A new SDN competitor for VMware and Cisco?
On the Gartner blog site, analyst Andrew Lerner took a look at the Cisco ACI versus VMware debate -- and he questions if a third player could be joining the game shortly. Recent industry news sparked Lerner's interest. Brocade's announcement of its flagship OpenDaylight-based platform, along with HP's announcement of its SDN app store, are signs that the race to SDN is not for two competitors alone. Lerner also includes a link to a Gartner white paper that examines how HP's app store needs an open platform to really make an impact.
Take a look at Lerner's post in its entirety to learn more about a possible third-party SDN competitor.
OpenDaylight and network policy
OpenDaylight's Helium release is sparking conversations regarding network policy, writes The Networking Nerd's Tom Hollingsworth. Helium contains group-based policies that enable network devices to talk to one another. That means Helium users don't necessarily need to be well-versed in code.
Helium is a sign that OpenDaylight will remain a policy-driven controller infrastructure. In Hollingsworth's opinion, the OpenDaylight controller will be the front-running controller for SDN, which means it will need to set a standard for network interaction -- or policy. Policy will play a crucial role in the future of network interfaces, which is what makes OpenDaylight so important.
Read more about OpenDaylight's impact on network policy in Hollingsworth's full post.
Applications and availability in the network
On the Plexxi company blog, author Mike Bushong looks at high availability for applications keeping in mind device uptime and design. App availability and experience tend to be the barometer for measuring infrastructure success. Meanwhile, it's biased to gauge performance solely by device uptime, he writes. Bushong added that the industry views availability in not the best way, by designing for correctness versus architecting for resilience.
Bushong also touches on application resiliency and how SDN provides a useful architectural framework for fast failover. In the end, Bushong concludes that architecting and implementing the "perfect” device in a perfect network is possible, but it does mean a change in design criteria.
Check out Bushong's applications and availability post in its entirety.