- Analysts are not employed by SDxCentral.
- Views and opinions expressed in analyst content belong solely to the author and do not reflect the views of SDxCentral, LLC.
- SDxCentral does not fact check analyst content. If you believe there is a factual error in analyst content, please notify firstname.lastname@example.org. Should we find factual irregularities, that article will be unpublished from the SDxCentral website.
Effective April 18th, 2019, the SDxCentral analyst blog syndication program has been terminated.
SDxCentral Statement about AvidThink, LLC
- Roy Chua, the founder of AvidThink, was a co-founder of SDxCentral. As of September 30, 2018, Roy is no longer affiliated with SDxCentral.
- The views expressed by AvidThink and Roy Chua are independent of SDxCentral and do not represent the views or journalistic principles of SDxCentral.
- As of April 18th, 2019, SDxCentral is no longer publishing AvidThink analyst blogs on the SDxCentral website.
Expectations were high as network operators and experts from enterprises, service providers, and vendors gathered in San Jose’s McEnery Convention Center for the Linux Foundation’s ninth annual Open Networking Summit (ONS). ONAP, the edge, and network automation were slated to be the hot topics, as well as advances in SDN and NFV.
It’s interesting to take a walk down memory lane and look at the first ONS in 2011, which was organized by the Stanford Clean Slate Program and the then Open Networking Foundation (ONF). The team in the Clean Slate Program evolved to become ON.Lab, which eventually took over the ONF. In that first conference the speaker line-up included:
• Guido Appenzeller, Big Switch
• Martin Casado, Nicira
• Matt Davy, Indiana/GMOC
• Stu Elby, Verizon
• Samrat Ganguly, NEC
• Saar Gillai, Hewlett Packard
• Jonathan Heiliger, Facebook
• Nick McKeown, Stanford/ONF
• Dan Pitt, ONF
• Jennifer Rexford, Princeton University
• Scott Shenker, UC Berkeley/ONF
• Steven Stuart, Google
• Rob Vietzke, Internet2
• David Ward, Juniper
And barely six months after that first ONS in Stanford, the second ONS went mainstream with Guru Parulkar as the ONS Chair. That show sold out and had a significant wait list. In any case, the prominent individuals above are familiar to many readers who’ve been following the SDN and NFV game for some time. Many of these industry notables are still involved in open networking, though almost all have seen significant role changes in the ensuing eight years. Just another sign of how fast our industry changes.
This year’s ONS 2019 had plenty of rich content in various tracks, from edge to AI to network automation. And yet attendance appeared lower than last year’s event (no, I don’t have the final counts for ONS 2019 attendees) and the buzz factor around the exhibits felt more muted. Part of the reduced attendance was due to the denial of visas for Chinese technologists looking to attend the show – an unfortunate outcome of the current geopolitical climate. Nevertheless, it was hard to not compare this with the recent OCP Summit in the same conference center, which had a lot more foot traffic in a larger exhibits area. And it’s unfair to compare ONS with DockerCon or KubeCon just in terms of sheer numbers.
Open Networking – Midlife Crisis?
Does this mean that open networking is in trouble? Does this spell doom for SDN and NFV? Now that the initial exuberance of the first five years has dissipated are we stuck with banality and incremental improvements? Or are we just going through a midlife crisis for open networking?
My thoughts are that technologists, like the general public, tend to focus on shiny new objects. In this case the new object is containers and perhaps serverless. The reality is that containers have wider applicability and the overall ecosystem size is larger – from application developers to cloud infrastructure to network infrastructure folks. On its own, network infrastructure has always been the realm of a much smaller group of people with a base that isn’t growing much more as we try to keep headcount the same or lower, moving to public clouds, or leveraging technology to scale network management.
In some ways open networking gets sandwiched between the underlying hardware, which is seeing a resurgence in innovation through disaggregation, NPUs, field programmable gate arrays (FPGAs), merchant silicon, and the application layer on top with containers and cool new cloud applications. And we’re stuck in the middle looking a little like the middleware layer of old with its uncool perception – and it’s a space in which it can be hard to make money. I will note that BEA, IBM, Oracle, Microsoft, Software AG, Tibco, and many other middleware product companies did build decent-sized businesses during their heyday.
The problems we’re working now are hard – orchestrating across multiple service domains; integrating across large swathes of both new and legacy telco and enterprise networking equipment; managing networks across private and public clouds; and rolling out 5G services with cloud architectures while transforming telco cultures. Makes a midlife crisis look pale in comparison.
The ‘Open’ in Open Networking
In any case, ONS did include a bunch of new content. From the blueprints put forward by the ONF that demonstrate how disaggregated and software-centric solutions apply to some of the hottest use cases at service providers like SEBA and ODTN; to new verification programs by OPNFV; to continued commitment to SDN and ONAP by carriers; to the announcement of a leaner approach to NFV called Lean NFV.
With regard to ONF, I know there are some who worry that “curated open source” isn’t really open source and suspect these reference designs from the ONF are merely product placement opportunities – like in a TV show or movie. I have two thoughts around this.
First, the meaning of “open source” has changed so much from the original intention that I’m not sure what it means any more. Second, perhaps we should laud the ONF for finding a way to remain commercially viable as it pursues its mission.
I’m not going to touch the debate of whether “curated open source” is still open source or whether we’re sullying the term. I’m also not going start another debate about open architectures and frameworks, open APIs, open standards, and open source – plenty has been said around that already.
For us at AvidThink, open source can be a business model (like IBM/Red Hat); a philosophy (like GNU); a community; a collaboration and sharing model; a recruiting strategy; or even a marketing strategy. Chris Wright from IBM/Red Hat has some good thoughts to share, so if you’re in the mood, find him at the next conference for a chat. Fundamentally, I would say “caveat emptor” before you get involved in any open source project or embed open source into your own codebase. Understand the ramifications and the underlying business model just as you would any other type of software.
Orchestration and Automation — Big Loop vs Small Loops
So, what did AvidThink learn at ONS this year? First, we’re starting to see maturity in open networking. Discussions between the various standards and open source groups are active, as was evident on stage – we know Arpit Joshipura of LF Networking was personally excited about this. (“Tweet this, Roy, you’ve gotta tweet this.”) We support collaboration though friendly competition between groups as it can be healthy and drive innovation.
We also learned that our views on ONAP are shared by a large number of ONAP participants. While all the major carriers in the world have committed to ONAP, deployment challenges still abound and there’s a lot of work to be done.
The other interesting trend we saw was what we’d like to call “big loop versus small loop.” While we’re working hard on the difficult orchestration and closed-loop systems across large systems and across entire service domains, we’re seeing small bits of automation in niche areas take shape. Tasks such as simplifying a local deployment, translating OpenConfig models into actual device configurations, augmenting these models temporarily with vendor-specific schemas to make them work, and then doing validation locally to make sure the intended config has been implemented. These simple little wins are what enterprises and service providers are working on. To that end, Itential, Frinx, Teoco Systems, and others like them are getting traction that will, hopefully, translate into larger wins longer term as we work toward larger orchestration loops.
There’s a lot more to cover, which I’ll save for another day. Suffice to say that open networking is starting to mature and while the momentum isn’t as obvious, I believe it’s making progress. Sure, containers are cool; I’m looking forward to DockerCon, and I know for sure that KubeCon+CloudNativeCon Europe in Barcelona, Spain at the end of May will be a blowout event. But someone’s got to dig the trenches, replace the plumbing, and clean out the sewers – infrastructure work is hard. And no, I don’t think open networking should eat, love, pray, or grow a mustache or buy a convertible – maybe it just needs a makeover.