Values Exposed at The BDEvent: A New Era in Data Storage is Dawning
Anyone who has ever witnessed a disaster knows that one of two things can happen. Either the area affected by the disaster can be devastated, never to recover; or, new life can spring up in its place. In many respect, the economic disaster that hit the entire nation and world hit the data storage industry equally hard. However the data storage industry is picking itself back up and, based upon what I saw and heard this week at The BDEvent in Palo Alto, CA, it has brought an end to one era in data storage while the dawn of another is now upon us.
This past week I had the opportunity to attend the three (3) day BDEvent, a conference that specifically facilitates business-to-business networking for those individuals and organizations in the data storage and electronically stored information (ESI) industries. While it is a relatively small gathering (~150 people), it is one of those conferences where almost all of those in attendance are regularly engaged in off-the-record conversations.
That said, there were enough on-the-record comments and presentations made during the course of the event that I can reference them in this week's recap blog. In fact, the analogy that I used at the outset of this blog came directly from some comments made by Dave Hitz, NetApp Founder and EVP, during the event's closing presentation.
His analogy pretty well summarized what I had been thinking during the course of the event: a new era in data storage is dawning. Technologies that met customer needs and budgets as recently as a couple of years ago are now viewed as inadequate, inefficient and no longer cost effective. Further, some technologies that have been incubating for years and quietly developing extensive user bases are coming out of the woodwork because of their value proposition. There were three that I specifically wanted to highlight who were in attendance at this event.
First is Coraid. It is a storage array provider that I covered some years ago for Storage magazine in a Trends article but have heard nothing from this company since. However there was Coraid's CEO Kevin Brown at the conference who explained that Coraid had now grown to the point where it had approximately 1100 customers.
Not 10 or 50 customers, mind you, but 1100. That's remarkable considering that the knock on Coraid then and now is that it uses a non-standard Ethernet networking protocol - ATA over Ethernet (AoE) - which is not used by anyone else (at least to the best of my knowledge)
The upside of using this protocol (which is delivered via an OS driver supplied by Coraid) is that you get all of the benefits and speeds of a FC SAN while only needing to put in place an Ethernet network. Since AoE is a layer 2 Ethernet protocol, it does not introduce all of the overhead of TCP/IP so companies can use the Ethernet cards that are embedded in servers and take full advantage of 1 Gb and 10 Gb Ethernet networks.
Now, granted, the use of AoE as a non-standard network protocol could be a barrier to adoption but for the 1100 customers using Coraid, they do not seem to care. As long as Coraid continues to give its customers a good value for the money, meets their application needs and does not try to take advantage of them, their customers probably could not care less whether the protocol is a standard or not.
Ocarina Networks is another company that is adapting to new demands from its customers. Originally it started out doing post-process deduplication of large image files (JPGs, MPEGs, etc.) that had been dormant for 30 days or more - great stuff! But now its customers and even OEMs (Ocarina did not say who) are coming to it and asking for it to do end-to-end data deduplication from primary disk to backup disk without ever reconstructing it. After all, once the data it deduplicated on primary storage, why reconstruct it to then deduplicate it again when it is backed up?
But here's the rub - no one was running around at this conference asking "Where is the deduplication standard?" No one was crying foul because SNIA or IEEE or whoever had not yet finalized a deduplication standard. The reason no one has - aside from the fact everyone else is guilty of the same transgression - is that deduplication saves customers money and solves real world problem. So again it was clear that customers do not care nearly as much about standards as they do about saving money and working with providers that they trust.
The third provider at this event that caught my eye was Netex. All we have heard over the last few years is that Riverbed is the best way to do WAN optimization using one of its Steelhead appliances. That's all fine and good but after talking to Netex, it appears it has come up with a better mousetrap. Rather than deploying a physical appliance that consumes additional rack space and power, Netex's HyperIP does the same things but it is deployed as a virtual appliance on an existing ESX server. This could give Netex a decided edge in competitive situations at a fraction of the cost of the Riverbed appliance.
Yet regardless of whether organizations use Riverbed or Netex, both are again delivering value as they accomplish approximately the same thing and neither one uses an optimization standard that is recognized or is supported by the other. However the customers they are supporting probably do not care about the standards, they only care about the business value that either of these technologies deliver as well as the upfront cost of obtaining the software.
These were not the only three that supported the thesis of my title. I saw emerging technologies from providers like ProStor Systems, Virsto Software, and Zetta that all re-enforce my belief that the dawn of a new era in storage is upon us. However it is not just the financial value that these companies are providing in terms of dollars saved that are contributing to their current success.
While that is part of it, they do not appear to be taking advantage of their customers and backing them into corners with exorbitant price hikes once they have them hooked on these non-standard platforms. So as long as they continue to treat them fairly both ethically and financially, why would their customers switch?
This brings me back to Dave Hitz's comments. He sees a new era in data storage and information management dawning right now for the reasons cited above. Yes, old ways no longer work so well. Yes, better technologies are emerging. But providers are being forced to become more transparent than ever before because of Web 2.0 outlets like blogs, forums, LinkedIn, Facebook and Twitter where, if a Tweet or a blog post goes viral, their credibility can grow or be instantly shot. This combination of factors is leading to rapid change in the storage environment - maybe even more quickly than I anticipated - and those users and providers that fail to recognize that this confluence of forces is occurring may be the next ones victimized by it.
This past week I had the opportunity to attend the three (3) day BDEvent, a conference that specifically facilitates business-to-business networking for those individuals and organizations in the data storage and electronically stored information (ESI) industries. While it is a relatively small gathering (~150 people), it is one of those conferences where almost all of those in attendance are regularly engaged in off-the-record conversations.
That said, there were enough on-the-record comments and presentations made during the course of the event that I can reference them in this week's recap blog. In fact, the analogy that I used at the outset of this blog came directly from some comments made by Dave Hitz, NetApp Founder and EVP, during the event's closing presentation.
His analogy pretty well summarized what I had been thinking during the course of the event: a new era in data storage is dawning. Technologies that met customer needs and budgets as recently as a couple of years ago are now viewed as inadequate, inefficient and no longer cost effective. Further, some technologies that have been incubating for years and quietly developing extensive user bases are coming out of the woodwork because of their value proposition. There were three that I specifically wanted to highlight who were in attendance at this event.
First is Coraid. It is a storage array provider that I covered some years ago for Storage magazine in a Trends article but have heard nothing from this company since. However there was Coraid's CEO Kevin Brown at the conference who explained that Coraid had now grown to the point where it had approximately 1100 customers.
Not 10 or 50 customers, mind you, but 1100. That's remarkable considering that the knock on Coraid then and now is that it uses a non-standard Ethernet networking protocol - ATA over Ethernet (AoE) - which is not used by anyone else (at least to the best of my knowledge)
The upside of using this protocol (which is delivered via an OS driver supplied by Coraid) is that you get all of the benefits and speeds of a FC SAN while only needing to put in place an Ethernet network. Since AoE is a layer 2 Ethernet protocol, it does not introduce all of the overhead of TCP/IP so companies can use the Ethernet cards that are embedded in servers and take full advantage of 1 Gb and 10 Gb Ethernet networks.
Now, granted, the use of AoE as a non-standard network protocol could be a barrier to adoption but for the 1100 customers using Coraid, they do not seem to care. As long as Coraid continues to give its customers a good value for the money, meets their application needs and does not try to take advantage of them, their customers probably could not care less whether the protocol is a standard or not.
Ocarina Networks is another company that is adapting to new demands from its customers. Originally it started out doing post-process deduplication of large image files (JPGs, MPEGs, etc.) that had been dormant for 30 days or more - great stuff! But now its customers and even OEMs (Ocarina did not say who) are coming to it and asking for it to do end-to-end data deduplication from primary disk to backup disk without ever reconstructing it. After all, once the data it deduplicated on primary storage, why reconstruct it to then deduplicate it again when it is backed up?
But here's the rub - no one was running around at this conference asking "Where is the deduplication standard?" No one was crying foul because SNIA or IEEE or whoever had not yet finalized a deduplication standard. The reason no one has - aside from the fact everyone else is guilty of the same transgression - is that deduplication saves customers money and solves real world problem. So again it was clear that customers do not care nearly as much about standards as they do about saving money and working with providers that they trust.
The third provider at this event that caught my eye was Netex. All we have heard over the last few years is that Riverbed is the best way to do WAN optimization using one of its Steelhead appliances. That's all fine and good but after talking to Netex, it appears it has come up with a better mousetrap. Rather than deploying a physical appliance that consumes additional rack space and power, Netex's HyperIP does the same things but it is deployed as a virtual appliance on an existing ESX server. This could give Netex a decided edge in competitive situations at a fraction of the cost of the Riverbed appliance.
Yet regardless of whether organizations use Riverbed or Netex, both are again delivering value as they accomplish approximately the same thing and neither one uses an optimization standard that is recognized or is supported by the other. However the customers they are supporting probably do not care about the standards, they only care about the business value that either of these technologies deliver as well as the upfront cost of obtaining the software.
These were not the only three that supported the thesis of my title. I saw emerging technologies from providers like ProStor Systems, Virsto Software, and Zetta that all re-enforce my belief that the dawn of a new era in storage is upon us. However it is not just the financial value that these companies are providing in terms of dollars saved that are contributing to their current success.
While that is part of it, they do not appear to be taking advantage of their customers and backing them into corners with exorbitant price hikes once they have them hooked on these non-standard platforms. So as long as they continue to treat them fairly both ethically and financially, why would their customers switch?
This brings me back to Dave Hitz's comments. He sees a new era in data storage and information management dawning right now for the reasons cited above. Yes, old ways no longer work so well. Yes, better technologies are emerging. But providers are being forced to become more transparent than ever before because of Web 2.0 outlets like blogs, forums, LinkedIn, Facebook and Twitter where, if a Tweet or a blog post goes viral, their credibility can grow or be instantly shot. This combination of factors is leading to rapid change in the storage environment - maybe even more quickly than I anticipated - and those users and providers that fail to recognize that this confluence of forces is occurring may be the next ones victimized by it.
1 Comments
Leave a comment
Newsletter Signup
Good points about standards in the dedupe industry. Right now, the solutions in the market are point products, meaning native data goes in, native data comes out, and disparate vendors products interoperate just fine, even though data gets reinflated each step of the way.
However, as the industry shifts towards end-to-end dedupe, to deliver value beyond just storage cost savings, the story changes. In this vision, dedupe may be applied at any point in its lifecycle, from app servers, to storage, to archival, to DR targets, and data stays shrunken as long as a human eye isn't looking at it.
In that model, any product that 'rehydrates' the data unnecessarily will stand out like a sore thumb, and vendors with the complete solution in place will be in a position of leverage (good or bad!). So there *will* be pressure for standards here to reduce lock-in risk and promote interoperability. The standards won't take the form of standard chunking and shrinking algorithms though. It's only necessary to by able to learn the file-maps, find the chunks, and perhaps be handed the decoder for unique algorithms, through a standards-defined access method (aka API).
In fact the algorithm guys at Ocarina have already worked on self-describing file wrappers, so a naive SW decoder can rehydrate something treated with proprietary compression.
For the time-being though, you're right that there is so much to be gained by dedupe that customers are willing to look the other way when it comes to standards. But as this space moves from 'new frontier' to stable-stakes-feature, standards pressure may escalate.