[Bismark-devel] bismark kickoff meeting 8AM EST FRIDAY (today)

[Bismark-devel] bismark kickoff meeting 8AM EST FRIDAY (today)

[Dave Taht]

It strikes me that now would be a good time for those that have been 
participating silently/indirectly (as part of the debloat-testing and 
uberwrt projects) to meet up with those pulling together the next phases 
of the recently announced bismark project [1][2]

While this is terribly early in the US, and terribly late to call a 
meeting, perhaps those in europe/australia that have an interest would 
be able to attend via skype or sip as well as on the #bismark irc channel.

With the permission of the participants the meeting can be recorded also.

My own portion of the agenda is short, in that I want to get a grip on 
how we can de-"heisenbug" the measurements that bismark based routers 
will be taking, and figure out the best paths for the additional 
packages and patches to (mostly) make it into openwrt.

originally I felt that we'd "cook" these in uberwrt, but...

I feel some schedule pressure in that the original measurements of SFB 
taken by dan siemons were not-so-good and as that code is in the 
mainline (rc2) linux kernel now, and not well tested, it would be very 
good to get something in place to actually test it before it makes it 
into a major release.

Similarly ECN+pfifo has been fixed in 2.6.39-rc1, but needs to be 
backported. There are other debloat-testing patches out there... and a 
raft of bandwidth testing packages[3] as well as perhaps a latency under 
load test, httpping, etc...

[1] http://www.gatech.edu/newsroom/release.html?nid=65059
[2] http://www.bufferbloat.net/projects/bismark
[3] http://www.bufferbloat.net/projects/bismark/issues

For those new to this sub-project, the email list archives and wiki 
contain what information as exists today.