 
| Subject: | |
| From: | |
| Reply To: | |
| Date: | Fri, 2 Jan 2004 09:27:10 -0800 |
| Content-Type: | text/plain |
| Parts/Attachments: |
|
|
Hello Folks @ 3000-l,
RE: FTP vs.. DSCOPY performance.
I hope everyone had a happy holiday season! I have returned to Atlanta from the wild wild west afer short stays in Tombstone, Sonna, Flagstaff & Phoenix Airzona!
I had a chance to run some test over the holidays and evaluate FTP vs. DSCOPY performance.... Here are my findings:
Test file set:
ACCOUNT= HOFMESTR GROUP= TEST
FILENAME CODE ------------LOGICAL RECORD----------- ----SPACE----
SIZE TYP EOF LIMIT R/B SECTORS #X MX
TEST1G8A 80B FA 13421772 13421772 3 4194304 * *
TEST1G8B 40W FB 13421772 13421772 3 4194304 * *
TEST1GBA 256B FA 4194304 4194304 1 4194304 * *
TEST1GBB 128W FB 4194304 4194304 1 4194304 * *
TEST1GBS 1B BA 1073741824 2147483647 1 4194304 * *
TEST2GBA 256B FA 8388608 8388608 1 8388608 * *
TEST2GBB 128W FB 8388608 8388608 1 8388608 * *
TEST2GBS 1B BA 2147483647 2147483647 1 8388608 * *
TEST5GBA 256B FA 20971520 20971520 1 20971520 * *
TEST5GBB 128W FB 20971520 20971520 1 20971520 * *
======================================================================
Kbytes /sec
1gb 1gb 2gb 5gb 1gb 1gb 2gb 5gb 1gb 1gb
bin bin bin bin asci asci asci asci byte byte
r 80 r256 r256 r256 r 80 r256 r256 r256 strm strm
---- ---- ---- ---- ---- ---- ---- ---- ---- ----
FTP 1670 1660 1522 1183 619 625 608 580 1672 1522
DSCOPY 871 866 858 844 865 881 880 852 fail fail
======================================================================
((Seconds /60) /n Gb) {Minutes to transfer 1Gb}
1gb 1gb 2gb 5gb 1gb 1gb 2gb 5gb 1gb 1gb
bin bin bin bin asci asci asci asci byte byte
r 80 r256 r256 r256 r 80 r256 r256 r256 strm strm
---- ---- ---- ---- ---- ---- ---- ---- ---- ----
FTP 10.4 10.5 11.5 14.8 28.9 28.1 29.0 30.4 10.5 11.5
DSCOPY 20.3 20.3 20.4 20.7 20.2 20.0 20.0 20.7 fail fail ======================================================================
Notes:
1) The test above are run on 989-400 MPE 7.5 w/8190 Mb memory
transferring to a 987-200 MPE 6.5 w/768 Mb memory over a isolated
100BT Full Duplex switch.
2) Your Mileage Will Vary (YMWV) - The documented performance numbers
are useful to evaluate the differences between DSCOPY and FTP in
a controlled (isolated) environment, but these numbers will change
with different systems, memory sizes, disc performance, network
interface cards and link traffic levels (In this test the
performance of the 987-200 is the obvious limiting factor).
3) Dscopy does not support 'byte stream' file transfers. Error
reported is: SOURCE FILE CANNOT BE BYTE STREAM RECORD TYPE
(NFT/3000 ERR 115).
4) With current patches FTPHD66/FTPHD67/FTPHD68 'byte stream' file
transfer is as fast or faster than binary transfers.
5) Dscopy of 'typical' 80 byte or 256 byte ascii files is 30-35%
faster than ftp.
6) Ftp of 'typical' 80 byte or 256 byte binary files is 30-45% faster
than dscopy With current patches FTPHD66/FTPHD67/FTPHD68 installed.
7) Ftp performance does drop significantly as the file record length
decreases for both ascii and binary file transfers.
Important fixes:
1. 8606331712 FTP Byte stream performance fix.
2. 8606328875 FTP Repair binary >4Gb.
3. 8606299250 FTP Repair >40Gb & >4Gb performance Improvements.
4. 8606269614 FTP Repair variable >3Gb.
5. 8606294908 FTP Repair Inbound Unix & Nt >4Gb.
Patches:
FTPHD66 for C.65.00 (Site Specific)
FTPHD67 for C.70.00 (Site Specific)
FTPHD68 for C.75.00 (Beta Test)
Note-1: When requesting the fix from the HP-Response Center, it is important to request the fix by SR number.
Note-2: Due to the lack of interest for the 6.5 and 7.0 fixes, they are still Site Specific. We typically require additional problem verification before a Site Specific patch is sent by the HP-Response Center.
I hope this helps.
Regards,
James Hofmeister
Email: <first>.<last>@hp.com
Hewlett Packard - Global Solutions Engineering (WTEC)
P.S. My Ideals are my own, not necessarily my employers.
---------------------------------
Do you Yahoo!?
Free Pop-Up Blocker - Get it now
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
|
|
