Re: [CUWiN] HSLS
David Young <[email protected]> Fri, 5 Nov 2004 21:38:17 -0600
| Newsgroups | gmane.network.wireless.cu-wireless |
|---|---|
| Message-ID | <[email protected]> |
On Sat, Nov 06, 2004 at 10:50:53AM +0800, Jong Hian Zin wrote: > Hi David, > > What is the status of the current HSLS/ETX in the CVS? Probably you > can tell us what are the things you're going to add/fix before the > roll out. Sure. Here are some outstanding tasks, just off the top of my head: The HSLS implementation currently uses IPv6 transport to route IPv6, only. Before the roll-out, we will make small changes to support IPv4. A developer is already working on that. I don't jitter Hello/LSU transmissions. I think it will be important to do that to avoid losing a lot of packets to collisions. Probably before roll-out. All the configuration is on the command line, today. We will add a configuration file. We haven't defined an HSLS MIB. I'm producing add/delete-route commands from the SPF results, now. Our Zebra interface is finished. Today I was injecting some routes into the RIB / FIB via Zebra. So we're probably about two weeks away from being able to run some trials on an HSLS/ETX-routed IPv6 network. Before roll-out, the daemon will be ready to operate efficiently on semi-broadcast (wireless) and point-to-point networks. (Someday I will optimize operation on wired ethernets using techniques borrowed from IS-IS, but I can postpone that until long after the roll-out.) There are 16 open PRs for HSLS in our internal bug-tracker. We anticipate finding and fixing many bugs between now and our mid-winter benchmark. Just for example, there is a conceptually simple "bootstrap" process for the link-state database that was a tad tricky to program; I will be testing much more carefully in coming weeks. I will probably find some bugs. There is trivial logic for purging old link-states and peers that remains to be programmed. Dave > > On Fri, 5 Nov 2004 18:12:12 -0600, David Young <[email protected]> wrote: > > On Wed, Nov 03, 2004 at 04:47:02PM -0500, Gopaul, Richard (Civ, ARL/CISD) wrote: > > > Hi, > > > > > > In your release notes it says: > > > > > > "The HSLS routing protocol and ETX route metric are not currently used. > > > " > > > What routing protocol and route metric do you currently use? > > > > We use OSPF with a hopcount metric. > > > > We are rolling out a larger testbed ahead of completing HSLS/ETX, and > > OSPF/hopcount performs about as badly as we expected. So we are looking > > forward to rolling out a reasonably bug-free HSLS/ETX implementation > > this winter. > > > > We can roll out sooner with your help! If you're interested in > > pitching-in on the development, the sources are available through both > > CVS and Subversion revision control systems: > > > > Browse CVS: http://cvs.sourceforge.net/viewcvs.py/wireless > > > > CVS checkout: > > > > cvs -d:pserver:anonymous at cvs.sourceforge.net:/cvsroot/wireless login > > cvs -z3 -d:pserver:anonymous at cvs.sourceforge.net:/cvsroot/wireless co cuw-trunk > > > > Browse Subversion: http://svn.cuwireless.net/svn/cuw > > > > Subversion checkout: > > > > svn checkout http://svn.cuwireless.net/svn/cuw/trunk > > > > Dave > > > > -- > > David Young OJC Technologies > > [email protected] Urbana, IL * (217) 278-3933 > > _______________________________________________ > > CU-Wireless mailing list > > [email protected] > > http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless > > Project Page: http://cuwireless.ucimc.org > > > > > -- > Jong Hian Zin > _______________________________________________ > CU-Wireless mailing list > [email protected] > http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless > Project Page: http://cuwireless.ucimc.org -- David Young OJC Technologies [email protected] Urbana, IL * (217) 278-3933 _______________________________________________ CU-Wireless mailing list [email protected] http://lists.cu.groogroo.com/cgi-bin/listinfo/cu-wireless Project Page: http://cuwireless.ucimc.org