re: http://www.garlic.com/~lynn/2009q.html#19 Mainframe running 1,500 Linux servers? http://www.garlic.com/~lynn/2009q.html#20 Mainframe running 1,500 Linux servers?
basic ESCON technology had been knocking around POK for quite some time before being released. One of the Austin engineers adopted it for RS6000 ... making it about 10% faster (220mbits/sec instead of 200/mbits/sec) and actually full-duplex (so it had quite a bit higher throughput) ... and used "Rochester" optic drivers ... which were quite a bit cheaper; http://www.garlic.com/~lynn/2009q.html#email870211 It made the SLA (serial-link-adaptors) for the RS6000 quite a bit better than ESCON ... but incompatible. Anybody wanting to do ESCON from RS6000 had to obtain it from completely different source. In the late 80s, LANL was pushing standardization of the 100mbyte Cray (half-duplex) channel as HIPPI and LLNL was pushing was serial technology they had as 1gbit FCS. We were somewhat involved in both activities ... examp http://www.garlic.com/~lynn/2006x.html#email920129 The Austin SLA engineer had started looking at exhancing SLA to 800mbits/sec but we convinced him to switch to working on FCS instead. There started to also be some ESCON participation ... looking to overlay ESCON half-duplex on top of full-duplex FCS. This was somewhat in line with work on "serial" HIPPI (i.e. moving Cray parallel half-duplex "copper" 100mbyte/sec channel to fiber). This old posts mentions jan92 meeting on doing 128-way RS6000 cluster scaleup http://www.garlic.com/~lynn/95.html#13 SSA Above also mentions moving Hursley 9333 disk subsystem (ran encapsulated scsi protocols over full-duplex "serial" copper, initially at 80mbits/sec ... full-duplex had several latency and throughput advantages). This turned into IBM's "SSA" technology. http://www-01.ibm.com/support/docview.wss?rs=505&uid=ssg1S1002348 SSA technology breaks 3000 IO bottlenecks http://www.3000newswire.com/subscribers/SSAPrimer.html We had suggested making it instead compatible with FCS protocol ... with an entry at 1/8th FCS over copper ... but allowing it to scaleup to full FCS over fiber (and be able to interoperate in FCS environment). recent thread/post mentioning some of the above: http://www.garlic.com/~lynn/2009q.html#0 Anyone going to Supercomputers '09 in Portland? http://www.garlic.com/~lynn/2009q.html#1 Anyone going to Supercomputers '09 in Portland? http://www.garlic.com/~lynn/2009q.html#2 Anyone going to Supercomputers '09 in Portland? and from long ago and far away (think.com hosted standards mailing list for both FCS and HIPPPI): Date: 24 Oct 90 04:59:59 EDT To: fc <[email protected]> Subject: October Fiber Channel Minutes TO: FC October 23, 1990 TO: X3T9.3 Fiber Channel Working Group Members FROM: Roger Cummings SUBJECT: FIBER CHANNEL WORKING GROUP MINUTES Please find attached a draft of the minutes of the ANSI X3T9.3 Fiber Channel Working Group of October 16 thru 18, 1990. Note that there are also fifteen Attachments to the minutes that relate to presentations at the meeting. The next Fiber Channel Working Group meeting will be held on November 1 and 2 as part of a Working Group week that is being hosted by Bill Spence of Texas Instruments at the Stouffer Hotel in Austin, TX. An announcement for this week is included. The following working group meeting will be held on the Wednesday and Thursday of the December plenary week (December 5 & 6) that is being hosted by AMD at the Hyatt on First Street in San Jose, CA. Note that the Working Group meeting will begin on Tuesday when the X3T9.3 plenary finishes, and thus will be a three day meeting. An announcement for the plenary week is attached. A schedule of X3T9.3 meetings (both plenaries and working groups) for the remainder of 1990 and the first half of 1991 is attached. Note that it is also intended to hold a three day Fiber Channel Working Group during the December plenary week. Hosts are still required for the August, October and December plenary weeks in 1991 and the November 1991 Working Group week. Note that ANSI has overruled as unconstitutional any attempt to limit discussion on a subject to those who have brought documentation on that subject. It has been decided to maintain a standalone Document Register for Fiber Channel, and the maintenance of such a register was begun at the October meeting. The register will be divided by subject and the numbering system used will be of the following format: FC-*/90-xxxRn where: * is the level to which it applies - 0, 1, 2, 3, 4, F or G (If it applies across levels then it is a G for General) xxx is the sequential number assigned by the Secretary n is the revision level, beginning at base 0 Note that future presenters are strongly requested to have such a document number and a page number on the top right-hand corner of each page of their presentation (this requirement is waived for the FC-0, -1 , -2, -3 and -4 documents themselves). They should contact myself as indicated below to obtain the specific sequential number. Regards 1990 SCHEDULE OF MEETINGS DATES EVENT LOCATION 10/29 - 11/02 Working Groups Austin, TX 12/03 - 12/06 Plenaries and Working Groups San Jose, CA 01/14 - 01/18 Working Groups Irvine, CA 02/18 - 02/20 Plenaries and Working Groups Austin, TX 03/18 - 03/22 Working Groups Cupertino, CA 04/22 - 04/25 Plenaries and Working Groups St. Petes, FL 05/20 - 05/24 Working Groups Wichita, KS 06/17 - 06/20 Plenaries and Working Groups Minneapolis, MN The format of the week for months containing only working groups is as follows: Monday SCSI Common Access Method Group Tuesday and Wednesday SCSI Working Group HIPPI Working Group Thursday and Friday Fiber Channel Working Group The format of the week for the months containing both plenaries and working groups is as follows: Monday Separate X3T9.2 (SCSI) and X3T9.3 plenaries (Monday portion of X3T9.3 devoted to IPI and HPPI). Tuesday Separate X3T9.2 (SCSI) and X3T9.3 plenaries (Tuesday portion of X3T9.3 devoted to Fiber Channel). Fiber Channel Working Group begins when X3T9.3 plenary ends (normally before noon). Wednesday Fiber Channel Working Group. Thursday Fiber Channel Working Group (December only). DOCUMENT REGISTER NUMBER TITLE FC-0/90-001R0 FC-0 Revision 1.7 002R0 High performance LW Laser SM 003R0 Militarized Optoectronic Devi FC-1/90-001R1.4 FC-1 Revision 1.4 FC-2/90-001R1.6 FC-2 Revision 1.6 002R0 Header Ideas and Proposal 10/ 003R0 Proposed Header Changes 10/2 004R0 FC Header Proposal 10/14 005R0 Some Thoughts on Headers 006R0 Proposed Header 007R0 An R_RDY Primitive 008R0 Moving the Link/Device bit 009R0 Time Frames FC-F/90-001R1 Fabric Requirements Revision 002R1 Event Log and Diagnostic Func FC-G/90-001R0 Laser Focus World Article 002R0 Operation Level Proposal Outl 003R0 ESCON Architecture Slides 004R0 ESCON Converter and Director -- 40+yrs virtualization experience (since Jan68), online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
