Skip to content

Latest commit

 

History

History
53 lines (41 loc) · 3.04 KB

net-util.md

File metadata and controls

53 lines (41 loc) · 3.04 KB
Raw

Network Metrics

Network Utilization

Task NP1: Network Utilization

  1. Assuming we have two machines connected with Gigabit Ethernet interfaces, what is the maximum expected throughput in kilo-bytes?

  2. For the following task you'll need two machines, or a partner:

    Machine/s Command Notes
    A + B sar -n DEV 2 Write down the receive/transmit packets/KB per-second. Keep this running for the entire length of the task
    A + B sar -n EDEV 2 These are the error statistics, read about them in man 1 sar. Keep this running for the entire length of the task
    A ip a Write down A's IP address
    A iperf -s This will start a server for our little benchmark
    B iperf -c <A's address> -t 30 Start the benchmark client for 30 seconds
    A iperf -su Replace the previous TCP server with a UDP one
    B iperf -c 172.30.0.251 -u -b 1000M -l 8k Repeat the benchmark with UDP traffic
    1. When running the client on B, use sar data to determine A's link utilization (in %, assuming Gigabit Ethernet)?
    2. What are the major differences between TCP and UDP traffic observable with sar?
    3. Start to decrease the UDP buffer length (ie. from 8k to 4k, 2k, 1k, 512, 128, 64).
      1. Does the throughput in KB increase or decrease?
      2. What about the throughput in packets?
      3. Look carefully at the iperf client and server report. Can you see any packet loss? Can you also see them in ifconfig?

Network Errors

While Linux provides multiple metrics for network errors including- collisions, errors, and packet drops, the kernel documentation indicates that these metrics meaning is driver specific, and tightly coupled with your underlying networking layers.

Tools

  • Most tools use /proc/net/dev to fetch network device information.
    • For example, try running sar -n DEV <interval> <count>.
  • Connection information for TCP, UDP and raw sockets can be fetched using /proc/net/tcp, /proc/net/udp, /proc/net/raw
    • For parsed socket information use netstat -tuwnp.
      • -t, -u, -w: TCP, UDP and raw sockets
      • -n: no DNS resolving
      • -p: the process owning this socket
  • The most comprehensive command-line utility is netstat, covering metrics from interface statistics to socket information.
  • Check iptraf for interactive traffic monitoring (no socket information).
  • Finally, nethogs provides a top-like experience, allowing you to find which process is taking up the most bandwidth (TCP only).

Discussion

  • What could be the reasons for packet drops? Which of these reasons can be measured on the receiving side?
  • Why can't you see the %ifutil value on EC2?
    • Hint: Network device speed is usually found in /sys/class/net/<dev>/speed.
    • Workaround: The nicstat utility allows you to specify the speed and duplex of your network interface from the command line: 
      nicstat -S eth0:1000fd -n 2