In-Depth

My Bad: More on SNIA's Silence

Does it take money and membership in SNIA to get that organization to consider a good product?

• By Jon William Toigo
• 08/05/2004

A couple of columns back, I wrote about hyperI/O, a vendor offering technology to track the number of accesses made to a file or dataset and the performance of the storage subsystem in responding to I/O requests. I saw the technology both as a potential enabler of true Information Lifecycle Management and as a provider of performance metrics that storage managers desperately need both to evaluate competitive products and to provide a basis for ROI analyses.

In summarizing comments from Tom West, CEO of the company, I inadvertently placed what might have been too much emphasis on his statements about SNIA’s disinterest in his technology. Specifically, I wrote, “West, whose burgeoning company develops products in the I/O measurement space, noted that he sought to contribute his technology to SNIA’s Storage Management Initiative Specification (SMI-S), an effort with which he is in philosophical agreement. But, he notes, his efforts have met with utter silence from that body, silence that he has interpreted as disinterest.”

I should point out that West sent me an e-mail—in advance of publication of the column, but unfortunately after deadline—in which he expressed his concerns that his comments regarding SNIA might be misconstrued. He said, “I have approached several folks associated with the SNIA in the hopes of introducing them to the hIOmon technology. However, hyperI/O is not an SNIA member and I have not formally approached any of the SNIA technical groups as regards contribution of the hIOmon technology to their SMI-S.”

“Basically,” West wrote, “I sought to garner SNIA interest in the hIOmon technology by introducing hIOmon to several SNIA members; my hope was that they would find it relevant to the SNIA efforts (such as the SMI-S with which I am in philosophical agreement as you mentioned) and perhaps solicit further interest in the hIOmon technology amongst the SNIA at large. Unfortunately, as noted, my efforts to gain the attention of these SNIA members were met with utter silence. As regards the official SNIA body itself, I suspect that they have not heard about hyperI/O and the hIOmon technology."

Maybe they hadn’t heard about it then, but they certainly have now. Alan Yoder, a Network Appliance guy by day and a SNIA Working Group Co-Chair by night, responded to the column with the following observation, “I do not see HyperI/O on the SNIA membership list. Given this, and today's litigious and predatory IP climate, claims of attempts to get involved there or ‘contribute’ technology are somewhat disingenuous,don't you think?”

Since Yoder asked, I must respond. In our view, the contribution of technology to a so-called standards-making effort is neither unprecedented nor disingenuous. Truth be told, it is done all the time in real standards groups like IETF or ANSI. Want examples? I can think of three off the top of my head.

Fibre Channel is one. IBM developed it, then presented it as a gift to ANSI. I don’t see any of the hundred or so vendors marketing FC-based products paying royalties to IBM for its intellectual property.

Another example is Cisco’s VSAN. This technology, currently proprietary and supported mainly on Cisco’s own equipment, was offered by ANSI member Cisco Systems to the T-11 Committee, where it was summarily rejected. It is now in the process of being validated as a standard through the “back door” at IETF, where MIBs are being written to support it: pretty neat trick for a non-standard. Brocade is said to be working on its own variation and will probably submit this, as it has its routing protocols, to T-11 for its standards track.

A third example is iSCSI, which EMC claims to own outright. As reported in this column in the past, IP Storage Working Group Chair David Black once stated on the IETF reflector that he felt compelled to let everyone know that EMC had a general patent on all network-based storage, including IP-based protocols (and two-cans-and-a-piece-of-string, for that matter). He was flamed so universally by the group that the observation went nowhere. As of this writing, EMC has made no moves to stifle iSCSI by bringing a suit against iSCSI vendors on the basis of patent infringement. Maybe this will happen in the future, of course, if EMC feels that other competitors jeopardize its iSCSI market position. Who knows?

So, the contribution of IP is not unprecedented. That leaves only the question of whether you have to pay to play. As West observes, his company is not a SNIA member. Nor, arguably, should he be. Most software guys I talk to feel that SNIA is not the right vehicle for their development efforts. It is dominated by hardware guys, regardless of how IBM, EMC, and others may be endeavoring to redefine themselves in the market. (Remember, by calling yourself a software company, you can sell stock totaling 30 or 40 times actual earnings without a Wall Street guy blinking an eye. Hardware companies get into hot water if their stock trades three or four times earnings.)

West’s reasons for not joining SNIA are his own. I do not know them. But, why should the clubby nature of SNIA be permitted to exclude a valuable technical adjunct to a SNIA effort. SMI-S, contrary to SNIA propaganda, is not a SNIA creature. It is based on the Desktop Management Task Force CIM/WBEM standard, and many DMTF folks are not SNIA members. It is also based on BlueFin, which was gifted to SNIA (more intellectual property, by the way) by a group of vendors who were compelled to create BlueFin by their leading customers. From where I am sitting, the organization is pretty accustomed to leveraging external non-SNIA stuff whenever it suits their fancy.

As a bureaucracy whose only purpose in the world is to grow, I have watched SNIA behave like the Blob in the old 1950’s Steve McQueen movie for some time. They have taken over other industry standards initiatives, starved them of member vendor support, and swallowed up their carcasses – effectively putting the efforts, however valid, to permanent sleep. They have moved into foreign geos, where vendor-agnostic storage conferences were held, starved them of their vendor sponsor dollars to get them to allow SNIA to take a decisive interest, then re-branded the shows as SNIA events with all the marketecture that goes along with the moniker. Now they seek to do the same thing with training, by creating a “consortium” that has a lot of little guys worried about their bread-and-butter going forward.

The bottom line is that West correctly reported his product capabilities, which makes a lot of sense to us. He also reported the non-responsiveness of SNIA members to his efforts to contribute. It may well be the case that he has great technology, but lacks the coin or the business case for joining the SNIA party. Does that automatically make him a non-player, or, worse yet, a nefarious vendor setting a trap for an IP infringement lawsuit downstream? I kind of doubt it.

The good news is that Howard Goldstein, a friend of this column, CEO of storage training company HGAI, and part of the SNIA Training Consortium, has seen the potential value of a product like hIOmon. He contacted us to connect him up with West in the hopes of adding value to his training programs through the use of the technology.

One small step for sanity in the SNIA universe. Comments? jtoigo@intnet.net