Priscilla, I read section 3.3 of RFC 793. Everything is clear...

Thanks for stressing out those important points of semantic.

Pierre-Alex

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
Priscilla Oppenheimer
Sent: Monday, May 27, 2002 1:59 PM
To: [EMAIL PROTECTED]
Subject: Re: TCP IP Sequence numbers [7:45131]


At 07:37 AM 5/27/02, Pierre-Alex Guanel wrote:
>I understand that when a host is sending an acknowledgement, it indicates
to
>the target the NEXT segment number that it expects to receive.

It's kind of a picky thing, but it's not accurate to say that the ACK
indicates the number of the next segment the host expects to receive.

TCP sequences bytes, not segments. The ACK number indicates the next byte
the hosts expects to receive. Of course, every segment has many bytes in
it. The Window field specifies how many bytes the host is ready to receive.


>When host A sends acknowledgement number X to host B, why does B sets its
>sequence number to X+1?

It doesn't. It sets the sequence number to the ACK number.

Where are you getting this info? It's not right.

The first thing I check when reading a description of TCP is whether they
know that bytes (not segments) are numbered. If they don't, I don't read
the rest. Also, be careful if they don't go beyond explaining the 3-way
handshake. Since the SYN, SYN ACK, and ACK packets don't actually have any
data bytes in them, the sequence number behaves a little differently. They
only go up by one, making it appear that segments are sequenced, rather
than bytes.

client sends: SYN initial seq number = 50
server sends: SYN initial seq number = 99, ACK = 51
client sends: ACK 100
client sends: bytes 51-100
server sends: ACK 101
client sends: bytes 101-150 ETC.

Actual size of packets depends on Max Segment Size (and how much data the
upper layer has to send). Actual number of packets sent before stopping and
waiting for ACK depends on the window size.

Check RFC 793, Stevens, Comer, many books other than "pass the Cisco test
quick" books. ;-)

Priscilla

>I would expect B to set its sequence number to X,
>since X is already the next segment number ...
>
>thanks
>
>Pierre-Alex
________________________

Priscilla Oppenheimer
http://www.priscilla.com




Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=45177&t=45131
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]

Reply via email to