Friday, September 23, 2011

The Cloudcast - Eps.21 - "The Network Stack of OpenStack" + Show Notes

Date: September 23, 2011

[APOLOGIES for the ECHO on Brian's microphone. These were not detected until after recording]

Q: Cisco announced a formal involvement with the OpenStack project in the spring of this year. Can you give us some background on why Cisco is involved and what’s been happening since that announcement?

Q: The big news this week was that the “Diablo” version of OpenStack was released. From some of the news stories and early discussions, it feels like maybe this is the version that could get OpenStack over the hump of being ready for a broader audience than earlier adopters. What is your take on the significance of this release and what does it have you excited?

Q: For Cisco, Diablo is a big deal because it’s the first official release that includes code that Cisco contributed to the OpenStack community and the first time “network” was it’s own project. Why is that a big deal for both Cisco and the community?

Q: For a while networking wasn’t a topic of alot of discussion, but as of late it seems to have gotten a ton of visibility because of emerging architectures (L2 vs. L3), competing standards for mobility (VXLAN vs. NVGRE) and things like OpenFlow. Do you expect the “network” area of OpenStack to gain a lot of community involvement, or is it still too early?

Q: When people traditionally think of Cisco, the phrase “open-source” doesn’t usually come to mind. But over the last couple years they have been assembling a number of people that understand the DNA needed to work with open-source communities. How are things changing at Cisco?

[NOTE: One error I made in this video is that there is support for block-based storage (iSCSI) via the "Nova" compute project. So OpenStack is not exclusive to Object-based storage, but the "Swift" project is Object-based.]