Author |
Message
|
Bill57 |
Posted: Mon Oct 14, 2002 6:52 am Post subject: Does MQ "inflate" message size accross a channel |
|
|
Apprentice
Joined: 26 Jul 2002 Posts: 35 Location: Atlanta, GA
|
I work in an environment where MQ is a gateway for non-MQ environments. Many of our customers have technologies that utilize very small messages and thus can be transmitted with relatively small bandwidth.
An engineer here believes MQSeries inflates the size of a message in transit across channels. For example a message of 250 characters may actually become 4K.
Except for the 500 or so bytes needed for the header, I don’t know how that would happen. Any ideas???
Bill _________________ Bill Anderson
MQSeries Developer |
|
Back to top |
|
 |
mrlinux |
Posted: Mon Oct 14, 2002 9:14 am Post subject: |
|
|
 Grand Master
Joined: 14 Feb 2002 Posts: 1261 Location: Detroit,MI USA
|
Well the channel is going to have overhead for checking the message and verifying that it arrives, I dont know any specifics. _________________ Jeff
IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries |
|
Back to top |
|
 |
pgorak |
Posted: Tue Oct 15, 2002 3:05 am Post subject: |
|
|
 Disciple
Joined: 15 Jul 2002 Posts: 158 Location: Cracow, Poland
|
Apart from MQMD header, the transmission queue header, MQXQH, is added by QM when the message is put on a transmission queue. The overhead resulting from this is less than 500 bytes.
Piotr |
|
Back to top |
|
 |
mrlinux |
Posted: Tue Oct 15, 2002 4:52 am Post subject: |
|
|
 Grand Master
Joined: 14 Feb 2002 Posts: 1261 Location: Detroit,MI USA
|
Well there is the acutal communications information like sequence numbers, batch information, and other handshakes. _________________ Jeff
IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries |
|
Back to top |
|
 |
Bill57 |
Posted: Wed Oct 16, 2002 7:54 am Post subject: |
|
|
Apprentice
Joined: 26 Jul 2002 Posts: 35 Location: Atlanta, GA
|
And the answer is.... the engineer was referring to the communication overhead involved in sending and receiving across an MQ channel. That overhead has to do with the sync point control between two channels. This requires lots of data packets to be exchanged in excess of the actual message.
When he first made his point, he phrased it by saying that MQ "blows up" a small message to around 4K. That confused me as to what he was getting at.
Anyway, I think the actual comms overhead is a little over 1100 bytes. So, indeed, for people trying to operate with low bandwidth, MQ presents a problem for them.
Thanks for entertaining the question...... sorry for the confusion  _________________ Bill Anderson
MQSeries Developer |
|
Back to top |
|
 |
mrlinux |
Posted: Wed Oct 16, 2002 8:44 am Post subject: |
|
|
 Grand Master
Joined: 14 Feb 2002 Posts: 1261 Location: Detroit,MI USA
|
MQEveryPlace has much smaller overhead for messages. _________________ Jeff
IBM Certified Developer MQSeries
IBM Certified Specialist MQSeries
IBM Certified Solutions Expert MQSeries |
|
Back to top |
|
 |
PeterPotkay |
Posted: Sun Oct 20, 2002 11:06 am Post subject: |
|
|
 Poobah
Joined: 15 May 2001 Posts: 7722
|
Here is an MQ message "on the wire". My network people gave me this.
The first four lines up to '00 00' just before 'TSH' in ASCII starts, is a
TCP header.
The TSH is the start of the MQ Channel header.
The XQH is the start of the transmission header.
The MD is the start of the MQMD, and then the real message follows that.
Code: |
ADDR HEX ASCII
0000: 00 02 a5 f0 b5 2c 00 50 8b c8 22 08 08 00 45 00 | ..¥ðµ,.P<È"...E.
0010: 05 dc 70 1e 40 00 80 06 b3 11 9d d1 4a b9 9d d1 | .Üp.@.EUR.³.?ÑJ¹?Ñ
0020: 4b 90 06 ac 05 86 00 42 08 bd e0 19 8c 30 50 10 | K?.¬.?.B.½à .OE0P.
0030: 21 98 de 1c 00 00 54 53 48 20 00 00 61 99 02 04 | !~Þ...TSH ..a(tm)..
0040: 30 00 13 dc a3 3d 21 70 01 00 22 02 00 00 b5 01 | 0..Ü£=!p.."...µ.
0050: 00 00 4d 53 48 20 3a f0 01 00 bd 5f 00 00 00 00 | ..MSH :ð..½_....
0060: 00 00 69 61 00 00 58 51 48 20 01 00 00 00 48 49 | ..ia..XQH ....HI
0070: 47 2e 50 43 2e 50 4c 30 30 2e 42 52 53 2e 50 45 | G.PC.PL00.BRS.PE
0080: 52 53 49 53 54 45 4e 43 45 2e 52 45 51 20 20 20 | RSISTENCE.REQ
0090: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 48 49 | HI
00a0: 47 49 44 47 51 31 20 20 20 20 20 20 20 20 20 20 | GIDGQ1
00b0: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
00c0: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 4d 44 | MD
00d0: 20 20 01 00 00 00 00 00 00 00 08 00 00 00 00 5e | .............^
00e0: 1a 00 00 00 00 00 11 01 00 00 33 03 00 00 20 20 | ..........3...
00f0: 20 20 20 20 20 20 00 00 00 00 00 00 00 00 41 4d | ........AM
0100: 51 20 48 49 47 42 52 53 51 31 20 20 20 20 3d 72 | Q HIGBRSQ1 =r
0110: 71 1e 00 17 0b b2 41 4d 51 20 48 49 47 42 52 53 | q....²AMQ HIGBRS
0120: 51 31 20 20 20 20 3d 72 71 1e 00 17 0b 72 00 00 | Q1 =rq....r..
0130: 00 00 20 20 20 20 20 20 20 20 20 20 20 20 20 20 | ..
0140: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
0150: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
0160: 20 20 48 49 47 42 52 53 51 31 20 20 20 20 20 20 | HIGBRSQ1
0170: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
0180: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
0190: 20 20 4d 51 4d 20 20 20 20 20 20 20 20 20 00 00 | MQM ..
01a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 | ................
01b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 20 20 | ..............
01c0: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 |
01d0: 20 20 20 20 20 20 20 20 20 20 20 20 20 20 1c 00 | ..
01e0: 00 00 4d 51 53 65 72 69 65 73 20 43 6c 69 65 6e | ..MQSeries Clien
01f0: 74 20 66 6f 72 20 4a 61 76 61 20 20 20 20 32 30 | t for Java 20
0200: 30 32 31 30 30 39 31 32 32 36 32 34 39 36 20 20 | 02100912262496
0210: 20 20 ac ed 00 05 75 72 00 33 5b 4c 63 6f 6d 2e | ¬Ã..ur.3[Lcom.
|
_________________ Peter Potkay
Keep Calm and MQ On |
|
Back to top |
|
 |
|