Today’s applications are connected both to users and other applications, increasing traffic and profoundly affecting performance. Recently, there has been much discussion around the Application Performance Management (APM) category. This got me thinking about new cloud migration versus legacy application tools. Is the industry being myopic and dealing with silos?
Wouldn’t it be amazing if the same software capabilities worked consistently across networks and servers? Sadly the deployment, maintenance and management of these legacy networks are limited by dated and monolithic operating systems of yesteryear. This gap is only getting wider. How does one truly scale across millions of composable machines and workloads?
Introducing Containerized Arista cEOS:
There is an industry quest for packing more applications on each server, known as “microservices”. This, and the acceleration of native cloud apps deployment, creates new needs for containers. Arista cEOS™, or containerized Extensible Operating System, is a packaged application that provides controls across the network. Arista cEOS is deployed as a container upon a platform such as Docker, while Linux provides the native OS for the network with the same baseline state mechanisms as NetDB. Shown in the figure below:
Containerization also has multi-faceted use cases, including when:
- The network switch can be managed as a server, resulting in uniformity and savings for network operators.
- The cEOS driven network can natively host other application containers. This allows Arista partners and customers to create applications specific to customized needs.
- cEOS can be deployed in public and private clouds to provide network functionality between virtual machines and containerized applications.
- cEOS can simulate large networks and their behavior under load to gain insights from this simulation in real deployment.
- cEOS can work with OCP (Open Compute Platforms) such as Microsoft Sonic, or Facebook Wedge, or HPE servers for consistency and visibility.
Customers have the ability to deploy EOS natively on a physical switch or virtual appliance for virtual networking (vEOS), or now as a containerized application (cEOS). We continue to work closely with our partners such as VMware and Docker to expand visibility via our tracers, such as VMTracer and Container Tracer, in a transparent, seamless manner across diverse workloads.
Summary: A Catalyst for Cloud Containerization
At Arista we see containerization as the frontier for data driven networking. The ability to isolate network interactions transforms the network from an IT cost center to a business driver. In this new software driven learning paradigm, EOS can self-learn the data (consisting of many micro work-x services), instead of approaching it as app silos. Welcome to Arista EOS 2017 for next generation cloud containerization!