for these functions.

So while organizations obviously do not want to lose data on their SATA storage systems (and probably will not because the data is protected with RAID), they also do not want to dedicate a full time person to manage the task of replacing failed HDDs in their growing disk farm. Yet storage administrators that manage large disk farms often complain about this so they are thinking more about this issue ahead of time and looking to buy storage systems that mitigate this problem.

Organization should therefore look for storage systems where replacing failed SATA HDD remains an occasional hassle and does not become a full time job. While this should not be considered a complete list, here are some features that organizations should look for in SATA storage systems to ensure high reliability of the SATA HDDs:

Select storage system manufacturers that have a history (5+ years) of working with SATA. Storage system manufacturers who started in the early days of ATA (pre-SATA) know how poorly those ATA drives were manufactured so they built mechanisms into their storage systems to account for those deficiencies. While today's SATA HDDs are much more reliable, they still have their own little idiosyncrasies that these manufacturers are more likely to know about and account for in the design of their storage system.
Manufacturer only uses enterprise SATA HDDs. Enterprise SATA HDDs are measured by different standards than SATA HDDs intended for the desktop. Enterprise SATA HDDs are tested using different workloads, have longer burn-in cycles, are rated for longer mean times between failures and include 5-year warranties. Desktop SATA HDDs have none of these features so they are more prone to failure.
Manufacturer stress tests the HDDs before deploying them in the system. Even enterprise SATA HDDs are not immune to failures so inquire as to how the storage system manufacturer stress tests the HDDs before they put them into the field. So for example, when the storage system is built and tested at the manufacturer's site, the manufacturer should ideally have software built into its storage systems that weeds out faulty drives. One test they can run is to perform random I/Os on individual HDDs and measure the time it takes to perform these tests. Response times that take too long are an indication that the HDD is prone to failure.
Manufacturer can manage HDDs when they are spun down. Since data stored on SATA HDDs is frequently not accessed after it is stored, newer enterprise SATA HDDs have the capability to spin down and go to 'sleep' without powering off However, the storage system needs to recognize and manage this behavior so it does not keep trying to spin the HDD up or incorrectly label it as a failed HDD.

Once organizations know about some of these finer points that SATA storage system manufacturers take (or do not take) to ensure the reliability of the SATA HDDs within their systems, it becomes easier to justify choosing one over another for these types of hardware benefits. For instance, Nexsan Technologies is a prime example of an organization that has a long history of working with SATA HDDs (10+ years) and has taken all of these steps and more to ensure the reliability of SATA HDDs on its many products which include SATABoy and SATABeast. Nexsan's CTO Gary Watson says, "Our customers frequently tell us that their Tier 1 storage systems often have higher HDD failure rates than their Nexsan systems with SATA HDDs."

Most organizations say that when they are contemplating the use of SATA HDDs that their primary concern is their reliability. In truth, most are more initially concerned about the protection and recoverability of their data which is a fear most SATA storage system manufacturers address through the use of RAID. But RAID only addresses concerns about data reliability, not hardware reliability, and as customers can find out after the fact, reliable SATA HDDs have a value that organizations may only appreciate and understand after they purchase an unreliable storage system.

Microsoft's Best Practices for Exchange Storage Configurations often Assume DAS not SAN

2009-06-04T10:00:00Z

Microsoft provides a large number of guidelines for how to properly configure storage systems that are used in conjunction with Exchange implementations. But an area where Microsoft often still comes up short is in providing best practices for configuring Exchange in conjunction with today's next generation storage systems. Many of Microsoft's storage recommendations are based on the assumption that organizations are deploying either direct attached storage (DAS) or storage systems with traditional RAID architectures. But with next generation storage systems such as the 3PAR InServ Storage Server that deliver features such as wide striping, old rules for Exchange storage configurations do not always apply.

To better understand how to balance existing Microsoft recommendations for Exchange storage configurations with new capabilities on modern storage systems, I recently had the opportunity to discuss these issues with Bill Plein, a storage architect on 3PAR's Strategic Accounts Engineering Team. Bill serves as the liaison between 3PAR and its clients in their Exchange deployments and endeavors to provide storage configurations that best satisfy customer requirements for capacity and performance while still falling in line with Microsoft recommendations which is not always easy to do.

A number of the problems that Plein encounters in the field stem in part from assumptions originally made when Microsoft developed its best practices for storage configurations for its release of Exchange 2007. As part of that release, Microsoft tried to correct some complexity and performance issues that arose from Exchange 2003 deployments where there were numerous suboptimal disk configurations. In particular, smaller organizations were encountering issues when deploying Microsoft Exchange on SANs as they had few, if any, best practices for managing Exchange on SAN implementations in place.

To try to improve this situation, Microsoft came out with a very strong framework around DAS for those shops that had smaller storage configurations. The problem that Microsoft specifically was trying to solve was to dispel the notion that a SAN was necessary to support Exchange. Microsoft found that in these smaller environments introducing a SAN actually added complexity and reliability problems to an Exchange implementation. To counter this "SAN is a requirement" argument, Microsoft developed many of its best practices from the viewpoint that organizations did not need to deploy a SAN in order to support Exchange and framed many of its recommendations on how to support Exchange using DAS.

The trick is that as you move from small to large organizations, Microsoft's recommendations for best practices for Exchange storage configurations do not change to account for the storage technologies that these size organizations use. Rather they are tailored for the lowest common denominator such as what you might encounter in a small organization. So when large organizations roll out Exchange and call in an Exchange expert to help them configure and implement it, the expert may view storage from this more limited DAS viewpoint and does not recognize new features that next generation storage systems offer.

A prime example of this is the Windows 'diskpart' command. Exchange best practices call for administrators to run the 'diskpart' utility to align a partition with the underlying physical disk. If you do not run this command from Windows, Exchange data may be written to two disks instead of one because of how Windows, by default, starts a partition on the last sector of a disk. However by running 'diskpart' before installing Exchange, you can avoid unnecessary disk or stripe crossings since 'diskpart' starts a partition on the first sector of a disk so there is an increased likelihood that all Exchange data ends up on a single disk drive which improves Exchange performance.

Unfortunately 'diskpart' assumes that it is working with a traditional RAID architecture such as RAID 1 or RAID 5 so it follows traditional RAID boundaries. This is not the case with storage systems like the 3PAR InServ Storage Server which uses wide striping as it by default disperses Exchange data across most of the disk drives in its system. So while putting Exchange data on wide striping almost always results in better performance than even what following Microsoft's best practices will deliver, it may not technically meet the specifics cited Microsoft's best practices for implementing and configuring Exchange in these enterprise environments.

The good news is that 3PAR provides a mechanism to meet Microsoft's best practices for Exchange deployments and still automate storage management using wide striping without the need to re-introduce the overhead typically associated with managing and configuring storage volumes using its template feature. I'll get into how 3PAR's template feature works in an upcoming blog so that users can follow Microsoft best practices for Exchange storage configurations, still leverage 3PAR's wide striping feature and even potentially further improve Exchange's performance running on a 3PAR system.