A few days ago, Arista Networks announced a couple of new cloud networking features that enable companies to move to a NetDevOps management model. For those not familiar with the term “NetDevOps,” here’s an explanation: As the name implies, it brings DevOps principals to network operations with the ultimate goal of running a network as agile as running cloud infrastructure.
The challenge is that network infrastructure was never designed with agility in mind. In fact, one could argue the traditional vendors purposely kept the complexity high to drive up the amount of service and support that networks required.
Go here to see eWEEK’s Resource Page on Intent-Based Networking.
Arista announced cloud networking capabilities that enable streamlined operations for running virtualization and containers in hybrid, multi-cloud environments. Its new CloudEOS operating system has an integrated VM that makes it easier to spin up network resources to create dynamic connections between on-premises and public clouds and between public cloud providers. This allows network traffic to be continually directed to the best path as determined by a number of network attributes, including analysis of the telemetry information available in EOS.
DevOps Tools Can Be Used with CloudEOS
Because CloudEOS, like traditional EOS has no hardware dependencies, the provisioning can completely be done in software through DevOps tools that support declarative models, such as Terraform. Security-conscious individuals need not worry about traffic traversing the Internet as Arista encrypts all traffic.
For organizations that have moved to a cloud-native operational model, CloudEOS also enables customers to set up the network operating system as a Kubernetes-managed container that provides a full network stack. Now as other containers are spun up, so too can the appropriate network resources. Typically, this might have a lag time of hours, days or even weeks, but through orchestration tools, the network resources can be spun up in conjunction with applications.
As is the case with all Arista products, customers can use the vendor’s CloudVision management tool, which is designed to be a single pane of glass for end-to-end visibility, telemetry, configuration management and orchestration. Think of CloudVision being to Arista what vCenter is to VMware customers.
CloudEOS also has native support for pay-as-you-go pricing through Amazon Web Services, Google Cloud Platform and Microsoft Azure. The benefit of this is it enables elastic cost models for network connectivity. Customers only pay for what they use and can be scaled in real time, so there’s no need to over-provision the network, which is the norm today.
Network Segmentation a Major Use Case for CloudEOS
One of the best use cases for CloudEOS is network segmentation. Cybersecurity continues to grow in importance--and not just with IT people, but also business leaders. This has increased the need to segment the network. The challenge in the past is that segmentation was very difficult to configure and required constant updating of the network. Arista’s CloudEOS can extend the reach of the network from the campus to the data center and then out to the public cloud using open standards. This obviates the need to deploy independent technologies to deploy segmentation in the data center, cloud and campus and then having to stich these together. One network, one segmentation plan.
The shift to cloud-native IT models has drastically changed the network. The long lead times that network engineers are gone, and the network needs to be as agile and dynamic as the rest of IT. Network operations must evolve to a NetDevOps model where resources can be provisioned with the needs of applications. Many network vendors have solved this for one part of the network, such as campus, cloud or data center, but Arista’s CloudEOS is a designed to bring consistency across the entire network.
Zeus Kerravala is an eWEEK regular contributor and the founder and principal analyst with ZK Research. He spent 10 years at Yankee Group and prior to that held a number of corporate IT positions.