/Images%20(Marketing%20Only)/Blog/Jayshree_Ullal_024.png)
Networking is undergoing a metamorphosis. Today’s operations are challenged to cope with the DevOps, NetOps, SecOps and CloudOps models that need consistent operations control. Why should enterprises care? How do you cope with decades of legacy and is change possible? Arista believes that the networking world is at the cusp of a transformation, significantly facilitated by the agile, dynamic and economic network models of the public cloud providers. They have proven the elegance of simple yet scalable designs that transform siloed networks for the data center, core, campus or branch PINs (Places in the Network) into east west PICs(Places in the Cloud). This new paradigm is a far cry from the traditional siloed network architectures that required applications to be assigned to specific servers or storage, causing fixed-function rigidity. Agility and high availability are pivotal foundations to building the new PICs.
As large data sets explode, the use of artificial intelligence applications, video and workflow traffic intermingled with breakthroughs in the mobile Internet from 5G to 400Gbps speeds, creates greater pressures on network portability. This has driven Arista to build a highly programmable state-driven software stack –Arista EOS® (Extensible Operating System) and network wide CloudVision®.
Campus is a Natural Extension
As data centers extend to the campus and branch, converting siloed PINs to PICs, applications require increased agility and elasticity, blending cloud native applications and enterprise networks via containers and APIs. The shift to IoT and microservices, be it physical, virtual or container-based with Arista CloudEOS and intelligent wired-WiFi edge connects a common spine with common protocols. This Universal Spine is enabled through CloudVision topology and inference management.
Changing Operation Models
Modern development tools today are closer to distributed systems architectures than before and need to understand failure patterns, network partitions/requirements and how their applications deploy and perform in far more dynamic environments. Traditional CLI based Network ops is lagging behind by decades and the model will only scale if declarative practices are adopted to augment mature DevOps. Ops teams are emerging ‘bottom-up’ as NetOps from IT infrastructure balance out this dynamic. The “operator” functions are rapidly becoming understood and accepted as critical roles within enterprise. CloudVision brings that network wide turnkey control.
Network wide Analytics
Migrating from old school appliances to next generation architectures requires network wide analytics. With switch-based DANZ (Data ANalyZer) and our recent acquisition of Big Switch DANZ Monitoring Fabric, (DMF) we can create overlays to CloudVision for network wide change control, automation and analytics. This is a primary driver of cost and efficiency on Day 2.
Arista Ahead
Until now the intersection of human, users, devices, machines and networking was somewhat loosely- coupled. Now one or many applications or scans can consume gigabytes of data, ingested from IoT devices such as video, storage, and real-time analytics with indexing and queuing views. Arista’s cognitive approach to high availability networking, built on cloud principles of availability, agility, and analytics, brings the combination of inference and real-time action. Troubleshooting shouldn’t be a finger pointing exercise, and live patching should be the norm instead of unplanned outages. The healthy flow of information to critical decision makers demands a cognitive network. Welcome to the new world of software defined cloud networking!
Reference:
Opinions expressed here are the personal opinions of the original authors, not of Arista Networks. The content is provided for informational purposes only and is not meant to be an endorsement or representation by Arista Networks or any other party.