Simulating a zero-size congestion window - TCP-IP

This is a discussion on Simulating a zero-size congestion window - TCP-IP ; How can one deliberately create a situation where the web client advertizes to the server a zero-size window ? The Linux kernel has quieing discplines which can delay and drop packets on the client, but this doesn't seem to be ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: Simulating a zero-size congestion window

  1. Simulating a zero-size congestion window

    How can one deliberately create a situation where the web client
    advertizes to the server a zero-size window ?

    The Linux kernel has quieing discplines which can delay and drop
    packets on the client, but this doesn't seem to be enough to get
    there.

    (We've also tried to upload large files in the background,and other
    such crude attempts)

  2. Re: Simulating a zero-size congestion window

    RonHarding@netcor.com wrote:
    > How can one deliberately create a situation where the web client
    > advertizes to the server a zero-size window ?


    Stop pulling data out of the socket buffer, and make sure the sender
    sends enough data to fill it (the window). If you cannot get the
    sender to send enough data to fill the receiver's socket buffer, you
    could hack the receiver to use a smaller socket buffer.

    > The Linux kernel has quieing discplines which can delay and drop
    > packets on the client, but this doesn't seem to be enough to get
    > there.


    The delays are before TCP, so all you do there is simulate higher
    latency.

    The only way to get a receiving TCP to send a zero in the window field
    is to have the application stop pulling data out of the socket.

    Now, if you increase the delay "enough" you can get the sender to
    run-out of window at various times if it sends enough data at one
    time, and it will stop to await the next window update from the
    receiving TCP. However, that will not have the receiver sending ACKs
    with zero in the window field.

    > (We've also tried to upload large files in the background,and other
    > such crude attempts)


    Uploading would only potentially cause the server to advertise a zero
    window to the client for the TCP connection(s) over which the upload
    takes place. Keep in mind that wrt advertised window, all TCP
    connections are independent.

    rick jones
    --
    The computing industry isn't as much a game of "Follow The Leader" as
    it is one of "Ring Around the Rosy" or perhaps "Duck Duck Goose."
    - Rick Jones
    these opinions are mine, all mine; HP might not want them anyway...
    feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

+ Reply to Thread