home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!uvaarpa!darwin.sura.net!paladin.american.edu!auvm!PURCCVM.BITNET!REXB
- Message-ID: <IBMTCP-L%93012810105254@PUCC.PRINCETON.EDU>
- Newsgroups: bit.listserv.ibmtcp-l
- Date: Thu, 28 Jan 1993 09:52:29 EST
- Sender: IBM TCP/IP List <IBMTCP-L@PUCC.BITNET>
- From: Rex Bontrager <REXB@PURCCVM.BITNET>
- Subject: Re: Multiple MVS Telnet Servers
- In-Reply-To: Message of Thu, 28 Jan 1993 09:21:57 EST from <TOMPKINS@AKRONVM>
- Lines: 28
-
- On Thu, 28 Jan 1993 09:21:57 EST Frank Tompkins said:
- >One last point, IBM does warn that this configuration is for "testing",
- >and may not be acceptable for production. From what I remember, this
- >was out of concern for the load on the "primary" address space, which
- >is now serving as a router.
-
- My bosses were concerned about the performance hit of playing router,
- so we ran some tests. Our tests had 0-4 users telnet into the secondary
- TCPIP so that TCPIP1 was playing router and TCPIP2 was supporting the
- telnet session. Each user continuously performed full-screen scrolls
- while we monitored the CPU usage of the two TCPIPs.
-
- The results: TCPIP1 and TCPIP2 used about the same resources while
- performing their respective functions; i.e., playing router and
- supporting telnet consume approximately the same number of cycles.
- Thus, a person who telnets into TCPIP2 uses twice the CPU
- (router+telnet) as someone who telnets into TCPIP1 (excluding the
- business which they are really trying to get done).
-
- We are running HAL v1.0.
-
-
- Rex Bontrager
- IBM Systems Programmer, IBM Postmaster
- Purdue University Computing Center
- Bitnet: rexb@purccvm
- Internet: rexb@vm.cc.purdue.edu
- Phone: (317) 494-1787 ext. 256
-