Hi Dan, Thank you for your e-mail. Yes. I agree that the Paxos commit is a good protocol to realize reliability.I am familiar with the HARC and Paxos commit since I worked together with Jon Maclaren and EnLIGHTened team for Japan-US interoperability experiment. I think Paxos is a little bit complicated for the basis of NSI architecture and protocol discussions. At this stage, it will be better to keep discussions on a simple protocol, and even 2PC might be too complicated. Thanks, Tomohiro 2009/12/3 Daniel S. Katz <dsk@ci.uchicago.edu>:
Hi,
It also might be worth looking at the first HARC paper (http://www.cct.lsu.edu/~maclaren/HARC/Papers/harc-gada-final.pdf), specifically figures 2 and 3, which show 2-phase commit and Paxos commit. Paxos commit is probably a better protocol for more general situations, though 2-phase commit works well when the "committer" is reliable.
Dan (wishing to have more time to participate in these discussions rather than just reading the emails some of the time)
On Dec 2, 2009, at 7:14 AM, Tomohiro Kudoh wrote:
Hi all,
here is an updated slide.
2009/12/2 Tomohiro Kudoh <t.kudoh@aist.go.jp>:
Hi all,
Here is a slide on 2-phase commit. Talk to you soon.
Tomohiro <2pc-updated.ppt>_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
-- Daniel S. Katz University of Chicago (773) 834-7186 (voice) (773) 834-3700 (fax) d.katz@ieee.org or dsk@ci.uchicago.edu http://www.ci.uchicago.edu/~dsk/