[BXXPwg] Definition of SEQ fields in TCP mapping document

Marshall T. Rose mrose+mtr.netnews@dbc.mtview.ca.us
Mon, 15 Jan 2001 19:04:03 -0800


> Okay. But it seems impossible for an outgoing SEQ's ackno to
> correspond to *exactly* the number of bytes that have been received,
> as the incoming and outgoing activity can be happening
> concurrently. Is it sufficient to say that the ackno must be *less
> than or equal* to the number of bytes received when the SEQ message is
> composed?

i'll quote my previous message:

> an implementation is free to implement whatever algorithms it finds useful
> providing that its external behavior matches the specification.

what this means is that if you've structured your implementation so that you
don't know the exact number when you generate a SEQ, then the external
behavior of your implementation should not indicate this. so, using a policy
like "less than or equal" is likely to keep up appearances.

/mtr