Who says there's no SDN scalability?
On the Network Heresey blog, SDN experts Teemu Koponen, Andrew Lambeth, Rajiv Ramanathan and Martin Casado, team up to tackle the common assertion that SDN is not scalable.
In the post, the team of experts present data taken over a two-year period from a production network virtualization environment that uses software-defined networking (SDN). The production environment, which is linked to in the post, was able to scale and still provide consistency in performance. The authors did find, however, that this type of flexibility in addressing scale meant that the protocol used between controllers was no longer standard. Therefore they emphasized the importance of standard interfaces around SDN applications and between controllers.
Read the full post on Network Heresy to discover more insight into the scaling properties of an application built using a SDN approach.
End-to-end QoS as an SDN use case
On the Keeping It Classless blog, Data Center Network Architect Matt Oswalt described how SDN can solve network Quality of Service (QoS) challenges. Silos exist between teams within an organization with a lack of communication, Oswalt wrote, yet this leads to fragmented QoS policies.
Oswalt argued that SDN "can and should take hold where the human element is the weakest link." SDN provides consistency where humans can't. Oswalt provided charts to show how SDN can centrally manage interaction with each node on the network, solving the problem of a lack of consistency and poor QoS.
View more of Oswalt's thinking behind QoS as an SDN use case.
Building racecars: Debating the practicality of OpenFlow
On the IP Space blog, author Ivan Pepelnjak took a hard look at the use of OpenFlow, writing that although it's appealing, what most OpenFlow "zealots" fail to mention is that only large companies -- like Google or McLaren -- are able to take on the task of becoming fluent in real-time, event-driven programming.
Pepelnjak compared using OpenFlow -- at this point -- to attempting to build a racecar. He advised companies to not ignore OpenFlow, but rather focus on more "down-to-earth" technologies.
Read more of Pepelnjak's argument about waiting on OpenFlow until there’s a solid controller.
Behold the benefits of telco SDN
On SDN Central, Dimitris Mavrakis, principal analyst with London-based Informa Telecoms & Media, broke down how SDN will affect the telecommunications industry. OpenFlow and OpenStack networking offer similar benefits to both IT and telecoms service providers, he wrote. In the long term, he said, SDN will allow for new service architectures and revenue opportunities.
Mavrakis said the benefits of telco SDN increase "towards the edge of the network" and are at their highest when facing the end user. Despite the failures of previous network technologies, telco SDN is arriving in the market at the right time -- when the telecoms and IT worlds are coming together.
Check out more of Mavrakis' view on how SDN is changing the telecom industry for the better.
Competing in a commodity world
On the Plexxi blog, author Michael Bushong challenged the concern that commoditized hardware will stifle profits.
Bushong predicted that commoditization will force companies to rethink the way they compete, which will be good from a customer perspective. As competition increases, due in part to SDN and the rise of start-ups in the switching space, there will be a divergence in the market. Vendors will come to realize that customer allegiance is to their business and service as opposed to a particular technology. That's a win for customers.
Read more of Bushong's thoughts on how to compete in a commodity networking world.