Blog

Filter by:

Blog Subscription Form

7 Seismic Shifts Remaking Networking

Charles Stucki
By Charles Stucki | 11.14.18 | No Comments

Two years ago, at the SDN NFV World Congress in the Hague, I spoke about applications needing an effective interface to the network. I was the Cisco product leader responsible for Tail-f, called Cisco NSO, and several platforms built on it. I wasn’t wrong, and those platforms continue to drive virtualization and automation within Cisco’s core networking businesses.

However, earlier this year, I - now an entrepreneur - along with my Bayware CTO, interviewed dozens of teams across several digitally-savvy industries who were in the thick of enabling networks to respond to the relentless deployment of new application and updates.

What we found was a set of seismic shifts that are remaking networking – well beyond virtualization and automation.

Here are 7 takeaways – about which I intend to share more in the future, so stay tuned.

  1. Applications. Software development and deployment takes all the spots at top of the technical agenda…provided that an enterprise can secure their digital assets from attack. Digital skills are software skills. Networking professionals are rapidly shifting their skills and orientation from specialized infrastructure operations to continuous software application deployment.
  1. Scale-out. To be more like cloud native systems, all complex systems are breaking down into smaller, simpler, easier-to-upgrade elements that are orchestrated into large-scale systems. This is a reversal for networking. As Vint Cerf points out, for decades routers were the heart of networking, and they got larger, more complex and harder to upgrade. Containerized microservices – including for network functions - are key concepts to explore further.
  1. Multi-cloud. The public clouds dominate primarily because they are technically and budgetarily outstanding platforms for developing scale-out-ready applications quickly.  But, to be responsive to customer needs, most teams are actively ensuring applications are portable, i.e. not tied to any one public or private cloud.  Linux and Kubernetes are key enablers to revisit. A network cannot be defined by place anymore.
  1. Agile. The data is in – see multi-year research on over 4,000 organizations by Forsgren et al. Teams that master the culture, technology and processes of lean, continuous and automated flow in development and deployment are wildly outperforming teams who don’t, on both velocity and stability, including security.  If someone tries to tell you Agile and DevOps are just buzz words, send them out of the room. Instead, start asking about how to bring networking and security into the Agile DevOps world.
  1. Zero-Trust. No application trusts any network. Determined attackers can compromise any perimeter. The application solution is, “building security into software development not only improves delivery performance but also improves security quality.” (see Accelerate by Forsgren et al.) The network solution to Zero-Trust is Default-Deny, that is to encrypt everything and default-deny access. The default is no network flows except as authorized by application services. I intend to talk about microsegmentation of applications in later posts.
  1. Identity not Address. Applications can’t rely on IP Addresses for service to service communication that achieve company and security policies, because IP addresses change at an accelerating pace. In an orchestrated environment, the average lifespan of a container is 12 hours. (Datadog, 2018) And every cloud and data center has a different schema for assigning and recycling IP addresses. Addressing has to be based on identity that is meaningful to and authenticated by the application. Talk about a disruption to networking!
  1. Observability. Also known as visibility, instrumentation, and telemetry. “Ultimately,” says Matt Klein of Lyft in his presentations - voicing what I heard from so many others, “for service-oriented architecture networking, robust observability and easy debugging are the most important thing. The network should be transparent to applications. When network and application problems do occur, it should be easy to determine the source of the problem.” This is a particular challenge and opportunity in hybrid and multi-cloud deployments.

In articles to come, my colleagues and I at Bayware.io expect to share more about how these seismic shifts create great challenges and intriguing opportunities for those of us who came from the networking world and are dedicated to meeting the future with software innovation.