There is little debate that open source is increasingly important across all of the major infrastructure areas: compute, storage, networking, and applications. But the role of open source, and in some cases even the purpose, is being changed as the major drivers shift from vendors whose primary objective is to carve out a business to the very users whose infrastructures leverage the open source components and tools.
As open source changes, it means that the surrounding landscape will have to change with it. And as with all change, this creates threats and opportunities for those who make their living in and around IT.
Open Source in Networking
Open source has actually played a key role in networking for several decades. Virtually every carrier-, enterprise-, or cloud-grade system available leverages (and has for some time) key open source elements. Whether it’s the underlying operating system or the SSL libraries, it is difficult to deploy a networking device today without having some open source component in your environment.
So even if nothing were to change, I would encourage any user of networking to begin developing an open source strategy if one does not already exist. Minimally, you should at least understand what open source exists within your environment if for no other reason than to track the known issues and vulnerabilities that tend to proliferate, at times quietly, through large swaths of the IT world. More optimally, you would develop a strategy for evaluating and qualifying key open components, and potentially even contribute to projects that advance your company’s interests.
The Rate of Open Source Inclusion Is Increasing
It’s not just that the major networking players have adopted open source strategies to accelerate development. The rate of open source inclusion is actually increasing.
The open source acceleration is driven by three primary dynamics: more mainstream understanding of the role of open source in production environments, the maturation of major open source projects, and the use of open source as a primary strategic lever.
I don’t need to spend too much time on the first two. It is observably true that open source is mainstream. The proliferation of Linux was enough to make this happen, and then we have added broader efforts like OpenContrail, OpenDaylight, ECOMP, and the like. And as these projects get deployed in high-profile networks, a lot of the risk that used to be associated with open source is suddenly rendered less potent.
But the use of open source as a major strategic lever deserves some special treatment.
Know Your Enemy
If you are a major service provider, who is the competition you worry about most?
AT&T certainly has to take steps to compete with Verizon. But Verizon represents a well-known foe. Like teams that face off in major rivalries in collegiate or professional sports, when you know your enemy, the battles are familiar, and the tactics are easily understood. It’s not that the competition is not fierce, but because it is familiar, the risk is not about the unknown so much as about execution. If you have wondered why AT&T has spent so much time focused on improving execution through a software focus, it’s because that’s exactly how you prepare to fight a longtime foe.
But the real risk lies not in Verizon. Or Sprint or CenturyLink or even Comcast. Those are all competition, which might be challenging at times, but the plays to run there are well understood. The risk—the unknown, if you will—comes from non-traditional warfare.
As broadband access improves and as access and WAN go through a technology refresh, we will see that more and more applications will perform reasonably well using regular broadband for connectivity. This is certainly true for consumers already, and it’s largely true for most smaller branches or for branches that primarily perform small functions.
With base Internet connectivity and some fairly basic services as all that’s really required in the home or on-prem, it’s a pretty small step to go from home assistant to gateway device for the likes of Amazon and Google.
If the existential threat is coming from the cloud players, then open source becomes an interesting lever to pull for AT&T. In fact, where they would normally fight for differentiation against their more traditional competition, they are now willing to open source. AT&T has pushed into the open source space in a big way with ECOMP. It’s a massive undertaking, and they are no doubt keeping some of the jewels under lock and key, but to put that much effort into an open source project signals a change in strategic disposition. Of course, it doesn't hurt that they can swell their developer ranks without hiring more people — a nice side effect of effective communities.
If they can use open source as a lever to bring down industry-wide costs, they can effectively arm the entire telco space with some of the tools required to make transport and managed services more cost effective. In doing so, they can help circle the herd while Google and Amazon attack from the outside. It is actually in AT&T’s long-term worst interests to allow these non-traditional players to gain any foothold from which they can build a better offering, even if that foothold is in someone else's backyard.
The Cloud Players Too
The cloud players are also contributing to the change in open source. Facebook is perhaps the best example with their efforts around the Open Compute Project, but virtually every major web scale and cloud property is engaged in this type of effort.
To understand how this works, we have to understand where the intellectual property is. For most of the web scale companies, the real jewels are less in the infrastructure and more in how it's managed. So as they innovate in the hardware or even the platform software, they can open that up for the world without really impacting their competitive moat. Essentially, they improve their negotiating position for some components while preserving the things that matter most to them.
And so we end up with a dramatically different open source landscape now that these players have solidified their operations and moved beyond.
It’s Just Starting
But the changes to open source are just starting. In fact, in the container space, we are seeing companies start to band together with efforts like Kubernetes and the Cloud Native Computing Foundation (CNCF). It might seem strange to see the cloud companies starting to align around a small set of common efforts.
But who is the primary enemy for the cloud properties? The obvious answer is that they are each others’ enemies. But the reality is that the biggest competition for cloud at this point is still non-consumption. That’s a fancy way of saying that people doing the same old thing is what they need to compete against. And if they collectively make it easier to move to the cloud, then they can all compete on the strength of their offerings.
In fact, as the cloud companies become more differentiated in their products, they will actually collaborate more in the open source arena. When cloud was just starting, they were all roughly after the same thing. Now that Microsoft is attacking with applications, Google with TPUs, and AWS with their sheer presence and market dominance, they will be freer to collaborate to take down traditional IT.
And to do that, they will undoubtedly step up their open source game above what is already very impressive engagement.
Clouds Love Open Source, Except When They Don't
Of course, this is all at a macro level. Cloud companies all love open source. This is true except when, of course, they don’t. Most of what the cloud companies love is their operational infrastructure. And insofar as open source supports that, they love it. And when it doesn’t, they love whatever they can get to make it easier to plug existing things into their frameworks.
So we will continue to see—at least in the networking arena—small battles like gRPC and OpenConfig that might not have ubiquitous end-user support, primarily because it fits into some existing framework. And those efforts will also impact open source because, depending on how the architects try to flex their muscles, we might very well see fractured communities emerge as people try to garner support for their secret sauce.
The Bottom Line
My major point in all of this is that open source is a major contributor to networking today, and it’s going to become more important over time. But the role of open source is not merely because someone wants something to be open. Increasingly, open source is going to be a strategic weapon for opening up markets and fighting B2B battles.
And this means that open source groups will increasingly become a battleground where proxy wars are fought between groups of people advocating for interests that are not always well-publicized. Minimally, I would think that consumers would want to at least survey the landscape to develop a point of view on how some of these battles will impact them. They could, for instance, signal the rise or fall of major efforts, which could render architectural decisions obsolete over time. As a means of risk mitigation, open source strategies simply have to exist.
The winners in all of this? The users always win when competition heats up. Minimally, it improves things like choice and flexibility. And more ideally? Someone will end up serving the entire industry out of what might originate as pure self-interest.