At Interop in Las Vegas this year, there was a virtual deluge of = vendors=20 touting their respective Software=20 Defined Networking-flavored products. I kept experiencing a vague = feeling of=20 d=C3=A9j=C3=A0 vu, of having seen this all before, but couldn't figure = out why. Then as I=20 sat in a product briefing I remembered why this all seemed so familiar: = the=20 wireless controller=20 market.
Wireless controllers are what I like to call proto-SDN. When vendors = released=20 the first lightweight access points with centralized management systems = and=20 controllers, it was akin to achieving networking nirvana. Centralized=20 management, control and visibility -- all in a handy-dandy Web = interface. What's=20 more, it supported IETF standards like lightweight access point protocol = (LWAPP)=20 and control and provisioning of wireless access points (CAPWAP).
I'm a bit skeptical regarding the = evolutionary=20 changes SDN promises the industry. We've seen it all before. The = industry had=20 its chance with wireless and blew it.
Michele Chubirka
Could life get any better than this? I raised my head and laughed = heartily at=20 those wired engineers with their painfully antiquated command line = interfaces. I=20 pitied them and their lack of power over their domain. I, Master of the = Radio,=20 could configure, troubleshoot, secure and manage my network without = (almost)=20 ever having to leave my desk, eating bonbons while those plebian wired = engineers=20 endured the drudgery of that backward hardware platform.
Paying for my overconfidence
I would pay for my hubris. Yes, I could manage and control everything = centrally, but at the cost of my independence. You see, while vendors = proclaimed=20 their support of open standards, I learned it was mostly fantasy. The = harsh=20 reality was that there was no real inte= roperability=20 between vendors' wireless hardware, so once you invested in one product, = switching to another meant a complete forklift or running multiple = systems in=20 parallel. Or let's say I wanted to use some access=20 points (AP) from another vendor for certain features or = capabilities: maybe=20 the APs were particularly good in high-density environments. Well, I was = out of=20 luck with fully managing that device unless I installed a parallel = ecosystem.=20 Oh, I always had SNMP,=20 but after tasting imported caviar, why would I go back to domestic?
And this is why I'm a bit skeptical regarding the evolutionary=20 changes SDN promises the industry. We've seen it all before. The = industry=20 had its chance with wireless and blew it. Will today's SDN promises of=20 transcendent centralized management and control bear fruit, or will they = just be=20 another precursor to vendor lock-in? After my wireless experience, I = feel like=20 I'm being offered leftovers. The question is whether they'll be as yummy = as that=20 turkey sandwich the day after Thanksgiving or as unappetizing as = two-day-old=20 pizza for breakfast.
About the author: Michele Chubirka, aka =E2=80=9CMrs. = Y,=E2=80=9D is a recovering=20 Unix engineer with a focus on network security. Likes long walks in = hubsites,=20 traveling to security conferences and spending extended hours in the Bat = Cave.=20 Believes that every problem can be solved with a "for" loop. She also = hosts a=20 podcast called Healthy Paranoia, a security feed of Packetpushers. You = can find=20 her blogs and podcasts at healthyparanoia.net or=20 http://packe= tpushers.net/author/securityprincess.=20 When not blogging or podcasting, can be found using up her 15 = minutes in=20 the Twittersphere or Google+ as @MrsYisWhy.
This was first published in June 2013