As part of a broader organisational restructure, data networking research at Swinburne University of Technology has moved from the Centre for Advanced Internet Architecture (CAIA) to the Internet For Things (I4T) Research Lab.

Although CAIA no longer exists, this website reflects CAIA's activities and outputs between March 2002 and February 2017, and is being maintained as a service to the broader data networking research community.


This page is part of the BART project.

Evaluating ASUS P4SGX-MX ATX Motherboards for the BART Project
CAIA Technical Report 031202B

Ana M. Pavlicic
Thuy Nguyen
December 2nd, 2003

Introduction

This technical report outlines the results obtained from testing a new BART project PC containing an ASUS P4SGX-MX ATX Motherboard. Several PCs are to be used in the BART project to simulate home users sending and receiving traffic across an ISP network. Tcpdump on the PC was used to capture packets and the time stamping of the packets investigated in this report. This technical report uses a similar setup and testing method as Mark Pozzobon's evaluation of CAIA laboratory computers.

The aim of the investigation was to find the optimal configuration for Ethernet interfaces so as to obtain the most accurate time stamp results. The optimal configuration involved manually setting the line speed on the interface. The addition of another NIC and its time stamp accuracy was also investigated. The optimal configuration for the off-board PCI Ethernet card involved manually configuring link speeds on both interfaces and connecting the unused interface to an active Ethernet link so that a carrier is present.

Equipment and Setup

The tests conducted involved two link speeds, 10Mbit/sec and 100Mbit/sec, with UDP packets 92 bytes + 4 bytes CRC long flooded to the PC under evaluation. Packet inter-arrival time set on SmartBits2000 was 125usec, 500usec, 1msec, 50msec and 100msec for each of the link speed tests. The inter-arrival time used by SmartBits2000 and the Tcpdump time stamps differ in the way they are measured. SmartBits2000 measures inter-arrival time as the time between the end of one packet and the start of the next packet. The Tcpdump -ttt option measures the time from the beginning of one packet to the start of the next packet. In this case, the size of the packet is included into the time interval. To learn more about SmartBits2000 inter-arrival time, tcpdump -ttt time stamps and time offsets see Mark Pozzobon's report. All time stamps should be the inter-arrival time of packets as sent by the SmartBits2000 plus approximatley 83usec (83.2usec) for the 10Mbit/sec line speed and 8.3usec for the 100Mbit/sec line speed.

Each inter-arrival time (125usec, 500usec etc.) was tested using a single SmartBits2000 card flooding packets to the PC's Ethernet port. Initially the interface configuration was allowed to automatically adjust to 10Mbit/sec or 100Mbit/sec traffic, then the line speed on the interface was explicity (statically) set at 10Mbit/sec or 100Mbit/sec (with SmartBits2000 also flooding at the explicit line speed). This was to investigate whether there is a difference when the interface is automatically detecting the line speed and when it is explicitly set to 10Mbit/sec or 100Mbit/sec. Figure 1 below illustrates test setup.


Figure 1: Hardware Setup


Tcpdump was run with the following command: "tcpdump -n -i interface -w filename" then packet time stamps placed into a separate file using the command: "tcdpump -n -ttt -r filename | awk -F " " '{print $1}' > newfilename" for analysis.

Test Set A: The first set of results were obtained using the following equipment:
  • Netcom System SmartBits2000 Multiport Performance Analysis System with one SX-7410 100Mb Ethernet TP card
  • Tcpdump running on the test PC with the following configuration:
    • Intel Celeron 2.4GHz (400) Single Processor
    • ASUS P4SGX-MX ATX Motherboard
    • Realtek RTL8100C 10/100 Mbps LAN on-board NIC (Ethernet interface sis0)
    • 256M PC2700 DDRAM Memory
  • Two CAT5 UTP Crossover cables
  • Separate PC running Netcom Systems SmartWindow v6.51 to control the packets being flooded by SmartBits2000
Test Set B: The second test set involved adding the following NIC:
  • Intel Pro100 Desktop NIC Adapter (Ethernet interface fxp0)

Results/Discussion

The following are the results obtained in the investigation by both test set A and B:

Test Set A: Interface sis0

The first step of the investigation was to test the in-built Ethernet card by allowing the interface to automatically detect the line speed as set on the SmartBits2000. This was followed by explicitly setting the correct line speed (either 10Mbit/sec or 100Mbit/sec) for each of the inter-arrival times tested. The sis0 interface had the following configurations for each inter-arrival time trial:

  • sis0 Automatic 10Mbps:
  • sis0: flags=8843 mtu 1500
    inet 136.186.229.238 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::20c:6eff:fed1:58f3%sis0 prefixlen 64 scopeid 0x1
    ether 00:0c:6e:d1:58:f3
    media: Ethernet autoselect (10baseT/UTP)
    status: active

  • sis0 Explicit 10Mbps:
  • sis0: flags=8843 mtu 1500
    inet 136.186.229.238 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::20c:6eff:fed1:58f3%sis0 prefixlen 64 scopeid 0x1
    ether 00:0c:6e:d1:58:f3
    media: Ethernet 10baseT/UTP
    status: active

  • sis0 Automatic 100Mbps:
  • sis0: flags=8843 mtu 1500
    inet 136.186.229.238 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::20c:6eff:fed1:58f3%sis0 prefixlen 64 scopeid 0x1
    ether 00:0c:6e:d1:58:f3
    media: Ethernet autoselect (100BaseTX)
    status: active

  • sis0 Explicit 100Mbps:
  • sis0: flags=8843 mtu 1500
    inet 136.186.229.238 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::20c:6eff:fed1:58f3%sis0 prefixlen 64 scopeid 0x1
    ether 00:0c:6e:d1:58:f3
    media: Ethernet 100baseTX
    status: active

The results obtained showed that explicitly setting the correct line speed produces the optimal time stamp accuracy:

125usec Packet Inter-Arrival Time
500usec Packet Inter-Arrival Time
1msec Packet Inter-Arrival Time
50msec Packet Inter-Arrival Time
100msec Packet Inter-Arrival Time

Test Set B: Interface fxp0 with sis0

In order to reduce interference from the unused sis0 and optimise time stamp performance on fxp0, the sis0 interface must be kept content with a carrier signal:

Interface Configuration for Optimal Time Stamp Performance On fxp0

The following set of results show that setting the interface line speed explicitly but NOT giving the unused sis0 interface a carrier signal does not improve time stamp accuracy. Although accuracy was improved using this practice when only the sis0 interface existed in the system, this is not true for the fxp0 interface. In the following results the sis0 interface was set to auto-detect the line speed and no carrier was present:

    sis0: flags=8842 mtu 1500
    inet 10.10.10.6 netmask 0xfffffffc broadcast 10.10.10.7
    ether 00:0c:6e:d1:58:f3
    media: Ethernet autoselect (10baseT/UTP)
    status: no carrier

while the fxp0 configuration changed as to mimic the configurations of the sis0 interface in Test Set A:
  • fxp0 Automatic 10Mbps:
  • fxp0: flags=8843 mtu 1500
    inet 136.186.229.237 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::202:b3ff:fee7:bc26%fxp0 prefixlen 64 scopeid 0x2
    ether 00:02:b3:e7:bc:26
    media: Ethernet autoselect (10baseT/UTP)
    status: active

  • fxp0 Explicit 10Mbps:
  • fxp0: flags=8843 mtu 1500
    inet 136.186.229.237 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::202:b3ff:fee7:bc26%fxp0 prefixlen 64 scopeid 0x2
    ether 00:02:b3:e7:bc:26
    media: Ethernet 10baseT/UTP
    status: active

  • fxp0 Automatic 100Mbps:
  • fxp0: flags=8843 mtu 1500
    inet 136.186.229.237 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::202:b3ff:fee7:bc26%fxp0 prefixlen 64 scopeid 0x2
    ether 00:02:b3:e7:bc:26
    media: Ethernet autoselect (100BaseTX)
    status: active

  • fxp0 Explicit 100Mbps:
  • fxp0: flags=8843 mtu 1500
    inet 136.186.229.237 netmask 0xffff0000 broadcast 136.186.255.255
    inet6 fe80::202:b3ff:fee7:bc26%fxp0 prefixlen 64 scopeid 0x2
    ether 00:02:b3:e7:bc:26
    media: Ethernet 100baseTX (100baseTX )
    status: active

Result obtained:

125usec Packet Inter-Arrival Time
500usec Packet Inter-Arrival Time
1msec Packet Inter-Arrival Time
50msec Packet Inter-Arrival Time
100msec Packet Inter-Arrival Time

Summary

As seen in this investigation, the initial set of results (Set A) showed that for both the 10Mbit/sec and 100Mbit/sec line speeds, explicitly setting the interface line speed resulted in the Tcpdump time stamps being more accurate than by letting the line speed be automatically determined. When allowing the line speed to be automatically detected by the interface, the maximum time stamp error was approximately 7msec away from the ideal time stamp value. By explicitly setting the line speed to be 10Mbit/sec or 100Mbit/sec this error was reduced to about only 15-50usec maximum. The investigation found that this was not true for the Intel Pro100 Desktop NIC, as specifying the line speed manually made no difference in reducing the large time stamp errors. For both cases where the line speed was configured to auto-negotiate and was explicitly set, the maximum time stamp error reached as high as 9msec. This was due to interference from the sis0 interface. By setting the sis0 interface line speed statically and giving it a carrier (plugging in a CAT5 crossover cable from the SmartBits2000), the time stamping accuracy improved to about 15-50usec maximum error.

Further investigation may include changing other computer components and testing a variety of NIC types. The results of this investigation could also be used to determine the type of hardware to be purchased for the ICE^3 project and how to configure and physically set up Ethernet interfaces so as to reduce the severity of time stamp errors.

Acknowledgements

Grenville Armitage, Director of Centre for Advanced Internet Architectures.





Swinburne Homepage Site Map Search Index
 

© Swinburne Copyright and disclaimer Privacy Feedback

Last Updated: Monday 5-Jan-2004 16:06:05 AEDT
URL:
Maintained by: Ana Pavlicic apavlicic@groupwise.swin.edu.au
Authorised by: Grenville Armitage garmitage@swin.edu.au

IndexSearchSite MapSwinburne Home Page