mussar@bcars53.uucp (G. Mussar) (06/09/90)
We have measured throughput for multiple, simultaneous memory-to-memory data transfers using TCP/IP between UNIX workstations, with CISCO routers connected via a 1- or 2-hop X.25 packet switched network, and with the routers directly connected by a dedicated link. Results indicate about 88% throughput efficiency with the X.25 network, compared to 95% throughput efficiency with dedicated link. There is a slight degradation (2-3%) in both configurations, when long transmission (e.g. satellite) delays (50 to 500 ms) are added. It is useful to set parameters for high efficiency transfer: e.g. X.25 windows to 7, packet sizes to 512 bytes, etc. -- ------------------------------------------------------------------------------- Gary Mussar |Bitnet: mussar@bnr.ca | Phone: (613) 763-4937 BNR Ltd. | UUCP: ..uunet!bnrgate!bcars53!mussar | FAX: (613) 763-2626