Simplifying and automating data centre network operations

“After extensive research Yahoo Japan Corporation chose the strategy and best practices of a large hyperscale company that had adopted a Clos fabric network to flexibly scale out their networking equipment to support our Hadoop infrastructure,” said Kenya Murakoshi,  Senior Manager, Site Operations Division, System Management Group at Yahoo Japan Corporation.  “Apstra streamlines Yahoo Japan Corporation’s network design, build and operations processes by abstracting the network configuration with a concept of intent, which was a new concept to us, but the benefits were clear. Apstra Intent-Based Analytics allow us to specify how we want our network to operate. We tell AOS what we want to happen vs. doing it ourselves. Apstra’s telemetry then provides deep insight into the state of our network and alerts us if there are risks like cabling or traffic imbalance issues that need to be addressed. Deployment of AOS at Yahoo Japan Corporation went smoothly without major issues.”

“Apstra offers the only multi-vendor offering that matched Yahoo Japan Corporation’s needs as we deploy network equipment from different vendors including white box products. We were impressed by the Apstra team’s passion to solve our problems. In addition, AOS allows Yahoo Japan Corporation to quickly automate, streamline, and replicate the building and operation of our Clos network fabric. We appreciate Apstra’s unmatched feature velocity, which is a huge differentiator, and we look forward to their growing feature set. We foresee a VXLAN/EVPN use case in the future with AOS, plan to replace more legacy 3 tier networks with Clos, and believe that Yahoo Japan Corporation will see even greater value from AOS in the future as we scale out our network. There is no real alternative to AOS in the market,” added Kenya Murakoshi.

“Yahoo Japan Corporation had a clear vision for the transformative impact of a fully automated multi-vendor Clos data center network operational model. They recognized that AOS and our declarative “intent-based” approach were the perfect fit to operate the company’s large Hadoop deployment, Clos environment, and the company’s requirement to increase east-west traffic,” said Mansour Karam, CEO and founder of Apstra. “Yahoo Japan Corporation has been a great partner to Apstra, validating our feature velocity and hardware inclusive approach. We look forward to partnering with them as they scale-out their full network infrastructure with multiple hardware vendors of their choice, including established vendors and open alternatives.”

Key requirements for Yahoo Japan Corporation included:

?       Operate and manage the large Clos fabric network that supports their Hadoop clusters with limited human resources (90 racks X 16 servers and thousands of nodes)

?       Rapidly deploy and replicate cross-platform telemetry services

?       Develop a multi-vendor network that is hardware-inclusive

?       Feature velocity to support Yahoo Japan Corporation’s constant flow of new Big Data offerings, including the automation of Day 2 maintenance workflows and powerful visibility to reduce Mean Time to Resolution (MTTR)

Key benefits reported by Yahoo Japan Corporation after Apstra AOS:

?       Ability to deploy multi-vendor network

?       Increased reliability (faster MTTR) with Apstra Intent-Based Analytics™

?       Reduction in design time (resource pooling) 

?       Reduction in deployment time through the ability with AOS to “click to deploy” multiple PODs from the same design in minutes, and with no human intervention

?       Reduction in management/complexity of the system, resulting in massive simplification of operations

?       Reduction in OpEx

?       Continuous Validation with cross-platform telemetry post deployment

?       Increased feature velocity including future features that could not be built in-house - including Day 2 maintenance operations across vendors, and intent-based analytics

?       Powerful Graph Live Queries

?       Excellent GUI providing a single pane view of the intent vs. network state

Big data and Hadoop are core technologies and strategic initiatives at Yahoo Japan Corporation which analyzes vast amounts of data obtained from its portal to constantly improve and develop new innovative Internet and ecommerce offerings. The company’s multi-vendor Clos network infrastructure connects hundreds of servers supporting thousands of nodes dedicated to supporting Hadoop across the enterprise, provides better network performance, and enables decommissioning many Hadoop nodes at once. A scale-out architecture and ability to easily, quickly, and affordably add capacity and reduce operational complexity to support the Hadoop network was critical. 

Multi-Vendor is Key

Yahoo Japan Corporation selected a switch vendor that also offered a proprietary management automation solution but the company also planned to deploy open, disaggregated hardware alternatives. Yahoo Japan Corporation is a big data-driven company experiencing a tremendous increase in network traffic and data. With the scale out traffic and growth in systems, the company was challenged to hire fast enough and recognized that automation was critical. The company has a large Hadoop deployment as a core technology and found that AOS is a perfect fit for the Clos environment and the company’s requirement to increase east-west traffic.