Software-Defined Networking (SDN): Does it Make Sense for MSPs?
There's been a new term being bandied about in various IT and data center circles lately: Software defined networking. Indeed, our editorial director recently queried some pretty smart MSPs about their plans around SDN and received blank looks (my reaction, too) and then assurances that it was all hype. But it made me wonder, just what is this software-defined networking (SDN) anyway? So I set off to find out. Here's what I learned.
First, of course, I checked out what Joe Panettieri wrote here about SDN: "SDN decouples control from hardware. The result could be easier-to-manage networks built on lower-cost hardware." Much technology has gone this way before. Could networking do it, too? Joe points out some vendor moves in SDN recently. It does seem like momentum is building around this.
ISO Model, Bob Metcalfe and Standards Wars
Then I got further down in the weeds and started thinking about the classic seven-layer ISO model and how SDN related to that. My wanderings eventually led me to an article about how Ethernet inventor Bob Metcalfe is acting as kind of an elder statesman diplomat to ward off a standards war between two factions around SDN.
Here's how Metcalfe described it to EETimes: [SDN] is being driven by the Open Networking Foundation (ONF), a group of “mostly computer companies like Google,” Metcalfe said. On the other side is the Metro Ethernet Foundation (MEF) which has investments in the carrier-class Ethernet systems that run most of today's network traffic and includes "carriers and companies like Cisco," Metcalfe said. If you go to the sites of both of these organizations, you'll notice that Cisco and a few other big players are members of both, hedging bets perhaps.
Intel's SDN Bet
Last week Intel made some moves around SDN at the Open Networking Summit unveiling reference designs that manufacturers could use to make SDN solutions. The IDG News Service provided this background on SDN and further insights around Intel's move and how it will affect service providers:
"Among other things, the Open Networking Platform will help to make virtualization feasible in service-provider networks, where requirements for low latency and voice jitter have prevented it until now, [Rose Schooler, vice president of the company's Intel Architecture Group] said. SDN promises to help service providers lower their capital, energy and operating costs, which are rising faster than their revenues, plus help them introduce new services more quickly, she said. Specialized, proprietary equipment with separate software stacks on each network device has hampered centralized management of enterprise and service-provider networks, according to Intel."
This sounds like mostly a play for MSPs working for big companies and for those building their own data centers. That said, if we are still in reference designs and standards wars, how far away is this technology really? My guess, which could be totally wrong, is that is a ways down the road. If some organizations are still resisting an upgrade from Windows XP, are they really going to be anxious to rip and replace expensive networking equipment that is designed to have a much longer lifespan than a PC?
And so I'll ask you the question that Joe asked several weeks back: Do you have any plans for SDN?
Like the idea, but no one
Like the idea, but no one ever talks about the hurdles bottlenecks.
How do you make a network Manger “decouple” from his “toys” and become a SOFTWARE GURU?? Not too easy to take candy from a baby, is it ?
Answer: In most cases right
Answer: In most cases right now the network manager won’t become a software guru. My personal opinion — SDN is a vendor hype wave. Unlike cloud — which is real — SDN is a vision statement that most network managers have yet to embrace.
-jp
[full disclosure: I work at a
[full disclosure: I work at a vendor in the SDN space so coming from a vendor perspective here]
SDN is more than just separating control and forwarding plane. SDN will ultimately enable a new class of applications that can work in concert with (not just alongside) the network. This new wave of capability will make automation easier (great for opex) and make new services possible (good for the top line). The end result? SDN will be big.
How big? It depends. IDC reported a $2B market size by 2018. We looked at it along with SDNCentral and came out with a much larger number. It is possible that SDN could make up $35B of the total $90B networking TAM by 2018. (http://www.plexxi.com/2013/04/sdn-market-to-reach-35b-by-2018/)
Don’t mistake that number with saying that deployments are close. But decisions to buy (based on RFP type responses) are increasingly including SDN as a meaningful selection criterion. By that measure, the financial impact could be huge.
-Mike Bushong
@mbushong
Mike: Thanks for the
Mike: Thanks for the disclosure and the comment. Let’s ignore the market research and forecasts for now and zero in on what you’re seeing in the market.
1. What types of customers are deploying SDN right now?
2. If you were going to give advice to MSPs and channel partners, what would you tell them to do amid the potential SDN opportunity.
I’m skeptical about all the hype. But I do value your readership and views.
-jp
Me; I work for an SDN that is
Me; I work for an SDN that is already bringing products to market – so yes, I do indeed have plans for SDN. 😉
– Mark
Mark/SDN_Student:
Thanks for
Mark/SDN_Student:
Thanks for stopping by and posting a comment. Tell us more.
1. What type of background do you have? MSP? Corporate IT? Consultant?
2. What are your SDN plans?
-jp
CISCO is moving fw with SDN
CISCO is moving fw with SDN as well. Will become a defacto standard in less than two years.