1. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100: developing network-aware operating systems New (9/01) DOE-funded (Office of Science) project ($1M/yr, 3 yrs) Principal investigators –Matt Mathis, PSC ( mathis@psc.edu ) –Brian Tierney, LBNL ( bltierney@lbl.gov ) –Tom Dunigan, ORNL ( thd@ornl.gov ) Objective: –measure and understand end-to-end network and application performance –tune network applications (grid and bulk transfer) Components –active network probes and passive sensors (leverage Web100) –network metrics data base –tuning daemon (WAD) to tune network flows based on network metrics www.net100.org

2. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100: applied Web100 Web100 –Linux 2.4 kernel mods –100+ TCP variables per flow Net100 –Add Web100 to iperf/ttcp –Monitoring/tuning daemon Java applet bandwidth/client tester –fake WWW server provides html and applet –applet connects to bwserver 3 sockets (control, bwin, bwout) server reports Web100 variables to applet (window sizes, losses, RTT) – Try it http://firebird.ccs.ornl.gov:7123

3. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100 network measurement Active measurement –Net100 probes at LBL, ORNL, NCAR, PSC, NERSC –scheduled set of path probes ( iperf with Web100 mods, traceroute, pipechar ) –local and centralized database ( netlogger ) –interface to other probers (NIMI, surveyor, Pinger, ?) Passive measurement –Web100 daemon records TCP info on designated flows Web100 data collected when flow terminates Web100 TCP info: losses, timeouts, reordering, cwnd, ssthresh, RTT,… use netlogger to report to central data base –other passive sensors (SNMP data, LBL’s tcpdump monitor, ?) Query tools –for dynamic application tuning –for network engineering and statistical studies

4. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100: tuning Work-around Daemon (WAD) Version 0 –use network performance data to tune flows –tune unknowing sender/receiver –config file with “tuning info” ? –Based on Web100/Linux 2.4 To be done –“applying” measurement info –adding more knobs to kernel –tune on non-Linux OS Related work –Feng’s Dynamic Right Sizing –Linux 2.4 auto-tuning/caching –Mathis TCP buffer tunning

5. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory TCP losses TCP is lossy by design Changing : bandwidths –9.6 Kbs… 1.5 Mbs..45 …100…1000…? Mbs Unchanging : –speed of light (RTT) –MTU (still 1500 bytes) –TCP congestion avoidance recovery after a loss can be very slow on today’s high delay/bandwidth links –proportional to MSS/RTT 2 Linear recovery at 0.5 Mb/s! Instantaneous bandwidth Average bandwidth Early startup losses

6. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100 tuning Avoid losses –use “optimal” buffer sizes determined from network measurements –ECN capable routers/hosts –reduce bursts (TCP Vegas, ?) Faster recovery –bigger MSS (jumbo frames) –speculative recovery (D-SACK) –modified congestion avoidance? Autotune (WAD variables) –Buffer sizes –Dupthresh (reordering resilience) –Del ACK, Nagle –AIMD –Virtual MSS –initial window, ssthresh non-TCP solutions (rate-based, ?) (tests with TCP-over-UDP, atou, NERSC to ORNL)

7. UT-BATTELLE U.S. Department of Energy Oak Ridge National Laboratory Net100 status Completed –network probes at ORNL, PSC, NCAR, LBL, NERSC –preliminary schema for network data –initial Web100 sensor daemon and tuning daemon In progress –TCP tuning extensions to Linux/Web100 kernel –analysis of TCP tuning options –deriving tuning info from network measurements Future –interactions with other network measurement sources –multipath/parallel path selection/tuning www.net100.org