Answer: Sure. There are a bunch of pocket successes so far out there, and a lot more will start happening soon. For example, StoreAge sells their software (think of it as a superclustered volume manager distributed across heterogeneous machines) so that it executes not only on servers but also on the Troika (now owned by Q-Logic) platform. I understand it works well.
Falconstor has all sorts of services that can execute on an "appliance" in the network, from migration to replication to continuous data protection.
Brocade has its Data Mobility software, which runs on their Rhapsody-based switch platform in the fabric that enables heterogeneous migrations. EMC's InVista [software on a switch] is also designed for migrations, at least initially.
You can say that guys like Acopia, Neopath and Rainfinity (now EMC) are network-based intelligence platforms designed to provide file-based migration capabilities (and global name-space, etc.).
IBM's SAN Volume Controller sits in the Fibre Channel fabric to provide block virtualization and some services. I could go on all day.
Perhaps the most shining success of network-based intelligence is the IP network itself. Talk about virtualization! It wasn't that long ago that you had to manually establish connections of each device on the net, but who does that anymore? Plug something in, it just works. You don't spend all day long understanding interdependencies, charting addresses in Excel and doing extensive planning before you add a node to your IP network -- but you sure do in storage. The storage world hasn't learned yet that the value is not in the interface but at a much higher level. That's why things still don't plug and play in storage.
My point in that previous article was that the pure play network guys haven't loaded intelligence directly on the switching infrastructure yet. They are all working on it, but they don't have it yet.
Brocade, McData, and Q-Logic are all shipping an off-switch intelligence platform, and Cisco is still committed to joining the fray. It's an interesting bit of fence-straddling they all have to do, however, since the original storage manufacturer has been responsible for selling all the storage switches, and a lot of them feel that by the switch guys loading their boxes with "smarts," it can take away from the value the storage guys bring. It's an interesting dance to say the least.
For example, look what happened when Brocade bought NuView (a software play using Distributed File System, or DFS, for global namespace functionality). EMC got very, very irritated, and immediately upped the ante with McData, who it was very, very irritated with last year when they bought CNT. The fact that Brocade probably has no intention of selling the NuView software as is but will likely incorporate the DFS engine into their switch was lost on the folks in Hopkinton. And, since EMC buys a couple hundred million dollars' worth of Brocade gear each year, they have the chance to make their feelings heard -- loudly.
At the end of the day, all the switch guys will need to be able to add significant value in terms of "intelligence" or they will get so marginalized that only one will survive (can you guess who?). It also happens to be better for society as a whole.
Ultimately the right thing always happens; it just takes time. It will be painful for some and a death toll for others, but you will see smarter networks -- in storage and other ways -- soon. In the age of impending virtualization everywhere, the glue to all is the network. Making it smarter isn't just good, it's necessary.
Steve Duplessie founded Enterprise Strategy Group Inc. in 1999 and has become one of the most recognized voices in the IT world. He is a regularly featured speaker at shows such as Storage Networking World and others, where he takes on what's good, bad - and more important - what's next. For more information about Duplessie or ESG, go to www.enterprisestrategygroup.com.
Find your next job with techworld jobs