We attended this year's spring Interop show and found it sleepy--and quite frankly a bit dull. Attendance was down, which is understandable in this economy. However, the big disappointment with Interop this year and in recent years is that it has lost its mojo. But we have a proposal to wake Interop from its doldrums next year--make interoperability testing for application delivery system (i.e. WAN optimization) vendors a centerpiece of the show.
A bit of history is needed here. Interop was the brain child of Dan Lynch, who envisioned it as a place for vendors to test the interoperability of their TCP/IP implementations. The Defense Data Network (DDN) had just become operational and there were thousands of hosts that needed to connect to the DDN with the new unknown protocol called TCP/IP. We know whereof we speak when it comes to Interop lore, since Peter has attended Interop since the very first meeting in 1985, and Rebecca since the early 90's.
All of the host vendors were still forcing their customers to use their own proprietary protocols like DECNET, IBM SNA, IBM 32xx, Burroughs Poll/Select, Banyan VINES, Xerox XNS, Novell NetWare, ARCNET, etc. The DoD had a huge problem trying to figure out how to implement or buy TCP/IP software for the existing hosts. A few companies had just emerged to supply the software. Interop was the venue for people to learn about TCP/IP--and more importantly see intercommunication among the vendors operating on the show net. That, for you youngsters out there, is why Interop has a show net.
An interesting thing happened as time went on. Sun Microsystems showed up at Interop in one of the early years with TCP/IP built in. This propelled Sun from a small player to a big success in DoD computing. It also caught the attention of the big computer vendors who suddenly started putting TCP/IP into their operating systems and stop stonewalling the new standard. Sun still had the lead when the Web came along. But by then the "slow-to-get-it" computer vendors were gone or played no role in the Internet wave.
Today there is little, if any, interoperability testing or demonstration at Interop. It is time to bring it back. We challenge TechWeb to return Interop to its roots and host ADS vendor interoperability testing by next year's show. We describe the need for this testing in yesterday's blog.
What a great way to celebrate Interop's 25th anniversary--with a revitalized agenda that moves the industry forward! Without this we fear that next year's Interop will be even more dull and headed towards extinction. How sad that would be.
The Leader in Network Knowledge
Rejuvenation
Well, the article is right on about the chaos that existed 3 decades ago in networking and my belief that TCP/IP was a neutral solution for the customers out there. It was fun testing all that stuff and the engineers needed to have fun. If today's application tangle needs a similar push, why not use Interop as a venue? i'd be happy to help.
Cheers,
Dan
Today's Interop
Peter, we're proud of our roots. But you're missing the point of Interop today: The network has grown to be the platform for all of IT and so has our mission. Interop is the largest, independent, business technology event with depth in many key technologies beyond networking. In fact, there was tremendous interest and energy at the Las Vegas event in a number of areas, including cloud computing, virtualization, data center, unified communications and mobility.
We're always interested in fostering interoperability. I'm happy to talk about how we do this around WAN optimization. From the comments, though, on your earlier post it's not clear that WAN optimization standards are seen as a priority. But let's discuss and see how we can work together. Cheers.
It Was There If You Looked Closely...
Peter has the right idea: With all due respect, Interop lacks the motive force that made it relevant in the early days. It was as much lab as trade show. Real things were discovered and IP technology moved forward as a direct result of the innovation and experimentation that characterized the original InteropNet. And while virtualization and cloud computing threaten to break the Internet as we know it (see last week's panel with Cisco, F5, VMware and Infoblox at the FiRe Conference - http://www.infra20.com/post.cfm/fire-infrastructure-2-0-panel-now-viewable-online) the network seems to be taking a back seat to (and may ultimately impede) the computing side of those initiatives.
There was, however, a Phoenix rising from the ashes in LV in May: At the Trusted Computing Group (TCG) booth, around a dozen vendors - including Juniper, Enterasys, McAfee, Wave, Hirsch, Lumeta, nSolutions and Infoblox - demonstrated disparate systems coordinating their actitivities using a new protocol, developed under the auspices of the Trusted Network Connect sub-group of TCG, called IF-MAP. The demos showed real, multi-vendor NAC systems; showed how physical access control (card-swipe) systems could be used to lock PCs off the network if their users had left a secured area; and showed how factory automation data could securely share an IP network with general IT traffic. By all reports, visitors to the IF-MAP demo at the TCG booth were wowed. And their excitement was reminiscent of the early days of Interop, when it was really about making things work together.
Interop played a meaningful role in the development and adoption of TCP/IP, which standardized and commoditized connectivity, with enormous impact. Who would have guessed that Cisco would replace GM on the DOW 30!!!??? Today, IF-MAP is poised to do for collaboration what TCP/IP did for connectivity. The potential impacts are no less profound and Interop can rightfully claim its role in having hosted the first MAP interoperability labs. What remains to be seen is if Interop will continue to be the premier venue for this next great wave in collaborative networked systems.
Interesting concept -- won't happen, of course
The concept of WAN acceleration product interoperability is interesting. The military has actually mandated this without really understanding what they mandated. So they specified that products must run SCPS as the transport layer protocol acceleration technology. If you were going to plan an interoperability test in the short term, SCPS would be a good place to start. It's a decent concept - an extended version of TCP that does things a bit differently than standard TCP if you allow it, and if it detects a SCPS partner on the otherside. It's not the greatest solution, but it is something of a standard supported by multiple vendors. And it really could use interoperability testing, not for demonstration, but to ensure that different implementations actually work together correctly and clarify the meaning of vague parts of the specification that have been interpreted differently by different vendors.
Beyond SCPS, there's not much that can possibly interoperate, so there's little to test. If we want interoperable solutions, we need to get together as an industry and say: here are the features that we want, here's the method that we're going to use to implement those features, and here's the specification for how the two sides will communicate needed information. This is really the IETF's job, but they've been conceptually opposed to the idea of any sort of middle boxes, and have refused to get involved.
At this point there's too much money involved for most vendors, at least the big ones, to willingly support an interoperable solution. It would take a big customer like the DoD with a mandate: lead the IETF working group, set strict deadlines, and then mandate that any product that doesn't meet the spec. can't be purchased.
Interop would be an interesting forum for the testing, but even if there was a mandate today, we wouldn't be ready with a spec and products to test by next year.
But back to the concept of bringing Interoperability testing back to Interop -- there has been ongoing interesting interoperability testing run by UNH. It's easier to get real work done on a college campus than in the middle of a big trade show in Vegas. But maybe we could recreate the original Interop around a week of testing at UNH? Would people come without the dancing girls and free candy?
Rejuvenate or resusitate?
It's a show. Long on entertainment based on the notion that vendor research is going on. I wonder how many Interop findings reports have ever been submitted to management? Once the granddaddy of shows, Interop may have simply passed into the retirement community phase. Relevance in the real world doesn't equate to vendor attendance and advertising. These are only a concern to the show management. Relevance is what new ideas are introduced. Today, a Twitter Post can achieve as much impact as a $1,000,000 Interop budget. Perhaps a name change: TwitterOp?
Post new comment