nanog mailing list archives
Re: MTU of the Internet?
From: gary flynn <gary () habanero jmu edu>
Date: Thu, 5 Feb 1998 10:31:32 -0500
My guess is that when you're traversing a network or NAP that is dropping 10-20% of its packets, smaller packets would have a higher survival rate and thus result in better overall throughput due to the lack of timer based retries. It probably has nothing to do with MTUs or fragmentation across the net. Its probably due to the lossy nature of our current environment (i.e. overloaded NAPs). A simple test should show this. Slap a protocol analyzer on the line and see how many retransmitted packets and/or response delays there are for various MTUs. For heaven's sake, don't change system software until we know what the *real* cause is! We've got enought kludges out there already :) Gary Flynn Network Analyst James Madison University
Current thread:
- Re: MTU of the Internet?, (continued)
- Re: MTU of the Internet? Marc Slemko (Feb 05)
- Re: MTU of the Internet? Phil Howard (Feb 05)
- Re: MTU of the Internet? Kevin A. Smith (Feb 05)
- Re: MTU of the Internet? Phil Howard (Feb 05)
- Re: MTU of the Internet? Marc Slemko (Feb 05)
- Re: MTU of the Internet? Phil Howard (Feb 05)
- Re: MTU of the Internet? Perry E. Metzger (Feb 05)
- Re: MTU of the Internet? Dan Foster (Feb 05)
- Re: MTU of the Internet? Marc Slemko (Feb 04)
- Re: MTU of the Internet? George Swallow (Feb 05)
- Re: MTU of the Internet? ken emery (Feb 05)
- Re: MTU of the Internet? Frank Kastenholz (Feb 06)
- Re: MTU of the Internet? Phil Howard (Feb 06)
- Re: MTU of the Internet? Stephen Sprunk (Feb 06)
- Message not available
- Re: MTU of the Internet? Jay R. Ashworth (Feb 06)
- Re: MTU of the Internet? Perry E. Metzger (Feb 06)
- Message not available
- Re: MTU of the Internet? Jay R. Ashworth (Feb 06)