160 likes | 289 Vues
EU DataGrid - Network Monitoring Richard Hughes-Jones, University of Manchester Thanks to a wireless connected Titanium in the garden@Houten. SURFnet. Manc. UvA. SuperJANET4. CERN. European Topology: NRNs & Geant. Gigabit Throughput on the Production WAN. Manc - RAL 570 Mbit/s
E N D
EU DataGrid - Network Monitoring Richard Hughes-Jones, University of Manchester Thanks to a wireless connected Titanium in the garden@Houten DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
SURFnet Manc UvA SuperJANET4 CERN European Topology: NRNs & Geant DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Gigabit Throughput on the Production WAN • Manc - RAL 570 Mbit/s • 91% of the 622 Mbit access link between SuperJANET4 and RAL • 1472 bytes propagation ~21s • Manc-UvA (SARA) 750 Mbit/s • SJANET4 + Geant + SURFnet • Manc – CERN 460 Mbit/s • CERN PC had a 32 bit PCI bus DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Window Mbytes for UDP BW 750 Mbit/s 460 Mbit/s Window for BW 1 Gbit/s Gigabit TCP on the Production WAN (01) • Throughput vs TCP buffer size • TCP window sizes in Mbytes calculated from RTT*bandwidth DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Gigabit TCP on the Production WAN Man-CERN • Throughput vs n-streams • Default buffer size slope = ~25 Mbit/s/stream up to 9 streams then 15 Mbit/s/stream • With larger buffers rate of increase per stream is larger • Plateaus at about 7 streams giving a total throughput of ~400 Mbit/s DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
UDP Throughput: SLAC - Man • SLAC – Manc 470 Mbit/s • 75% of the 622 Mbit access link • SuperJANET4 peers with ESnet at 622Mbit in NY DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Gigabit TCP on throughput Man-SLAC • Throughput vs n-streams • Much less than for European links • Buffer required: rtt*BW (622Mbit) = ~14 Mbytes • With larger buffers > default, rate of increase per stream is ~ 5.4 Mbit/s/stream • No Plateau DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
UDPmon & Ripe MAN-SARA from 25 May 02 UDPmon Loss Throughput Mbit/s MAN – UvA ~750Mbit/s RiPE 1 way UvA RAL ms RiPE 1 way RAL UvA ms DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
iperf Pinger & UDPmon UK-Lyon From 25 May 02 – 04 Jun 02 UDPmon Loss Throughput ~140 Mbit/s Man Lyon Ping Man Lyon 40 ms rtt Iperf TCP Throughput ~ 47 Mbit/s Man Lyon 1048576 byte buffer DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Work on End Systems: PCI: SysKonnect SK-9843 • Motherboard: SuperMicro 370DLE Chipset: ServerWorks III LE Chipset • CPU: PIII 800 MHz PCI:64 bit 66 MHz • RedHat 7.1 Kernel 2.4.14 • SK301 • 1400 bytes sent • Wait 20 us • Sk303 • 1400 bytes sent • Wait 10 us • Frames are back-to-back • Can drive at line speed • Cannot go any faster ! Gig Eth frames back to back DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
PCI: Intel Pro/1000 • Motherboard: SuperMicro 370DLE Chipset:: ServerWorks III LE Chipset • CPU: PIII 800 MHz PCI:64 bit 66 MHz • RedHat 7.1 Kernel 2.4.14 • IT66M212 • 1400 bytes sent • Wait 11 us • ~4.7us on send PCI bus • PCI bus ~45% occupancy • ~ 3.25 us on PCI for data recv • IT66M212 • 1400 bytes sent • Wait 11 us • Packets lost • Action of pause packet? DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
UDP Throughput: Intel Pro/1000on B2B P4DP6 • Motherboard: SuperMicro P4DP6 Chipset: Intel E7500 (Plumas) • CPU: Dual Xeon Prestonia (2cpu/die) 2.2 GHz Slot 4: PCI, 64 bit, 66 MHz • RedHat 7.2 Kernel 2.4.14 • Max throughput 950Mbit/s • Some throughput drop for packets >1000 bytes • Loss NIC dependent • Loss not due to user Kernel moves • Traced to discards in the receiving IP layer ??? DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
UDPmon UDPmon UDP UDP IP IP Eth drv Eth drv Gig Switch HW HW • No loss at switch • But Pause packet seen to sender Packet Loss: Where? • Intel Pro 1000 on 370DLE • 1472 byte packets • Expected loss in transmitter ! N Lost N Gen IpDiscards N Transmit N Received DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Interrupt Coalescence: Latency • Intel Pro 1000 on 370DLE 800 MHz CPU DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Interrupt Coalescence: Latency • Intel Pro 1000 on 370DLE 800 MHz CPU DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester
Interrupt Coalescence: Throughput • Intel Pro 1000 on 370DLE DataTAG/WP7 Amsterdam June 2002 R. Hughes-Jones Manchester