School server networks fail to perform fairly or effectively because the loss-based congestion control methods in TCP networking are not a good match for the wireless environment.
Wireless access points or devices with too much buffering memory make the problem worse. A case of more not being good. Former OLPCer Jim Gettys warned us about bufferbloat years ago. I'd like to hear from anyone deploying the new BBR congestion-base controls, which should be available in 4.9 kernels. It is a server-side feature, so there's no great need for it in laptops that operate as clients. References: http://queue.acm.org/detail.cfm?id=3022184 https://lwn.net/Articles/701165/ http://queue.acm.org/detail.cfm?id=2071893 -- James Cameron http://quozl.netrek.org/ _______________________________________________ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
