A new age of openness is coming upon us. At least that’s what we’re being told. For instance—“The reign of closed solution suites is over, shifting to the rise of open, heterogeneous software ecosystems.” Maybe it’s my 30 years in the information technology business (how many people remember Thomas-Conrad ARCnet hardware?), but I’m not convinced. It’s worth taking a moment to consider the case.

On the positive side, there is a huge movement towards openness in many areas of the IT world. It is slowly becoming possible, for instance, for mid-scale operators to disaggregate their routers and switches into multiple parts, each purchased and managed to obtain the best bang for the buck. In this regard, the importance of the router (or switch) as an appliance certainly seem to be on the wane. The open source movement, standing on the shoulders of open standards, certainly seems to be making huge strides. There are now a number of open source routing stacks available (including Free Range Routing, forked off of Quagga).

Various flavors of *NIX are available through open source that are production grade, and many companies are contributing large and important projects (such as KAFKA) to the community. These open source projects form the backbone of the cloud, in fact; cloud providers largely build their services on open source software and white box hardware. Open19 is accelerating the move towards commodity compute and storage, as well, making the white box buy much more compelling for mid-scale operators. The existence of large-scale, widely available development platforms is making a lot of companies ask: Why buy hardware from a name brand vendor when you can rent a cheaper version that someone else maintains?

But all the roads in the world do not lead to open software systems. There are several counter movements that need to be watched carefully if we are to see the whole picture.

The first to note is the Software-Defined Wide Area Network (SD-WAN) movement. While it might be fairly invisible to hyper- and web-scale operators, it is “in-your-face” for last mile and transit providers. SD-WAN is taking the wide area world by storm, with most transit and last mile providers either scrambling to keep up, or partnering with an existing company in the space. More importantly for the question this post is asking: SD-WAN is based on completely closed, proprietary solutions that do not interoperate with anything else.

The second to note is the serverless movement. At first glance, serverless is just another stage of the cloud. First, you remove the storage, then the compute, then the network, and, finally, the operating system. But there is a more important point in the serverless revolution. To go serverless, you must move your applications into an API controlled by a provider. While these applications might well interoperate with other applications and systems, they must do so through the facilities provided by the operator. Serverless is, to put it more familiar terms, at least for someone who used to work on mainframes and minis, a pretty mainframe’ish version of the cloud. Again, the very definition of a closed, proprietary system.

So there are several examples of movement towards open software running on commodity hardware. There are, at the same time, several examples of movement towards closed systems running on what is essentially proprietary hardware bundled with software. Neither case is a “pure appliance” play, but both cases rely on strong vertical integration to create a new way of solving old (and sometimes new) problems.

What does all of this tell us? If you hold one set of facts steadily in view, it appears the days of proprietary solutions are over. If you hold another, it appears they are just beginning. The result seems to be this: open software is not going away anytime soon, certainly enough. At the same time, to declare the proprietary “wars” as won, and the age of the closed system over, seems a little ahead of its time—if there ever will be a time in the IT world when such systems will cease to exist.

After 30 years, the lesson seems to be the same: the more things change, the more they stay the same.