For the last several years, I’ve written end-of-year posts to predict the future of network virtualization. This year, I’m expanding my scope to cover what I expect for networking more broadly. As I suggested at VMworld this year, networking is undergoing a major transformation, and this will become more clear in 2017. This post is a little different from some of the other prediction posts you may have read from VMware, which cover everything from the Third Industrial Revolution to hyper-converged infrastructure to security. Put simply, we think networking is different. We’re not just seeing a few modest changes that will take place in networking; we’re looking at a fundamental shift.
The Future of Networking Lies in a Shift of Power
To see the future of networking, we need only look at one of the dominant trends in IT: the shift in power toward lines of business and developers. With the developer becoming increasingly important and empowered to choose their own tools, and lines of business adopting public cloud capabilities to become more agile, networking is no longer solely the province of the IT department. Developers need networking capabilities to build and deliver applications; at the same time, IT needs networking (and security) capabilities to manage risk, compliance, cost, and so on.
Networking is no longer about buying the best piece of hardware for my data center—it is about delivering the critical capabilities to meet the needs of applications, no matter what infrastructure they run on or who owns that infrastructure. In other words, networking needs to move from being part of the infrastructure to being independent of the infrastructure.
This is not an incremental change to networking—it is a fundamental shift. Networking must simultaneously meet the needs of two distinct communities: developers deploying applications on a heterogeneous collection of infrastructure, and IT managers trying to apply an appropriate level of control to ensure compliance and security without impeding developer productivity.
The Effect of This Fundamental Shift
The effect of this shift will manifest itself in a variety of ways. One is increasingly sophisticated options for container networking. We’ve already seen a proliferation of approaches to networking for containerized applications (e.g., CNM/libnetwork and CNI). These approaches need to mature to the point where developers are able to easily consume the networking services that they require, while IT managers are able to ensure appropriate security controls. For example, not only do developers need connectivity among the microservices that make up an application, they will typically need other networking capabilities, such as load balancing or secure access to corporate servers holding relevant data. Simultaneously, IT is looking for networking services such as firewalling and visibility into data access patterns. To this end, pluggable networking for containers seems a likely path to enable a range of options supporting the needs of both developers and IT.
Similarly, as businesses face the reality that their applications will run in multiple clouds, both public and private, there will be a need for networking and security services that can readily be deployed consistently across these disparate cloud environments. In concert with the rise of new application frameworks, this drives the need for networking capabilities to be truly independent of the underlying infrastructure on which the application runs—independent of whether the infrastructure is public or private, independent of whether the application is virtualized, containerized, or bare metal.
The increasing independence of networking services from underlying infrastructure also leads to the rise of the software-defined WAN (SD-WAN). Networking in remote branch offices will be independent of the capabilities of the traditional WAN providers, responsive to the needs of the business (e.g. direct access to cloud services from the branch), and delivered in software.
Furthermore, this transformation of networking increases the relevance of open source software. In particular, to provide useful networking services to developers using public clouds and open-source container frameworks, it will be essential to offer networking capabilities that easily integrate with those environments, such as Open vSwitch. Open source networking for the data center (e.g. OpenSwitch) is also likely to gain prominence.
What’s to Come
In summary, 2017 is poised to be a year in which networking undergoes a fundamental transformation. Networking is no longer about which vendor provides the best equipment to the traditional buyer of networking, the IT department. Developers require a rich set of networking capabilities that enable their applications. IT requires networking capabilities that enhance developer productivity, no matter what application framework or deployment environment they choose, or where the applications are ultimately deployed, while providing IT with appropriate controls for security and compliance. I think the pieces are in place to meet that challenge— of course we’re working on some of them—and we’ll see this transformation gather steam in 2017.
Read more from VMware executives on what’s to come for technology innovation in 2017 and beyond.