home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!crdgw1!rdsunx.crd.ge.com!rdsunx!barnett
- From: barnett@grymoire.crd.ge.com (Bruce Barnett)
- Newsgroups: comp.protocols.tcp-ip
- Subject: Why would a Sun take 1.5 seconds to retransmit a missed packet?
- Message-ID: <BARNETT.93Jan7091846@grymoire.crd.ge.com>
- Date: 7 Jan 93 14:18:46 GMT
- Sender: usenet@crd.ge.com (Required for NNTP)
- Reply-To: barnett@crdgw1.ge.com
- Organization: GE Corp. R & D, Schenectady, NY
- Lines: 19
- Nntp-Posting-Host: grymoire.crd.ge.com
-
- I am examing some data collected from a medical network, and noticed
- some unusual timings.
-
- A Sun is transmitting large files (3.5Mbytes average) to an archive
- system while receiving images from a scanner.
-
- Packets are missed by the archive system, so the Sun has to retransmit
- the missed packets. I have noticed large delays (0.5 - 1.5 seconds)
- while the sending machine has to "back up" and retransmit packets that
- were previously sent but never acknowledged. 35% of the sessions had
- inter-packet delays > 1 second. The receiving buffer size is 24K.
-
- Why would there be such a long delay? I would expect the data is in
- mbufs, so the time to resent the missed packets should be small.
- From the timing it looks like the data was occasionally on a disk.
- If we wanted to remove these conditions, am I correct in assuming the
- sending system needs more RAM or perhaps more buffers?
- --
- Bruce Barnett <barnett@crd.ge.com> uunet!crdgw1!barnett
-