MUD-Dev
mailing list archive
[ Other Periods
| Other mailing lists
| Search
]
Date:
[ Previous
| Next
]
Thread:
[ Previous
| Next
]
Index:
[ Author
| Date
| Thread
]
Re: [MUD-Dev] Transport layer (UDP vs TCP)
- To: mud-dev#null,net
- Subject: Re: [MUD-Dev] Transport layer (UDP vs TCP)
- From: cg#ami-cg,GraySage.Edmonton.AB.CA (Chris Gray)
- Date: Wed, 18 Mar 98 18:46:01 MST
- >Received: by ami-cg.GraySage.Edmonton.AB.CA (V1.17-beta/Amiga) id <8vzy#ami-cg,GraySage.Edmonton.AB.CA>; Wed, 18 Mar 98 18:46:01 MST
[Ben Greear:]
:For instance, I send a 512 byte UDP packet. Can the receiver ever just
:receive the first 256 bytes or so? Surely not, as this would make it
:basically impossible to code anything without writing another layer on
:top of it...
My recollection (its been a few years since I did our UDP stuff at work),
the packets are not broken up by UDP. If they are too long, they can be
completely dropped, silently, but I don't believe you can get a portion
of them. They can also be delivered out of order, and duplicated. For
strictly ethernet, we've found that you can safely use packets of about
9K. If you want larger, UNIXes seem to have sysadmin configuration tools
that increase their maximum packet sizes. Our stuff runs better with
bigger packets, so we try to do that for timing runs.
--
Chris Gray cg#ami-cg,GraySage.Edmonton.AB.CA
- Thread context:
- Re: [MUD-Dev] (fwd) Functional Security, (continued)
- Re: [MUD-Dev] Transport layer (UDP vs TCP),
Chris Gray cg#ami-cg,GraySage.Edmonton.AB.CA, Thu 19 Mar 1998, 01:54 GMT
- Re: [MUD-Dev] SfD: Clientside Caching,
Chris Gray cg#ami-cg,GraySage.Edmonton.AB.CA, Thu 19 Mar 1998, 01:54 GMT
- SIMULATING FUTURE HISTORIES: THE NAU SOLAR SYSTEM SIMULATION & MARS SETTLEMENT,
J C Lawrence claw#under,engr.sgi.com, Thu 19 Mar 1998, 00:42 GMT
- (fwd) INFO: [client] Chaco looking for new parent for Pueblo,
J C Lawrence claw#under,engr.sgi.com, Wed 18 Mar 1998, 18:07 GMT
[ Other Periods
| Other mailing lists
| Search
]