1
|
- Warren Matthews, Les Cottrell, Paola Grosso
- SLAC/SCS
- Prepared for BaBar Data Distribution meeting, SLAC February 26 2003
- www.slac.stanford.edu/grp/scs/net/talk/babar-dd-feb03.htm
|
2
|
- SLAC Connections to the Internet
- Usage
- Performance
|
3
|
|
4
|
|
5
|
|
6
|
|
7
|
- PingER – low impact RTT, loss etc
by ping
- IEPM-BW – higher impact throughput
- Netflow – passive, characterization of traffic
|
8
|
|
9
|
|
10
|
|
11
|
|
12
|
|
13
|
|
14
|
|
15
|
- 9/01 105 Mbps 30 Streams: SLAC-IN2P3;
102 Mbps 1 Stream CIT-CERN
- 11/5/01 125 Mbps in One Stream (modified
kernel): CIT-CERN
- 1/09/02 190 Mbps for One stream shared on 2
155 Mbps links
- 3/11/02 120 Mbps Disk-to-Disk with One Stream
on 155 Mbps
link (Chicago-CERN)
- 5/20/02 450-600 Mbps SLAC-Manchester on OC12
with ~100 Streams
- 6/1/02 290 Mbps Chicago-CERN One Stream on
OC12 (mod. Kernel)
- 9/02 850, 1350, 1900 Mbps Chicago-CERN 1,2,3 GbE Streams,
2.5G Link
- 11-12/02 FAST TCP: 940 Mbps in 1 Stream
California-CERN;
9.4 Gbps in 10 Flows California-Chicago
|
16
|
- More bandwidth, less cross traffic
- Tune TCP/IP parameters
- New TCP stacks
- Tom Kelly's Scalable TCP,
- Caltech's FAST TCP
- Sally Floyd's High-Speed TC
- Can achieve ~1Gbps trans-Atlantic
- SLAC and collaborators hold the world record (Nov-02)
|
17
|
- Linux memory leak
- Linux TCP configuration caching
- What is the window size actually used/reported
- 32 bit counters in iperf and routers wrap, need latest releases with
64bit counters
- Effects of number of packets in queue for NIC (txqueuelen)
- Routers do not pass jumbos
- Some applications have problems with big windows
- Firewalls
|
18
|
- A ubiquitous network monitoring infrastructure
- Grid Middleware
- Guide Applications
- Troubleshooting
- Resource Allocation
- Set expectations
|
19
|
- Multiple high-speed connections
- Optimize routing
- Backup paths
- Well Engineered Networks
- Negligible packet loss, Good Round Trip Times
- Traffic doubled on ESnet every year since 1990
- Need continued upgrades to keep up with demand
- Transmitting TBytes/day today
- Room for improvement (bottlenecks at edges)
- Needs funding &
- Improved TCP or other protocols
- TCP works up to Gbits/s rates, but needs wizards today to make it
perform
|
20
|
- High-performance Monitoring
- www-iepm.slac.stanford.edu/bw/
- TCP stack evaluation and 10GE testbed
- www-iepm.slac.stanford.edu/monitoring/bulk/fast/
- Measurement reports
- www.slac.stanford.edu/comp/net/bandwidth-tests/antonia/html/slac_wan_bw_tests.html
- Land Speed Record
- www-iepm.slac.stanford.edu/lsr/
|
21
|
|