home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!airgun!zeus.wg.waii.com!allen
- From: allen@zeus.wg.waii.com (Rusty Allen)
- Newsgroups: bit.listserv.ibmtcp-l
- Subject: Re: IBM's MVS TCP/IP problems
- Message-ID: <1738@airgun.wg.waii.com>
- Date: 29 Dec 92 21:43:59 GMT
- References: <IBMTCP-L%92122811274412@PUCC.PRINCETON.EDU>
- Sender: news@airgun.wg.waii.com
- Organization: Western Geo. - Div of Western Atlas Intn'l Inc., Houston, TX
- Lines: 41
- Nntp-Posting-Host: zeus.wg.waii.com
-
- Ron Lane rightly defends MVS TCP/IP but the problems with MVS TCP/IP is
- "The Death of a thousand cuts" rather than the "Shot to the Head". I point
- out a new feature that Ron listed.
-
- SMF records are very poorly implemented with
- MVS TCP/IP FTP V2R2.1 First I hate to point out that MVS TCP/IP FTP is an
- IBM PRODUCT and thus should have a IBM assigned SMF records number in
- the IBM reserved range 0-127.
-
- Speaking of SMF numbers I also note that there are seven events that can
- produce smf records and thus eat up SEVEN different SMF numbers in the
- LIMITED SMF user range. If the better SMF Standard record header with subtype
- had been used ( see MVS/ESA SMF GC28-1628-2 page 8-3 ) then only one
- smf record type number would be needed with subtypes for each event.
- I hear someone saying why not use one SMF record number and not worry
- about subtypes. One record number would be fine if there were few
- smf records but there are millions and millions of records. Standard
- smf utilities can filter on type/subtype/subsystem that is in the STANDARD
- smf Header with subtype.
-
- Well I seem to be talking about the records layout that MVS TCP/IP FTP
- is creating. I see in TCP/IP Planning and Customization (sc31-6085-02
- pg. 172-173) that everything is a displacement off +0. I remember the days of
- smf type4/5 records before smftype30 records with their indirect
- referencing to the data sections in the record. Every time new data
- needed to be added to the record all our accounting systems had to be changed
- first. There was not any version number in the old smf4/5 records, so there
- was not any to tell what level of code generated the records. I expect
- that over time the information in the TCP/IP FTP SMF records will change.
-
- I conclude that IBM TCP/IP is like a one legged man in a B*T kicking contest.
- "They fall down a lot and Hurt themself mostly". I will be chairing
- "C817-Understanding the IUCV sockets Interface of TCP/IP for MVS" and
- "C830-TCP/IP: Concept of Routing in a Heterogeneous Network" at SHARE_80
- in S.F. Mar 1-5. I hope to see some of ya'll there (ya'll is texan for
- "You Guys").
- --
- M.R. Allen My comments are my own, not my company's.
- Western Geophysical - A division of Western Atlas International,
- A Litton/Dresser Company
- DOMAIN addr: allen@wg.waii.com
-