As I explained in a previous blog post, most leaf-and-spine best-practices (as in: what to do if you have no clue) use BGP as the IGP routing protocol (regardless of whether it’s needed) with the same AS number shared across all spine switches to implement valley-free routing.
This design has an interesting consequence: when a link between a leaf and a spine switch fails, they can no longer communicate.
For example, when the link between L1 and C1 in the following diagram fails, there’s no connectivity between L1 and C1 as there’s no valley-free path between them.
Read more ...Who’s ready for another Cumulus content roundup? This month we kept busy with step-by-step guides, tips and tricks, and even kicked off a new series, Tales from the field! Grab a seat and settle in to learn more about everything from EVPN to ZTP, and how they benefit you. Happy reading!
From Cumulus Networks:
We stand with open source: We’ve come a long way from where we began as a company. Find out why we’re passionate about open source networking, what we’ve brought to the table in the last couple of years, and our goals for the future.
5 tips for transitioning to 100G networking: Are you ready to increase your networking speed and pack more computing into the same space? You’re going to want to read these 5 tips on making the transition to 100G networking as smooth as possible.
EVPN behind the curtains: Confused about EVPN? Read this blog post for a step-by-step understanding of EVPN and what it means to you.
Tales from the field: Best practices for initial provisioning (Part 1): Let’s dive into ZTP! Read part one of “Tales from the field,” a series, and find out how to use automation tools to Continue reading
The company provides DevOps automation tools focused on AWS and Azure public cloud environments but expects to announce some work with Google Cloud later this year.
After a hot, dry June day in Richmond I boarded an empty train. It being just myself and a couple opposite we struck up a conversation, awaiting the train’s departure. Over the next twenty minutes, as the train slowly worked through a procession of sun-baked stations, I discovered that my companions worked in IT, but in a much earlier time. The white-haired gent confessed to be a C programmer back in the 60s. We talked about software, home lab specs, and genial conversation flowed from mutual interests. However, when I idly announced ‘Microsoft have really changed, Windows has greatly improved’ the C gent gave my comments short shrift and cut them down with:
‘They can’t change, they’re too big.’
No matter, it was a serendipitous meeting that enriched a dull train journey.
That encounter stuck with me, especially the rejection of Microsoft, and I was reminded of it last week in a brief exchange with David Bombal on twitter. The #netdevops guys put together a video about using Windows as a developer environment and David mentioned some negative comments they had received Continue reading
As one of the first hyperscalers in the world, Yahoo, part of the Oath conglomerate that also includes AOL and that is owned by telecom giant Verizon, knows a thing or two about running applications at extreme scale. …
How OpenStack Lassoes Yahoo’s 4 Million Server Cores was written by Timothy Prickett Morgan at .
Both cloud-native and traditional organizations must implement a security practice that offers automation, abstraction, flexibility, visibility, and is application-aware.
RingCentral deployed ThousandEyes monitoring platform last quarter to add an additional layer of visibility over its global network and services.
“The path forward is an open platform,” said Adva’s Prayson Pate, adding that an open platform will meet the full-potential of an agile and flexible SD-WAN.
VMware ranked No. 2 in the HCI software vendor list while Dell remained the market leader for branded hyperconverged products.
Vodafone said Juniper’s Contrail Service Orchestrator is a key part of the platform because it provides customers with a self-service portal.
While deep learning models might not be able to simulate large-scale physical phenomena in the same way purpose-built supercomputers and their application stacks do, there is more research emerging that shows how traditional HPC simulations can be augmented, if not replaced in some parts, by neural networks. …
Deep Learning Infiltrating HPC Physics Domains was written by Nicole Hemsoth at .