In-Depth

Uber-SAN Switch? Don't You Believe It

Can a new switch from Brocade ready unite different SANs based on non-interoperable switches from different vendors?

• By Jon William Toigo
• 12/09/2004

Far be it from me to criticize fellow industry observers—we all need to make a living in these difficult times. A recent column, however, by a technology trade press pundit got my goat. Well known for his Pollyannish views of really bad storage products from brand name vendors/advertisers, the writer in question gave yet another glowing review, this time to Brocade for its recently announced Silkworm Multiprotocol Router.

He referred to the product as a “Superswitch” and touted its value in uniting SANs heretofore separated by, more than anything else, vendor greed. The efforts of vendors (notably Brocade) to become the leader of the Fibre Channel world have fostered the implementation of Fibre Channel standards on switch platforms in ways that guarantee non-interoperability between the products of competing vendors. Now, Brocade is wrapping itself in the rhetoric of inclusion and seeking to provide a path to SAN unification that has all the credibility of, say, a Ukrainian election.

The columnist did not convince me. After what I seriously doubt was very extensive testing, he rated the technology a whopping 8.6 out of 10. However, he made no mention of comparable technology products from competitors such as Cisco Systems, which have been in the market for a lot longer than the Brocade offering. Nor did he mention bridging solutions from companies such as Crossroads or ATTO, or virtualization software solutions such as DataCore’s SANSymphony, which cost a lot less and do not require joining oneself at the hip with a particular vendor.

According to the story, Brocade’s “uberSwitch” promises to unite disparate SANs with greater ease than was possible before. Good news for those who, as a result of bad karma or poorly conceived mergers, confronted multiple SANs that had about as much chance of working and playing together as two toddlers with one Tickle Me Elmo. LSAN, promises Brocade in its best Mel-Gibson-as-William-Wallace brogue, will heal the wounds and unite the SANs.

So will VSANs from Cisco, or SAN Volume Controllers (SVCs) from IBM, or Storage Virtualization Routers from EMC, or even TagmaStore from HDS—if you want a different kind of vendor lock-in. Brocade’s Router is just more of the same marketing spin that has helped to create the litany of problems confronting just about everyone trying to ride herd on Fibre Channel fabrics today. Braveheart they ain’t.

The nonsense around the multiprotocol router is more of the same. At the same time, Brocade has endeavored to heal the shortcomings of FC SANs with in-band management technology embedded in its switches, what the company claims is delivering “more intelligent SAN management capabilities.” In-the-wire management has been a deficit of FC fabrics since they were first (and incorrectly) termed “SANs.” You had to run a third connection to every device in the fabric—usually an IP network—to actually manage the devices interconnected by dual redundant FC links.

This out-of-band management kludge was a product of two things: first, the design of the FC protocol itself, which made no provision for in-the-wire management or any other “IP stack-like functionality” in Fibre Channel; and, second, the refusal of the Fibre Channel Industry Association (now a part of SNIA) to acknowledge that a “management deficit” even existed in its over-hyped interconnect. Brocade is now claiming to have fixed this previously unacknowledged issue, albeit with a proprietary in-band management scheme that only works if you buy only Brocade switch products.

There is no mention of whether this management scheme works across SANs united by the Brocade Multiprotocol Router. The favorable review by the Pollyanna pundit didn’t see fit to address this. My guess is that customers who deploy the device will have to find out for themselves the hard way.

One Christmas gift that has come early this year is DataCore Software’s new SANSymphony release. Touting “virtual capacity,” the product gathers together all of the capacity identified to SANSymphony’s virtualization service—including “allocated, but unused” capacity—and parses it out to the applications that need more space in accordance with user policies. SANSymphony is trying very hard to mask the inefficiencies of underlying storage topologies and to get to the real value: managed capacity provisioned on an as-needed basis to hungry apps.

From where we are sitting, that’s something to crow about. That’s real uberSAN technology. Your views are welcome: jtoigo@intnet.net