enger@BLUTO.SCC.COM (04/04/88)
Bill: You wrote recently that, Even worse is that fact that in 4bsd this [end-to-end checksum] is a host wide option, so if I have a single SLIP based client of my NFS fileserver and I want him to get checksummed packets, I have to checksum packets for all Ethernet only local users as well, users who are more likely to be concerned with SPEED [my emphasis] and not with data corruption on the Ethernet. It occured to me that we now live in a new era: A.J. (after Jacobson). An era in which two small Sun workstations can obtain a TCP throughput of 8 Mbps, checksums, headers and all; and without fully consuming the CPUs! I am NOT commenting one way or the other on the SLIP controversy. Rather, I would like to offer the thought that the impact on performance caused by running with end-to-end checksumming enabled may be greater than it need be. Perhaps implementers could follow Mr. Jacobson's fine example by revising their algorithms and optimizing their code. Even when only modest throughputs are required, a user would still benefit from the reduction in CPU utilization. Bob