Every year pundits make their predictions for the next year, based largely on the technology cycles they see, knowledge of the market and its history, and the formation of coffee grounds left on the bottom of their coffee cup. That last one is usually the most accurate, by the way. It’s the one I’ve used the most over the past, well, many years in technology.
Some things, however, are pretty clear given the shifts and movement across the entire data center. Software-defined everything is going to be a big deal because there’s no other realistic way to scale operations to keep up with demand generated from the exponential growth in applications needed to support the API economy.
That API economy is not restricted to developers. We tend to use that phrase to refer to the Internet of Things, mobile applications, and even modern integration methods, but there’s a second, lesser known API economy that will be absolutely essential to the business and the data center that supports it.
It’s the infrastructure API economy, and it’s going to continue to be a significant factor in not just operationalizing the data center in order to scale the systems on which just about every business depends today. It’s also a major driver of the technology that’s needed across the data center. Whether it’s the network, application services, security, or storage, the impact of the infrastructure API economy can be seen in increasing adoption and demand for frameworks and tool chains that focus on comprehensive data center automation and orchestration.
With that in mind, I’m going to go out on a fairly sturdy limb and offer a few predictions for 2016:
1. Enterprise NFV will become more attractive as SDN continues to struggle. SDN has made very little inroads since it was hailed as the “game changer” for network infrastructure. There’s a variety of reasons why that’s the case, but suffice to say that a lack of clear definition and benefits has failed to convince most enterprises that the investment in replacing existing network infrastructure is worth it.
Network functions virtualization, however, has been making great strides in the data center primarily by focusing on exactly what enterprises crave: scalable network and application services provisioned with alacrity through automation and orchestration. Efforts to adapt NFV for enterprise use have already begun, and it seems a fairly good bet to expect those efforts to ramp up in 2016.
This doesn’t mean SDN is dead and, in fact SDN will likely live on in modified forms. One could reasonably argue that enterprise NFV is a form of SDN, after all, as are many solutions and frameworks that initially grew out of the rise of SDN.
2. OpenStack will dominate data center orchestration. A few years ago, it appeared as though OpenStack might go the way of the dodo. Plagued with upgrade and migration issues, as well as significant changes in direction across “modules,” it seemed as though only those enterprises with excess operational and capital budgets would be likely to adopt. But over the past year we’ve seen significant strides made by OpenStack not only in stability and support from those with a non-trivial interest in the infrastructure API economy, but in adoption, too.
The momentum is likely to continue and push OpenStack to the top of the DC orchestration stack as there are few others -- commercial or open source -- that offer the same breadth of data center stack support in what is becoming an if not mature then certainly almost mature DC orchestration system. Don’t be surprised to see some combination of OpenStack and NFV become the “enterprise NFV” orchestration system of choice.
3. Infrastructure as code will move into production. The notion of “anything” as code involves the ability to encapsulate a configuration and essentially “replay” it for provisioning purposes. The concept comes squarely out of DevOps, where developers and operators have been using the idea to better automate the processes that govern building and delivering an application to production, but also is seen in SDN-related technologies.
This notion will extend into production and will eventually impact every part of the infrastructure, but in the next year expect to see this idea play out in network and application services closest to the application where agility and speed of deployment are becoming increasingly imperative. The use of templates “as code” to provision network and application services close to the application will increase consistency while decreasing error rates in production, leading to faster, repeatable, and more stable deployments.
Whether or not you agree with these predictions, I think it’s likely you’ll agree that automation, orchestration and the infrastructure API economy are going to drive technology and solutions in 2016 because there are few other options for the data center. If it’s going to keep up with business growth, it’s got to be a lean, mean resource machine that scales as elegantly and quickly as its very attractive alternative: the cloud.
The coffee grounds have spoken.