Arista expands extensible operating system

Arista Networks has announced the next major expansion of the Arista EOS® network stack with the introduction of the EOS Network Data Lake (NetDL™).

  • Thursday, 4th November 2021 Posted 3 years ago in by Phil Alsop

Coupled with the AI/ML-driven Autonomous Virtual Assist (Arista AVA™) and a broad ecosystem of industry leaders, Arista is extending the EOS network stack architecture to provide a high-fidelity data lake capability for the next era of data-driven networking. This novel foundation supports a wide variety of applications to interact, analyze, and enrich network data for best-of-breed enterprise solutions.

 

The combination of EOS NetDL and AVA with validation from industry leaders provides high-value integrations that improve end-user and network operator experiences, simplify operations, and enable a shift from reactive troubleshooting to proactive, predictive, and prescriptive outcomes.

EOS Network Data Lake - The Next Frontier

Arista EOS has over fifty million ports deployed in the campus, data center, routed WAN, and in carrier and cloud environments. EOS NetDL builds upon the established second-generation of the Arista EOS NetDB core publish-subscribe state model with datastores and analytics for network data sources such as alerts, flows, full packet capture, control plane traffic, and device state streaming. In addition to network-centric state from Arista devices, EOS NetDL also enriches network data with third-party and external data integrations, enabling a wide range of applications to process, analyze, and derive operational insights and predictions from this data set.

 

NetDL provides a single source of network data ‘truth’ and a common sensor/collector architecture that enables forensics and analytics for threat hunting, network packet brokers, network detection and response, network performance monitoring, and application performance monitoring. By collaborating with a variety of industry leaders, Arista is able to deliver powerful production customer benefits. These data-driven network models can be transformed into insights that deliver actionable operational outcomes.

 

“Arista is entering the third generation of its flagship software stack. Developing a network-based data lake foundation from the ground up on our existing network state database makes Arista EOS NetDL a differentiated network and data-centric operating system. This compelling multi-tenant and multi-modal data stack is future-proof and engineered to enable customer innovation,” stated Ken Duda, Founder and CTO at Arista Networks.

 

NetDL Enhances a Broad Industry-Wide Ecosystem

Arista’s EOS NetDL architecture is supported by a broad industry ecosystem of reputable leaders committed to customer solutions and integrations with data-driven networking.  The EOS NetDL ecosystem of solutions and partners delivers multiple options of adding high-quality enrichment to the core networking data set and providing contextual value, rich analytics, and turning data into actionable insights across a broad range of client operational use cases.

 

Key integration examples with NetDL ecosystem alliance partners include: enhanced network segmentation and cloud security integration, Quality of Experience (QoE) with collaboration applications, validated designs and templated deployments, full inventory integration that verifies production devices are accurately tracked, and live chatops/bot integration enabling real-time data consumption and interaction.

 

AI/ML Enrichment with AVA

Building upon EOS Network Data Lake, Arista is also extending AVA capabilities for AI-assisted workflows. Arista AVA is an AI-enabled decision support system that combines cloud scalability with the codified expertise of real-world network and security operations experience. AVA utilizes the breadth and high-fidelity depth of enriched data that exists within EOS NetDL to deliver targeted AI-based capabilities that enhance network operations and engineering with real-world use cases leveraging advanced pattern-matching capabilities.

 

This data-driven network approach helps to improve the efficiency, availability, and security of modern network operations. As an example, AVA can leverage network state data in NetDL to monitor the performance of optics to predictively model a critical component failure and generate proactive alerts before the performance and availability of the network are impacted. The combination of Arista AVA and EOS NetDL provides predictive and prescriptive intelligence for data-driven networks.