In Company News, Featured, Industry Insights, Network Evolution, SDN

I’ve opined in the past about the impacts of SDN on the networking market at large, covering topics like commoditization and pricing mix and suggesting that there would be increasing pressure to size the SDN market as companies jockey for position and investors and analysts look for a scoreboard they can use to assess progress.

As the old adage goes, if you want something done right, do it yourself. So Plexxi has joined up with the premier SDN online community SDNCentral and one of the top venture capital teams Lightspeed Ventures to size the SDN market.

The big punch line is that SDN is going to have a much larger impact on network spend and in much shorter time than other industry estimates have shown. We expect that between 30 and 40 percent of total networking spend could be influenced by SDN over the next six years, which means the total SDN market could reach $35B by 2018.

We have published the full results as an infographic, and a report that outlines the measurements and assumptions can be found here.

As we were going through the data gathering and analysis, there were a few major takeaways for me:

  • The SDN market is growing much larger and much faster than anyone had anticipated. I think the numbers most cited by folks show the SDN market at something like $3.5B in 2015. We concluded that the market will grow to nearly 3x that size by 2015 and to a full 10x that size by 2018.
  • The number of SDN companies is a lot higher than I had thought. SDNCentral has some 225 companies registered as SDN companies. Even though some of these are certainly SDN-washing, it still means that we have seen an industry with near 0 SDN companies in 2009 climb to over 200 in just 3 years.
  • The amount of VC money being poured into this space is massive. Would you believe that the total VC investment has increased 50x over 3 years? Sure, some of this might be dumb money, but it means that a lot of people are betting a lot of money on the growth of this space.

This last point is particularly interesting to me. My personal conclusion here is that the competitive landscape is shifting in the kind of way that we only see maybe once in a generation. What has been a virtual monopoly for a couple of decades is quickly becoming a hotly contested space. The new technology has brought with it an influx of new money and new people, which will ultimately deliver new solutions based on new ideas. We haven’t seen this competition perhaps ever, and it should serve both customers and investors well. The new companies will loosen the stranglehold that years of a dominant incumbent have brought to networking. That will force everyone – including Cisco – to return to innovation and value as a means of competing. This can only spell goodness for consumers.

And the fact that SDN is acting as a conduit for adjacent spaces to enter the networking foray could make things really interesting. Obviously EMC and VMWare have been on a collision course with Cisco since they acquired Nicira, but Oracle has been active in the SDN space of late. And companies like IBM might leverage their systems integration skills to also enter the space. We could see a colossal clash of the titans before much longer, which would alter even more this rapidly changing space.

And then juxtapose this upon changes that are already happening within the IT space more broadly. The convergence of IT silos, the rise of software-defined everything, and the emergence of a DevOps population could add different twists to this story. For the first time that I remember in my dozen years in the industry, we will see real strategic work at play. No more chasing the next MPLS or multicast feature; we should see real product moves shaping the industry.

Oddly enough, this will actually make it harder for customers in the short run. Until the winners and losers are sorted out, customers will need to decipher the various SDN messages to determine what will ultimately emerge. But I feel confident in saying that whoever the winner, whatever the solutions, we will wind up in a dramatically different place than we started. And I for one cannot wait.

Showing 9 comments
  • Kristin Niemi

    Great visualization of the industry. Always an interesting perspective to see it in graphics rather than in just graphs.

  • Sanjay Kapoor

    Great work. Very thorough and nicely done.Congratulations!!!

    I would agree that for L4-7 networks services, NFV market sizing can be estimated as a % of the total L4-7 Network Services TAM, since Virtual L4-7 Network Functions are replacements for the otherwise spend on Physical infrastructure.

    But, when it comes to Network Virtualization spend, in Public/Private Cloud DC’s, not sure if this is quite that straightforward. Network Virtualization, specially of the Overlay type, is additive spend, to the underlying switch fabric and infrastructure. Not really a replacement.

    The benefit of such an overlay, is “Agility”, in provisioning, a VPC, within minutes vs Days. An alternate way to quantify the sizing of Network Virtualization spend, would be to take a % of spend on OA&M of the current physical network infrastructure.

    It is common knowledge that SP’s spend 4x-5x of CAPEX on OPEX, no wonder why they don’t make much money. If Network Virtualization, can help eliminate such spend in the Cloud DC, would they pay some % of that 5X for Network Virtualization Software. I WOULD !!!

    • mike.bushong

      Sanjay, sorry for the slow response. Lots of reasons, no excuses.

      I agree that overlays are additional spend. In my mind, the question is this: does SDN create new networking spend? Or is networking spend more tightly correlated with IT spend broadly (which will track with GDP)?

      I tend to believe that money does not get created; it gets moved. So as overlays emerge, that money either comes from broader networking (in which case it is part of the networking TAM, but a separate carve-out), or it comes from somewhere else in IT. It is true that some of this will come as opex is reduced, essentially converting spend from heads to products that are more efficient.

      The interesting thing here is that to get the opex, you have to deploy the technology. That deployment takes effort. So in the short term, opex actually might go up, and then dip back down on a different trajectory. This means that selling it internally is going to be paramount for companies who see this as strategic. For companies that already treat IT as strategic (Google, Facebook, etc), the transition should be easier. For companies who view it as a cost center, the transition is going to be hard.

  • claudiu.cismaru

    SDN will became a defacto standard in ~two years; will be interesting to see the competitive landscape by then. A mix of networking giants (i.e. CISCO), software large caps like Oracle and startups, each one with an eye on connecting their adjacent products. Some of the drivers that one may anticipate having a mid term impact: existing hardware’s 3-5 years life cycle, plus tons of easy/printed money in the bank, plus some early adoption at Google for their internal network and the numbers anticipated by the study above may become reality. Many thanks for sharing BTW.

    • mike.bushong

      I agree with you for most of the large enterprises and above. I think it might take a little longer for SMBs who handle most of their procurement through channel partners optimized to shorten call times to maximize revenue. And I completely agree with you about Oracle (and would add IBM). I believe the battle to watch will be the professional services battle. The DevOps movement could eventually look a lot like massive ERP projects where initial plans need to be managed by consultants. This will favor a bit those companies who have strong business practices around professional services. Should be interesting!

  • Dipesh shrestha

    Mike, very thoughtful article. I liked your thoughts around how competitive advantages ps organization could bring to the table. It’s been year since you wrote this article. What is your current thoughts around adoption of sdn and competitive landscape of sdn?

    • mike.bushong


      This is a great topic to revisit. Rather than write out a lengthy response here, I’ll add this to next week’s blog rotation–probably Monday’s.


pingbacks / trackbacks

Leave a Comment

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Start typing and press Enter to search