Listen, Subscribe & Follow:
This Weekly Show episode was recorded live at IETF 96 in Berlin in July 2016. Greg Ferro and several guests discuss the state of routing protocols such as BGP, and explore different approaches to routing, like Facebook’s Open/R initiative.
They also debate issues around telemetry, network disaggregation, and whether enterprises should participate in the IETF to influence vendor product development.
Joining Greg in the conversation are Jeff Tantsura, Director Product Management, Head of Technology Strategy Routing at Ericsson, and the chair of the Routing Area working group at the IETF; and Dr. Tony Przygienda, a distinguished engineer at Juniper Networks and chair of the Bit Indexed Explicit Replication (bier) working group at the IETF.
Also weighing in is Russ White, network architect, author, and blogger.
If you like routing and are ready to get nerdy, this one’s for you!
Sponsor: Paessler
Check out PRTG Network Monitor from Paessler. PRTG simplifies the daily work of IT professionals by offering them a smart network monitoring solution to ensure their business-critical IT infrastructure runs smoothly with 100% uptime. Get an exclusive free network consultation when you download the 30-day trial of PRTG network monitor. To participate, simply visit www.paessler.com/packetpushers.
Sponsor: ThousandEyes
ThousandEyes delivers visibility into every network your organization relies on—from your data center to the cloud. You can quickly and precisely pinpoint the root cause of network issues— wherever they occur — and then share your insights with your vendors and customers. Sign up for an account at ThousandEyes.com/packetpushers to monitor 3 locations for free, and choose a ThousandEyes t-shirt.
Share this episode
The “spark” protocol being used in Open/R is not Apache Spark, it’s a simple and lightweight hello protocol used to detect neighbors and establish adjacencies, very similar to BFD. Just wanted to make the correction to avoid future confusion =)
Pablo, thanks for correction. In this case I was inferring wrongly. I was working off what was published and available then and since the emphasis was on unequal cost paths and ZeroMQ streaming into KV store my natural reaction was (given probably also what I was looking into those days) that we’re talking about Spark to heft the seemingly large amount of computations (as I stated, a mind-mumblingly great piece of technology). In the podcast I mildly nudged the delay issue on such a solution though and I was also still thinking how you manage to have the data on the cluster replicated (without using something like Kafka or extensively writing through into HDFS).
I re-read the newest version of what you published and things are clearer and to be frank, look more mainstream and less revolutionary 😉
To my own defense, I made it sound in the podcast as well if ZeroMQ would do framing, Thrift was never mentioned and should have but as you saw we were really all over the place having as much fun as exchanging information and without Jeff nudging us back we would probably have circled the moon.
Hope otherwise it was all worth your time to listen into …
Is there a download of open/r in some form available for my personal interest. Googl’ing Facebook Open/R ends you up in either R (data mining) mostly 😉