A recent survey of more than 13,000 developers, published by research firm VisionMobile, noted that developers (and in particular enterprise developers) aren’t always in control of the environments in which they develop and surely aren’t in control of the environment in which they are deployed. Yet despite being told that public cloud is going to eat IT, 44% of developers rated private cloud as the most popular cloud development platform, ahead of AWS (16%) and Microsoft Azure (13%)” according to the survey.
It’s easy to get caught up in the private versus public cloud argument, but the reality is that if 44% of developers are targeting private cloud, then 56% are targeting public cloud. In your org maybe it’s 50-50, or 60-40, or even 80-20. Regardless of the ratios, the reality is that the current state of the enterprise is a hybrid environment and it’s going to stay that way for a while. For a variety of reasons, organizations are maintaining a presence in the data center even as they begin to take advantage of public cloud. In the long run, the winners are going to be organizations that learn to maximize benefits while minimizing costs across the entire cloud spectrum: private, public, and SaaS.
This actually has a pretty profound impact on data center network engineers who must adapt and architect solutions that deliver services and connectivity in the context of a multi-cloud (hybrid) environment. The data center is no longer constrained by physically defined perimeters; it’s now inclusive of external systems and services. The data center is now a fluid border that changes with each application deployed in a new cloud environment external to the data center.
Further complicating the task of modern networkers is the growing need for parity of security policy and performance across these new and very ethereal boundaries. Both remain high on the list of challenges cited by those trying to deliver apps in hybrid environments. The network and application services that securely deliver applications must be enabled for deployment in an increasingly cloudy world and that means networking pros must ready themselves to do that across both public and private environs.
Ultimately, this requires adopting a mindset that is cloud first, not so much in location or private versus public, but in terms of enabling and supporting cloud in all its forms. Being a cloud-first networking team means thinking less about port density and more about provisioning velocity. It means less focus on VLANs and more on visibility. Being cloud first requires considering what it means to fit into an increasingly cloudy world that spans not just the corporate data center, but the Internet itself.
That might involve a number of things: deploying services in a public cloud; developing a greater mix of hardware and software (virtual) network services to better serve the growing private cloud and traditional corporate application portfolio; and decentralizing some network services while outsourcing others to the cloud, as a service. It definitely involves considering how well a network or application service platform will fit into (and enable) your private -- and public -- cloud environment. And it absolutely requires considering programmatic characteristics like APIs and templates and integration with management and orchestration systems.
Cloud has been one of the most disruptive trends of the century, and it continues to be disruptive as the aftershocks of partial adoption begin to impact the network and security silos inside IT. But it’s also an opportunity to reinvent the corporate network as less of a confined, connected physical construct and more of a flexible, adaptable virtual set of services defined and managed programmatically across all environments.