Five Networking Thoughts You Should Be Thinking
As Jordan Novet aptly styles it, “Yes, people really are still debating the definition of SDN” (software-defined networking). While “IT Ops” labeled SDN a “Trend to watch …” late last year, the industry as a whole still hasn’t figured out the relative roles of virtualized networking, commodity hardware, network function virtualization (NFV), software-defined datacenters (SDDC), programmable networking, and other promises. No one is in a position to sell “solutions” yet
Start action now
That’s the first and most important reality in the current networking landscape. Despite this–despite being so clearly in what Novet and others identify as “the FUD phase”–it’s time for you to start taking action on SDN. SDN will eventually be indispensable. Even more than that, parts of it have already crystallized. You can start to profit from the slices of SDN that make sense now, well before the whole domain has been adequately defined. At the very least, it’s a good time to analyze your own networking. Much of the promise of SDN has to do with fitting technology to business requirements; a natural starting point for you, therefore, is to inventory your existing traffic, assets, and artifacts. Do your biggest daily networking headaches have to do with internal CIFS bandwidth? Inter-campus security? Latency in delivery of HTTPS to external end-users? However SDN shakes out in commercial practice, you’ll be in a better position to leverage it with clear understanding of your own needs.
Distinguish commercial trends from long-term technical advances
While you’re doing this, keep in mind the third important reality about networking in 2013: coverage and commercial “buzz” have to do with monetization, which is distinct from technical value. Tom Nolle, president of CIMI Corporation, frequently touches on this in his occasional networking blog. In the long term, SDN must converge on what provides technical benefit; over the next year or so, though, most SDN maneuvering has to do with vendors staking out territory and rent-seeking. This divergence doesn’t mean you give up on networking purchases; just keep in mind that your motives won’t always be aligned with vendors’.
IPv6 on the way
Fourth: IPv6 is on the way. As frustrating as the rollout has been–and it’s sure not what I would have dictated, were I King of All Things Networked–IPv6 is also part of the future, and it’s time to deepen whatever experience you have with it. It really, really, is time to make yourself good at IPv6.
Faster and faster
Finally, it’s time to practice thinking about faster networks. An important part of good engineering during the last decades has been to recognize that faster hardware solves certain scaling problems quicker than software rewrites, simply because organizations are so slow and cumbersome at application maintenance. We’re well into an era where networking is like that: terabit ethernet, power-over-ethernet (POE), 802.11ac wireless networking, next-generation NFC and Bluetooth, advanced intra-rack networking, and other standards distinct from SDN are all making this a great time to re-run analyses of your plans for the next two-to-five years. Look for game-changers; they’re out there.