TCP packet size?

[Grant Edwards]

In article <39487D5B.FC96BE3F at rpgarchive.com>, chris wrote:

> Is there a way to ensure or predict packet size when using TCP?

Nope.

TCP/IP guarantees that the bytes get there in the right order.
It neither knows nor cares about any sort of boundaries between
bytes.

> I'm developing a Multi-user application that functions much
> like a muli-player game.  The applications are sharing xml
> data.  The xml data can get very big 3000-6000 bytes.
> Unfortunately, when the size of the network message gets over
> about 2000, its to get broken up into smaller packets.  So,
> when I send a 3000 byte message it gets cut up into 2-3
> packets.  Anther problem I get is when I send two small
> messages back to back, they are sometimes received together.
> Is there something I'm missing here?

That's the way TCP/IP works.  You're using a stream protocol to
try to send datagrams.  That's like trying to juggle tennis
balls while wearing shoes on your hands.  Shoes are a damn fine
invention, but you've got to use them for the right thing.

You want to use a datagram service instead of a stream service.

There's UDP, but it's unreliable, so you've got to add your own
ack/retry scheme -- bit of pain.  On Linux "man socket" claims
to support reliable datagrams:

   SOCK_SEQPACKET

       Provides a sequenced, reliable, two-way connection-
       based data transmission path for datagrams of fixed
       maximum length; a consumer is required to read an entire
       packet with each read system call.

That sounds like what you want, except it's not implemented for
address type AF_INET (which is probably what you're using).
Bummer.

There's also plain old unreliable datagrams (SOCK_DGRAM), and
reliable, unordered datagrams (SOCK_RDM). Probably the only
ones you can depend on being available are SOCK_STREAM (TCP/IP:
which is what you're using) and SOCK_DGRAM (UDP/IP: unreliable
datagrams).

>I tried using headers and breaking up messages into smaller chunks, but
>the size of the receiving packets doesn't always match what I sent.  I'd
>appreciate any help or suggestions on how I can tackle this problem.
>I'd be happy to send someone my network code.

Maybe you can find somebody who's already written a reliable
datagrram layer that runs on top of UDP/IP?

-- 
Grant Edwards                   grante             Yow!  Yow! We're going to
                                  at               a new disco!
                               visi.com