Changes between Version 19 and Version 20 of Internal/DesignNotes


Ignore:
Timestamp:
Apr 10, 2006, 4:46:44 AM (19 years ago)
Author:
kishore
Comment:

—

Legend:

Unmodified
Added
Removed
Modified

  • Internal/DesignNotes

    v19 v20  
    181810000 packets were sent with an interval of 10ms between consecutive packets. The figure above(on the left) shows the number of packets lost per receiver. From the figure, all receivers except for one (node14-3) receives all 10000 packets. The same behavior was exhibited in another run of the same experiment. This seems to indicate that there might be something wrong with node14-3. The figure (on the right) shows the number of packets successfully received at each receiver.
    1919
     20We also saw the same behavior when we used 5000 packets instead of 10000.
     21
    2022=== Experiment 2 ===
    2123
    22 [[Image(Internal/DesignNotes:num-pkts-lost.png)]]
    23 [[Image(Internal/DesignNotes:num-pkts-recvd.png)]]
     24[[Image(Internal/DesignNotes:num-pkts-lost.3.png)]]
     25[[Image(Internal/DesignNotes:num-pkts-recvd.3.png)]]
    2426
    25 10000 packets were sent with an interval of 10ms between consecutive packets. The following figure shows the number of packets lost per receiver. From the figure, all receivers except for one (node14-3) receives all 10000 packets. The same behavior was exhibited in another run of the same experiment. This seems to indicate that there might be something wrong with node14-3.
    26 
     2710000 packets were sent "as fast as possible". There was no "sleep" statement between two send events at the sender. The figures above show the number of packets lost per receiver (on the left) and the number of successful packet receptions (on the right). It is unclear from current experiments as to where this loss is occurring. Some sources of this loss could be
     28 * buffer (between user-space and kernel-space, between kernel and driver, between driver and card) overflows at the sender or receiver?
     29 * the switches in the network cannot handle the load? Seems unlikely.
     30 
    2731=== Experiment 3 ===
    2832
    29 [[Image(Internal/DesignNotes:num-pkts-lost.png)]]
    30 [[Image(Internal/DesignNotes:num-pkts-recvd.png)]]
     33[[Image(Internal/DesignNotes:num-pkts-lost.2.png)]]
     34[[Image(Internal/DesignNotes:num-pkts-recvd.2.png)]]
    3135
    32 10000 packets were sent with an interval of 10ms between consecutive packets. The following figure shows the number of packets lost per receiver. From the figure, all receivers except for one (node14-3) receives all 10000 packets. The same behavior was exhibited in another run of the same experiment. This seems to indicate that there might be something wrong with node14-3.
    33 
    34 === Experiment 4 ===
    35 
    36 [[Image(Internal/DesignNotes:num-pkts-lost.png)]]
    37 [[Image(Internal/DesignNotes:num-pkts-recvd.png)]]
    38 
    39 10000 packets were sent with an interval of 10ms between consecutive packets. The following figure shows the number of packets lost per receiver. From the figure, all receivers except for one (node14-3) receives all 10000 packets. The same behavior was exhibited in another run of the same experiment. This seems to indicate that there might be something wrong with node14-3.
     365000 packets were sent "as fast as possible". There was no "sleep" statement between two send events at the sender. The figures above show the number of packets lost per receiver (on the left) and the number of successful packet receptions (on the right). It is unclear from current experiments as to where this loss is occurring.
    4037
    4138= Design Notes for nodehandler =