As organizations evolve to meet the challenges of ever increasing demands on infrastructure, it is critical that the network no longer get in the way. Now is the right time for a solution that is catered to the needs of Networking for Cloud Native environments.
The way things are…
Many organizations are managing development, operations, and network as three distinct pillars – with little overlap and a set of deliverables that vary greatly. This structure breeds contempt and creates an environment where it is unlikely that everyone is “on the same page”. Not only do these teams have different types of engineers, with different skillsets, but they are all using different tools and managing the environment in unique ways – with little sharing of knowledge and ideas. Finally, each team is measured and subsequently rewarded according to different metrics – the devs are measured on how quickly they can build a new feature or application, ops on how quickly they can stand-up a new environment, and the network team is judged on how many hours of uninterrupted up-time they have provided. This misalignment of goals yields an atmosphere where agility is not rewarded and forces teams to say “no” to change as anything new might compromise their ability to deliver or causes increased risk to the up-time.
It’s getting better…
With the advent of DevOps as a set of operational principles – forward momentum was made in making infrastructure operations more streamlined and harmonious. Using DevOPs and CI/CD pipelines – compute operations can now adopt principles of application development to continuously improve the environment. This cycle of constant improvement and progress has proven invaluable for the compute world, but has distinctly left the network in its own silo.
Many attempts have been made to apply the principles of DevOps to the network – but these mainly focus on applying tools purpose-built for one thing to crudely manage another. Instead of extending the existing DevOps tools to manage the network natively, most efforts focus on trying to represent network models and principles in a “server-like” manner. This falls short in acknowledging that network devices are inherently different from their compute and storage counterparts and need modeling and operational principles that match these needs. A top-of-rack switch is not just a server with 48 or more ports.
Network for Cloud Native
With the rise of public cloud platforms, organizations are adopting management principles and toolchains to manage on-premise deployments in a similar fashion. The only thing left behind is the network. The only way to remove the network from its prescribed silo is to elevate it to the same level as the rest of the environment.
The network needs to be managed using the same Cloud Native tools that are being used to deploy containerized infrastructure. The network has to be built using an OS that adopts these principles and presents the network using models and paradigms that are native to the operators of Cloud Native environments. This solution cannot use add-ons, bolt-ons, extensions, or other proprietary and bespoke solutions – the network needs to be managed using Cloud Native tools, natively.