Every now and then, someone generates clickbait content recycling the “networking engineers are a dying breed” hype.

Here’s what I wrote as a response to a particular instance of that stupidity in 2013. Not surprisingly, networking engineers are doing well in 2022, marketing VPs are still generating nonsense, and the author of that article probably still keeps spreading it.

In related news (from July 2022 perspective):

  • NEC ProgrammableFlow disappeared
  • So-called SDN controllers are everywhere (including campus and wireless), but they are mostly just another layer of complexity on top of the networking infrastructure. We called those things “network management systems” before SDN was cool.
  • We’re still dealing with VLANs

Response: SDN’s Casualties

An individual focused more on sensationalism than content deemed it appropriate to publish an article declaring networking engineers an endangered species on an industry press web site that I considered somewhat reliable in the past.

The resulting flurry of expected blog posts included an interesting one from Steven Iveson in which he made a good point: it’s easy for the cream-of-the-crop not to be concerned, but what about others lower down the pile. As always, it makes sense to do a bit of reality check.

  • While everyone talks about SDN, the products are scarce, and it will take years before they’ll appear in a typical enterprise network. Apart from NEC’s Programmable Flow and overlay networks, most other SDN-washed things I’ve seen are still point products.
  • Overlay virtual networks seem to be the killer app of the moment. They are extremely useful and versatile ... if you’re not bound to VLANs by physical appliances. We’ll have to wait for at least another refresh cycle before we get rid of them.
  • Data center networking is hot and sexy, but it’s only a part of what networking is. I haven’t seen a commercial SDN app for enterprise WAN, campus or wireless (I’m positive I’m wrong – write a comment to correct me), because that’s not where the VCs are looking at the moment.

Also, consider that the my job will be lost to technology sentiments started approximately 200 years ago and yet the population has increased by almost an order of magnitude in the meantime, there are obviously way more jobs now (in absolute terms) than there were in those days, and nobody in his right mind wants to do the menial chores that the technology took over.

Obviously you should be worried if you’re a VLAN provisioning technician. However, with everyone writing about SDN you know what’s coming down the pipe, and you have a few years to adapt, expand the scope of your knowledge, and figure out where it makes sense to move (and don’t forget to focus on where you can add value, not what job openings you see today). If you don’t do any of the above, don’t blame SDN when the VLANs (finally) join the dinosaurs and you have nothing left to configure.

Finally, I’m positive there will be places using VLANs 20 years from now. After all, AS/400s and APPN are still kicking and people are still fixing COBOL apps (that IBM made sexier with XML and Enterprise JavaBeans integration).

Latest blog posts in What Is SDN? series

19 comments:

  1. New technologies will most likely contribute to job growth, not loss; as you've pointed out, the key is to adapt.

    VMware did not cause a downsize in the server team for most companies - rather there is a shift in focus - instead of spending time rolling out big irons, they optimise, monitor and find new ways to use the scalable virtualised environment now, with the same or increase in the number of staff.

    SDN, in time to come, is going to achieve the same thing - instead of folks rolling out physical switches, they will spend most of their time monitoring, optimising and finding new ways to use the network.

    p.s. In the wireless LAN, I think architectures using Centralised Wireless controllers with protocols like CAPWAP look very much like SDN - and they have been around for some time. I guess that's missed opportunity for the marketing folks to name them Wireless SDN (wSDN).
    Replies
    1. Actually, VMware did downsize the server team. As the consolidation ratios rise, server teams are managing more instances with fewer personnel. A few years ago, the admin to instance ratio was 1:10. Now it is closer to 1:100 (or some are even higher). It also drove the skillset up the ladder allowing virtual administrators to pick up roles in storage and networking.

      Networking will be pruned to a certain degree - partially form improved technology and partially from the "sharing" of duties with other teams.

      Good point on the wireless controllers. They have a working SDN implementation to which datacenter networking is headed.
    2. I don't normally post but I have to after your last line. I can just see a press release at CiscoLive now... Prime NCS is now Prime SDN.
  2. Get IPv6 IN and VM admins will piss in their pants.....
  3. @Anon

    There is actually a difference between an explicit downsize and a more favorable employee to device ratio. The total number of whatevers (VMs, networking devices, whatever) continues to climb in absolute terms. Despite several rounds of convergence, this macro trend remains true.

    Similarly, despite new technology trends, the absolute number of people to man these devices continues to rise.

    SDN should mean the two (devices and people) shouldn't have to grow at the same rate, but that is actually a different point than suggesting an industry-wide reduction in force. Such a workforce action is difficult to take for a lot of reasons, the least of which are technical.

    Rather than frame the discussion around who will lose their jobs (people will retire and self-select out over time anyway), I'd rather focus on who will make a killing because they are evolved. I do think a DevOps person with networking engineering knowledge will cash in. They should have first mover advantage in a changing workforce.

    -Mike (@mbushong)
    Plexxi
  4. Some more good news for network engineers.

    (1) Someone still needs to build the underlay network -- SDN doesn't mean there will no longer be boxes, ports and wires. It just means there will be less people thinking you're an idiot for taking so long to deploy a machine.

    (2) A good deal of time is actually spent by network operators in trouble shooting application problems. Most app folks don't [want to] know anything about what happens after their program calls an API for service. There will always be a need for specialists that understand what happens between the program on one end and the program on the other.

    (3) When your SDN ain't working, it might be a good insurance policy to have a specialist on hand.

    (4) SDN may actually mean fewer jobs for project managers involved in deployment management rather than fewer network specialists. Most shops hardly had enough network specialists in the first place.

    (5) I personally don't know any network engineers that only provisioned vlans so I think most network engineer jobs are quite safe. If all you do is program vlans, then you're probably really a server admin. ;)
    Replies
    1. Get this...even if a single phy port is requried..Cisco will be IN and SDN out....

      So they need to build networks without a single phy port....
  5. HP has an SDN solution for data center and campus. In fact, they rolled out SDN for campus before data center.
  6. OK, please help me understand the difference between marketing and reality. My definition of "rolled out product" is something that is documented (including data sheet/install/config docs) on vendor's public web site.

    Networking/Technology/SDN path on HP web site brought me to this page:

    http://h17007.www1.hp.com/us/en/networking/solutions/technology/sdn/index.aspx#tab=TAB4

    ... where the only "SDN" management product listed leads me to an IMC plugin, which can provision ToR switches based on VM information gleaned from VMware.
    Replies
    1. Hard to find the product/application detail on the web, but HP has the "Sentinel" SDN application that turns the network into a distributed botnet filter and denies at the campus access layer if trying to go to rogue/malicious websites. Switch redirects DNS requests (OF matching on port 53) to Tipping Point IPS, or there is some back end integration somehow, and validates DNS requests on ingress to network rather than at Internet Perimeter.

      Distributed policy enforcement at the edge...in the campus.

      Sounds like a solid app, but not sure I'd call it an "SDN Campus Solution."

      -Jason
    2. Please help me understand the following "fact sheet":

      http://h20195.www2.hp.com/v2/GetPDF.aspx/4AA4-0792ENW.pdf?jumpid=em_r1165_ww/en/large/eg/Solutions_for_HP_Virtual_Application_Networks/sdn/business/Apr_2013

      It says "OpenFlow switches available now", "VAN and VCN in limited beta" and doesn't even mention the availability of Sentinel. There's also nothing in the "products" section of their web site.

      It might be that I'm missing some nuances of English language, but in school they told me that you usually use present tense (has) for present, not future.
    3. Ahh, okay. I saw a customer speak about Sentinel twice. Same customer at ONS and Interop...guess maybe they are still on beta code. To me, "has" seemed appropriate...but apparently not...thanks for keeping the facts and grammar in check :)
    4. HP`s SDN solutions will be based upon exH3C´s switches with Comware OS version 7, which is currently only available on some of their latest switches, e.g. A5900, A10500, A11900, A12900, ...
    5. OMG ... and that coming from a company that has been touting its OpenFlow leadership for years. Priceless.
    6. Re Anon @10:00 .. HP currently has supported OpenFlow switches from its traditional ProCurve portfolio in 29 switches... At end of April 2013, it announced soon to be support for 40. These are the "exH3C's switches with Comware OSv7" ..

      These ProCurve switches are the ones that HP started providing commercial support on in March 2012, and we have been adding more products ever since. We have hundreds of customers running both full and hybrid OpenFlow deployments and lots of beta testers of our controller and SDN applications as well.

      Sentinel is not yet available for purchase.. If you want it, we can run beta trials with you. Speak to your local HPN team for details...

      Full disclosure > HP employee
  7. @Ivan

    Have you looked at Cyan's products?

    SP focus but still applicable to enterprise WAN/campus LAN.

    http://www.cyaninc.com/
    Replies
    1. Visited their web site. Full of marketing fluff, no hard facts I could digest. No documentation or data sheets online.

      I'm so not interested.
  8. There is some interesting research showing a decoupling between job growth and productivity:
    http://www.technologyreview.com/featuredstory/515926/how-technology-is-destroying-jobs/

    In the last decade, productivity continues to increase while job growth is not.

    This leads me to believe, that supplanting network engineer jobs (that spend the bulk of their time manually provisioning configurations that will be handled by SDN type solutions) with new SDN engineer jobs will not necessarily mean more jobs will be created. It's very probable that fewer network related jobs will be created, due to the efficiencies and increased productivity gains from future SDN solutions.
    Replies
    1. It does not say anything about network specific jobs..getting automated...it talks in Generall...
      BTW we know how to play with the Tide.Duh...

      You may get outdated and outlcassedd....by the time you read my respone..you may be out..begging...
Add comment
Sidebar