Рет қаралды 370
Sean Hafeez and Carly Stoughton discuss the causes of data center meltdowns and nightmares. No question there are data center meltdowns and disasters such as how does a network engineer handle multiple network switch failures, massive data center outages, large outages of applications, data center power outages, and other data center nightmares? The lesson taught by Google's data center and facebook's data center is not to build a snowflake. Conventional wisdom is that modern data centers should standardize on an L3 Clos, shouldn't use bare metal if you don't have to, use VMs or containers, every pod, and data center should look and act the same and applications should move around between them. If you can agree to build data centers the same then you need a way to stitch them together at a management level so you can specify the intent of how you want them to run and then you can avoid these nightmare data center issues. Apstra Intent-Based Networking for the Data Center enables non-web scale companies to run their data centers like the webscale companies like Google and Facebook. Apstra has built a commercial version of the tools and operating procedures that Google and Facebook have built over time and at huge cost to them.