| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| SSL Performance over different Cipherspecs |
« View previous topic :: View next topic » |
| Author |
Message
|
| BBM |
 Posted: Thu Sep 28, 2006 3:44 am Post subject: SSL Performance over different Cipherspecs |
 |
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
Hi,
We have conducted some fairly extensive performance testing for our UAT SSL connections with some surprising results. The set up is as follows:
Two queue managers connected via 1 SDR->RCVR channel on the same box. We originally tested over the LAN and latterly WAN connections but found variations in network performance skewed the results.
Here are the results for batches of 250000 persistent messages sent with various cipherspecs. the bit we don't understand is why TRIPLE_DES_SHA should be faster than NULL_SHA or even no SSL!! We're wondering whether TRIPLE DES compresses also. The first five results refer to the five batches - the sixth result is an average of all five.
Any thoughts or has anyone any conflicting results?
TRIPLE DES SHA US 11 Min 30 Sec, 11 Min 20 Sec, 11 Min 10 Sec, 11 Min 15 Sec , 11 Min 15 Sec, 11 Min 18 Secs
TRIPLES DES SHA US with SSLCAUTh 11 Min 40 Sec, 11 Min 35 Sec, 11 Min 30 Sec , 11 Min 35 Sec, 11 Min 35 Sec, 11 Min 35 Secs
NULL_SHA, 11 Min 45 Sec, 11 Min 45 Sec, 11 Min 50 Sec, 11 Min 40 Sec, 11 Min 40 Sec, 11 Min 44 Secs
NULL_SHA with SSLCAUTH, 11 Min 50 Sec, 11 Min 30 Sec, 11 Min 40 Sec, 11 Min 35 Sec, 11 Min 30 Sec, 11 Min 37 Sec
RC4_SHA_US, 11 Min 50 Sec, 12min 28secs, 12 mins 26secs, 11min 55secs, 12 Mins 42secs, 12 mins 28 secs
RC4_SHA_US with SSLCAUTH, 12 Min 35 Sec, 12 Min 30 Sec, 12 Min 20 Sec, 12 Min 10 Sec, 11 Mins 50 Sec, 12 mins 17 sec
WITHOUT SSL 11 Min 55 Sec, 11 Min 45 Sec, 11 Min 20 Sec, 11 Min 35 Sec, 11 Min 55 Sec, 11 Min 42 Secs |
|
| Back to top |
|
 |
| interactivechannel |
| Posted: Tue Oct 03, 2006 6:02 am Post subject: |
|
|
Voyager
Joined: 20 May 2003
Posts: 94
Location: uk
|
Did the same thing and recall TDES_SHA was quicker than NULL_SHA. Compression is a good guess.
However in my tests, no SSL was significantly quicker. Also my queue managers were distributed. |
|
| Back to top |
|
|
| BBM |
| Posted: Wed Oct 04, 2006 12:22 am Post subject: |
|
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
Thanks for that. It's interesting you saw the same behaviour - I'll follow up with IBM.
Cheers
BBM |
|
| Back to top |
|
|
| hguapluas |
| Posted: Fri Oct 06, 2006 11:50 am Post subject: |
|
|
Centurion
Joined: 05 Aug 2004
Posts: 105
Location: San Diego
|
| Have you tried the AES options? There are 2, one 128, one 256. Would be curious to see your results with AES as that is supposed to be faster than Triple-DES but I have not been able to confirm this. |
|
| Back to top |
|
|
| BBM |
| Posted: Sat Oct 07, 2006 9:50 am Post subject: |
|
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
Hi,
Thanks for your reply. We didn't test AES as Triple DES was the company standard.
However, having seen your comments we are going to repeat our tests for both flavours of AES and see what difference it makes. Judging by our results so far it will probably be 100 times faster than no ssl at all!
Many thanks
BBM |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
