We rounded up the best of this week’s SDN blogs, which explored Prescriptive Topology Manager, SDN controller clusters, how IT infrastructure differs, and everything there is to know about OVSDB.
Learning about Prescriptive Topology Manager
On his personal blog site, Packet Pushers co-founder Ethan Banks delves deeper into Prescriptive Topology Manager or PTM. Banks first heard of PTM when reading about the Cumulus Linux 2.2 release. Put simply, PTM allows for the simplification of operations workflow, while reducing risk by prescribing how a network is supposed to operate. In other words, PTM is a cabling verification tool that makes sure that the physical cabling done on a network matches a predefined plan.
Banks expand on PTM and what it's used for while also exploring Link Layer Discovery Protocol (LLDP) and DOT. DOT is a plan text graph description language that can be used by both humans and computers. Banks includes illustrations of what a DOT graph may look like. He also embeds a YouTube video to better help explain PTM and DOT in its entirety.
Take a look at Banks' post looking at PTM, LLDP and DOT.
Are SDN controller clusters a single-failure domain?
On his IpSpace blog, Ivan Pepelnjak writes that SDN controllers -- whether OpenFlow-based or not -- are basically a single failure domain. A controller cluster will protect against hardware failures, but it won’t protect against software failures. Pepelnjak explains how an active or standby controller cluster works, adding that there’s added complexity around the clustering of software.
Pepelnjak suggests distributed systems versus centralized ones in order to address these issues, as well as loosely coupled systems like Border Gateway Protocol SDN compared to tightly coupled ones, such as OpenFlow control.
Learn more about SDN controller clusters in Pepelnjak's post.
The similarities of IT infrastructure
Greg Ferro talked about IT infrastructure and how little it differs from market to market on his Ethereal Mind blog. As a freelance engineer for the past 15 years, Ferro has recognized that infrastructure is strikingly similar across enterprises. In fact, most companies face the same problems and can use the same solutions. And as a result, Ferro writes, it's unfair for vendors to sell platforms that are customized or unique to a particular customer.
Check out Ferro's post in its entirety, explaining why IT infrastructure doesn't differ from market to market.
Everything you need to know about OVSDB
In his SDN protocols series, Matt Oswalt does a fourth installment looking at the Open vSwitch Database Management Protocol, or OVSDB, on his Keeping it Classless blog site. Oswalt covers what the protocol can and can't do -- and he writes that, if you're new to OVSDB, it's best to think of it as any other configuration API, like NETCONF or even proprietary vendor configuration APIs.
Oswalt covers how OVSDB is different than OpenFlow, exploring the advantages and drawbacks of control and management.
Take a look at Oswalt's full post to learn everything you need to know about OVSDB.