Live From SNW - Day 1

Our semi-annual storage pilgrimage lands in Phoenix this year. Nice change of venue for Storage Networking World (SNW). As we did with SNW in the spring and VMworld last month, I will be doing updates throughout the three days. As always I ask your patience as we are doing this rapid fire.

Our next briefing was with the SNIA group driving the Cloud Storage Initiative. They are trying to round up the herd of cats and bring some organization to the term. No small task. This is important for SNIA because clearly cloud storage is driving new technologies and is bringing new companies to the storage fold that normally have not been part of the community. 

The big change of course in cloud storage is not so much the technology as it is the way that storage is delivered via an on-demand model. It is also a shift in administration. As one of the SNIA representatives put it, the disk drive has been commoditized but not the administration of that disk drive. The cloud seeks to commoditize that storage administration as well. 

Part of the work of this SNIA committee is to develop a common cloud API set that applications can interface with. I believe that APIs are going to be an important part of cloud adoption going forward and attempts at a standard deserve consideration. It will be interesting to see how the rest of the community responds to the effort.

Next up on Monday's briefings was Sepaton. Sepaton continues its focus on providing disk based data protection for large enterprises. Sepaton's target customers tend to start at 50TB and up. To address this market they have a clustered or grid based architecture that scales to meet the performance and capacity requirements of these enterprises. They recently announced delta based replication to go along with their deduplication technology. Our discussion focused more on what was coming next.

Sepaton plans on leveraging its content aware architecture to scale beyond basic data protection capabilities to be able to provide data management to additional non-primary storage data repositories. Since Sepaton understands the data that it is currently backing up, it is not a far leap for them to be able to do something with that data. That could be tasks like indexing, retention or even migration to different tiers for secondary storage.

These capabilities are critical when you consider the landscape in which Sepaton competes. Enterprise players are still trying to figure out their right ingredient for high end disk based backup and I think are discovering that it has to do more than just copy data from disk to disk. You have to do something with that data when you get it.

HP was next up on the briefing calendar. While they are not announcing anything major at the show, I had time to sit down with them and discuss their storage strategy which was valuable for me because I have been confused in the past. My interpretation is that HP Storage now is broken down into two broad groups: platforms and unified storage. The platform group is the classic traditional storage platform. 

The unified storage group is made up of products from a series of HP acquisitions like Lefthand Networks, Ibrix, and Polyserve as well as internal products like their disk to disk and VTL solutions. The key here is that the products in the unified group are typically built on an X86 infrastructure. They are delivering these services either on classic servers like the ProLiant, their blades, virtual machines or an integrated storage platform. 

HP is in an interesting position. With their c-Class BladeSystem as a starting point, HP could provide a rack where storage applications, virtualization solutions and applications are provisioned on an as-needed basis, almost like a cloud in a rack.

AutoVirt - First up was AutoVirt. They have created a software based global name space focused solely on the Windows platform. 85% of unstructured data is on the Windows platform, especially in AutoVirt's target market; mid-market data centers. The product provides storage managers the ability to quickly and easily create a global name space. With the name space in place, storage managers can move folders and files between the different Window shares transparently to the users. 

AutoVirt is announcing their 3.0 release at the show. In addition to migration and file movement, it adds the ability to replicate and provide file level archive. Replication brings a certain HA flavor to Windows based file servers. With the name space in place, if one server is not available the user is re-routed to a backup server. The archive function allows the movement of data to a specific archive tier and then sets that data to read only, providing a base level of compliance. 

The market that AutoVirt is focused, mid-tier data centers, is in pain when it comes to data management. They are at the point where they can't keep buying more storage and ignore the problem and there is too much of it to move to a different storage tier via manual brute force.