Answer the question
In order to leave comments, you need to log in
Is the fashion for counting checksums of network packets going away?
Once upon a time, I wrote a network analyzer on my knee (in PHP), to which you can “feed” a file with a record of network packets, and at the output get the contents of TCP streams assembled from individual IP packets.
Recently I fed a similar file to my analyzer, and was surprised to find that almost half of the packets were lost (although in the file itself, when viewed with "eyes", the characteristic HTTP headers present inside supposedly lost packets are easily detected).
Having fed the “broken” file to the Network Monitor itself (which wrote this file), I noticed that in all outgoing IP packets, zero is written in the “checksum” field, and BAD-checksum is written for the checksum field of the TCP segment.
That is, it was my own analyzer that dropped the packets, since it honestly checked the checksums.
But, despite the “left” checksums, all network applications (the packages of which I tried to catch) worked fine for themselves.
Has anyone else experienced similar things in their practice?
Answer the question
In order to leave comments, you need to log in
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question