COMPUTER SUPPORT FOR PROJECTS PERINSUR AND CAPER

Document Type: 
Collection: 
Document Number (FOIA) /ESDN (CREST): 
CIA-RDP85B01152R000700910029-2
Release Decision: 
RIPPUB
Original Classification: 
K
Document Page Count: 
4
Document Creation Date: 
December 21, 2016
Document Release Date: 
June 26, 2008
Sequence Number: 
29
Case Number: 
Publication Date: 
June 29, 1983
Content Type: 
MEMO
File: 
AttachmentSize
PDF icon CIA-RDP85B01152R000700910029-2.pdf144.52 KB
Body: 
Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 ? ROUTING AND RECORD SHEET SUBJECT: (Optional) DD/A Registry Computer Support for Projects FRO EXTENSION NO. ODP-83-904 D ODP / 2D00 Hd s DATE q . TO: (Officer designation, room number, and building) DATE OFFICER'S COMMENTS (Number each comment to show from whom RECEIVED FORWARDED INITIALS to whom. Draw a line across column after each comment.) EO/DDA 7D18 Hdqs. 983 ; `~~~~ 2. ADDA 7D18 Hdqs. 3. DDA 7D18 Hdqs. 5 JU 1983 e -4. 7. 8. 9. 10. 11. 12. 13. 14. 15. FORM 610 USE PREVIOUS I-79 EDITIONS - Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 DD/A Registry ODP-b3-904 2 9 JUN 1983 NF;t,i0RANDUiyi FO1i: Director of eersoniiel F'hOi'A : SUEJ1sCT: iiEH'LELI'NCE: Director of Data rrocessine, STAT Coruputer Support for Projects Your kemo, Same Subject, Ltd. 21 Plar.' b3 (ODP-63-456) 1. We have investigated all of the reasonable, orations for ssuu_~uoting your two Hicrodata Reality applications, at two different sites. Both applications currently run on a Microdata 8000 system with an older 6000 system being used for development work. cost-effective o tion would be to move the 8000 system to support and e the memory of the 6000 system to 128K and use it for The disadvantages of this alternative, however, are the questionable reliability of the 6000 system and its inability to provide good response at high concurrent user levels. 2. Attachment A presents the tabulated results of a series of benchmark tests run on-both the 6000 and 8000 systems by ODP in late, 1980. Another series of benchmarks were run recently which re- verified these earlier results. A significant point to be noted from the Attachment A tabulation is that the performance of the 6000 series machine begins to degrade significantly beyond the six concurrent user level even at its maximum memory configuration (128K bytes). 3.- Since it is our understanding that th -application will have to support up to nine concurrent users, using the 6'000 system for this application may'present performance problems during peak loading periods. Because of the performance problem just cited and because of the potential for significant hardware outage on the 60Q0 (it is a very old machine which uses ferrite core memory), Processing recommends that, if funds can be ied, a second Micro Data 8000 system be acquired to support when it..moves. STAT STAT STAT SIAI STAT SIAI STAT STAT Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 0 ? l'tnis would leave tine current 6000 anu b000 machines in their current location to support Iand development activities. SIAI 4. ata bUUU system can be acquirea at an approximate [-TAT bytes of memory for the 6UdO would cost about) Since ODP STAT did riot anticipate or budget for increased ilicrodata Reality capacity requirements, no funds ar.e available for this purpose in USP. 11.1 S P n "I~ fund this increase as a cost of moving its personnel STAT Please let us know how you wish to proceed. STAT cc: deputy Director for Administration, w/att Approved For Release 2008/06/27: CIA-RDP85BO1152R000700910029-2 Approved For Release 2008/06/27: CIA-RDP85BOl 152R000700910029-2 Attachment A 13 February 1981 Benchmark Results All programs used in the benchmark display menus at the terminal. The menus are "filled in" automatically with data that has been stacked in the terminal's input buffer. In a user environment, the ENTER key would be hit after each field in the menu was "filled in" and again after the entire menu is complete. Average response time was computed, therefore, as the total time for- a complete menu cycle divided by the number of fields in the menu plus one. Since the time for a_-complete menu cycle includes the time required to display the menu, the values in the table below may be higher than the response time that the user would actually experience. -------------------?----------------=--?------???--?....???-? ? Average Response Time (seconds) Users 6000-64K 6000-128K 8000-256K -------------------------------------------------- m ------- ? 1 ..1.0 10.0* 3.0 2 3.4 6.0 3.0 3 4.7 6.0 3.7 4 8.0 6.7 4.9 5 10.8 7.4 5.2 6 18.0 9.0 6.1 7 24.2 11.4 7.2 8 - 11.7 . 8.1 ? 9 - 14.2 9.4 it _ ~ - ? 11.2 1.. - 19.4 - `~ - - 15.9 ---------------------?--------------------------- -------- * An anomaly probably associated with system start-up procedures. Approved For Release 2008/06/27: CIA-RDP85BOl 152R000700910029-2