PIRARD%vm1.ulg.ac.be@CUNYVM.CUNY.EDU (Andr'e PIRARD) (04/12/91)
On Wed, 10 Apr 91 04:13:05 GMT <tcp-ip-relay@NIC.DDN.MIL> said: >I've heard some interesting things about SLIPs but don't know >how they work or how to install one. Can anyone enlighten me??? I'll comment about the other often asked questions: whether to, why so slow... SLIP performance (on low bit rate lines) is that of the (sending) TCP's. This holds for any slow line, but is aggravated by SLIP's unreliability. Indeed, a TCP sending data has to figure the timeouts used to resend data carefully. Too short makes it a plague of duplicate packets. Line errors may get the value occasionally very high and get things slow if the TCP is sluggish in reducing the timeout value. Some rough figures (9600 bauds MNP modems) when FTP receiver is: SunOS 4.0.3 1.1 Kb/sec (presumably any BSD 4.3) VM TCP/IP 1.2.2 (IBM) .75 CUTCP .25 (could be improved with other parameters, but had better do without such customization). All hosts were on Ethernet, SLIP was between routers. MNP means that the impact of line errors was almost nil (but another cause for loosing packets got me mad for some time). MNP made the throughput very slightly better, nothing to speak of, but shows an annoying echo latency in terminal mode that doesn't exist without it. I guess it's because of a time value it uses to wait for enough bytes to assemble into packets to minimize the checksum overhead, and this has no reason to be for what's the noninterrupted flow of datagrams. Is there a way to get rid of this latency? Andr'e PIRARD SEGI, Univ. de Li`ege B26 - Sart Tilman B-4000 Li`ege 1 (Belgium) pirard@vm1.ulg.ac.be or PIRARD%BLIULG11.BITNET@CUNYVM.CUNY.EDU