Monday, May 2, 2011

Arista Opens its Big Data Switches for Cloud Developers

Arista Networks is opening its network operating system to enable native third party development. The new capabilities will enable applications that can leverage real time visibility of the switch state, in-service software upgrades of network-based tasks, and fully-automated testing and regression.

The company said this open software framework paves the way for more advanced, cloud-based applications. Cloud stack developers, big data companies, academics, and independent software developers around the world are working on EOS applications, including, Cloudscaling, Kyoto University, Morphlabs, Nimbula, and Solace Systems.

Arista is also creating "EOS Central", an online community where customers and developers can track projects and share resources, such as a growing reference library of scripts and APIs.

As part of EOS Central, Arista is showcasing a downloadable extension called "CloudVision", which combines web-scale message passing platforms with the EOS architecture to deliver a management capability that is extremely flexible: topology agnostic, in-band or out-of-band, featuring a publish/subscribe architecture, and supporting orders of magnitude more devices than competitive offerings. CloudVision is built with open-standard protocols and open source software - all modifications have been checked in on GitHub to continue to fuel innovation in message-bus architectures.

Arista noted that is two-tiered data center architecture scales to over 18,000 10 Gbps attached hosts, thereby providing the power for massive cloud deployments.

"Customers recognize that legacy enterprise vendors have failed to deliver cost-effective, simple, and scalable cloud solutions. Now they are turning away from closed, inefficient, slow, complex, and overpriced systems," stated Randy Bias, chief technology officer at Cloudscaling. "Arista EOS offers our customers an open environment with a modern software architecture enabling automation, scalability and overcoming the network challenges for cost-effective cloud deployments."

See also