| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 867.1 | Version of DECmcc? | TOOK::ANWARUDDIN | Anwar | Fri Apr 05 1991 14:11 | 2 |
| What version of DECmcc are you using? Any other details you can provide
will be helpful.
|
| 867.2 | The last DECmcc Version... | ROM01::NINNI | Don't worry...be happy | Mon Apr 08 1991 11:54 | 7 |
|
I'm using the SSB version of DECMCC v.1.1, VMS 5.4, DECnet VAX
Extension (Wave 1).
Do you need other information (CDD version,....)?
Giuseppina
|
| 867.3 | Any suggestions...? | ROM01::NINNI | Don't worry...be happy | Wed Apr 17 1991 09:26 | 7 |
|
Have you any news? I'm very interested wich is the reason. I discovered
this situation, using the graphic display and copying a big file on
that circuit. In normal situation (not too traffic) the displayed datas
are good.
Thanks in advance
Giuseppina
|
| 867.4 | Can't reproduce it.. | TOOK::ANWARUDDIN | Anwar | Thu Apr 18 1991 12:33 | 7 |
| re: .3
We have not been able to reproduce the problem. We will try
copying a large file, as you did, and see if it can be reproduced.
We will post a note about the findings.
Thanks for your patience.
|
| 867.5 | OK, I will wait | ROM01::NINNI | Don't worry...be happy | Fri Apr 19 1991 13:42 | 3 |
| OK thank you
I will wait
|
| 867.6 | More info please.. | TOOK::ANWARUDDIN | Anwar | Fri Apr 19 1991 18:13 | 15 |
| We have had no luck reproducing the problem. Could you please
provide the following information?
1. Type of circuit and speed
2. Size of file
3. Duration (Export interval)
Could you just do a SHOW STATISTICS for the circuit while copying
the same large file & for the same duration and see what the utilization
is?
You don't have to export data.
Thanks a lot.
|
| 867.7 | comments/questions | JETSAM::WOODCOCK | | Mon Apr 22 1991 14:48 | 33 |
| This will prove to be one of the more important problems you folks work
out. This seemed to be one of the biggest complaints about NMCC (data
integrity) and one reason why net managers are eager to migrate to MCC.
Maybe past experience from NMCC can help nip this one.
While using NMCC I have never personally experienced percentages over
100 while viewing histograms. But while using NTA on top of NMCC I have
seen numbers over 100% (with counter timer at 65000). In reading over the
docs a few areas of interest could lead to the prob, the formula,
'normalization' of data, and the handling of time stamps and the
extraction of data from rdb.
In doing a small test of the formula using PA it appeared to be dead
accurate compared to what was documented. But the counter timer had not
zeroed the counters in the middle. Our counter timers are set to 65000
so it's only once every 18 hours. The handling of zeroed counters could
have an effect on accuracy. What is the value of base note's counter timer
and did the counters zero during the testing?
Both NMCC and MCC advertize some sort of normalization of data. I'm not
big as a statisticion but how this is handled would certainly have an
affect on accuracy. Does this normalization of data get used when the
request times fall between poll times? And if so does the formula reflect
what the user requested for time interval or what was returned by MCC
by this normalization process? Could it be possible that what gets passed
(bytes) into the formula goes back further into time because it fell
between polls but then the user interval is used (somewhat shorter) as the
duration time?
Hopefully some of these questions could touch a right button.
brad...
|
| 867.8 | The answers... | ROM01::NINNI | Don't worry...be happy | Wed Apr 24 1991 07:15 | 887 |
|
We discoverd the problem the first time, without any file transfer.
We are working on two circuits at different speed:
-FNAL circuit at 64 kbps
-RTROM circuit at 768 kbps
In the test with file transfer, we were in the following situation:
1. Circuit at 768 kb/s
2. The dimension of the File to transfer: 3406 blocks
3. Statistics duration: with 30'', after with 1', etc.)
We had always the same problem.
We are using the following counter timer values:
- at executor level: 43200 sec
- at circuit level : 1800 sec
In order to collect statistics we used the following command:
> show node4 rtgar3 circuit rtrom outbound utilization, for start (-0:00) -
> duration 0:1:0, at every 0:2:0, to file prova.dat
I have obtained the following results, also without the file transfer:
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:10:17 Statistics
Outbound Utilization = 7.38 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:12:09 Statistics
Outbound Utilization = 31.05 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:14:09 Statistics
Outbound Utilization = 86.09 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:16:09 Statistics
Outbound Utilization = 152.93 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:18:09 Statistics
Outbound Utilization = 205.35 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:20:09 Statistics
Outbound Utilization = 231.98 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:22:09 Statistics
Outbound Utilization = 11.01 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:24:09 Statistics
Outbound Utilization = 11.99 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:26:09 Statistics
Outbound Utilization = 15.52 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:28:09 Statistics
Outbound Utilization = 12.16 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:30:09 Statistics
Outbound Utilization = 20.02 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:32:09 Statistics
Outbound Utilization = 12.45 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:34:09 Statistics
Outbound Utilization = 2.78 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:36:09 Statistics
Outbound Utilization = 1.83 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:38:09 Statistics
Outbound Utilization = 3.64 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:40:09 Statistics
Outbound Utilization = 9.06 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:42:09 Statistics
Outbound Utilization = 8.9 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:44:09 Statistics
Outbound Utilization = 19.1 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:46:09 Statistics
Outbound Utilization = 23.12 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:48:09 Statistics
Outbound Utilization = 26.08 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:50:09 Statistics
Outbound Utilization = 33.44 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:52:09 Statistics
Outbound Utilization = 8.33 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:54:09 Statistics
Outbound Utilization = 5.88 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:56:09 Statistics
Outbound Utilization = 4.8 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 11:58:09 Statistics
Outbound Utilization = 5.65 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:00:09 Statistics
Outbound Utilization = 4.16 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:02:09 Statistics
Outbound Utilization = 5.48 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:04:09 Statistics
Outbound Utilization = 6.27 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:06:09 Statistics
Outbound Utilization = 5.81 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:08:09 Statistics
Outbound Utilization = 6.1 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:10:09 Statistics
Outbound Utilization = 6.88 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:12:09 Statistics
Outbound Utilization = 9.64 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:14:09 Statistics
Outbound Utilization = 24.27 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:16:09 Statistics
Outbound Utilization = 38.92 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:18:09 Statistics
Outbound Utilization = 54.36 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:20:09 Statistics
Outbound Utilization = 71.83 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:22:09 Statistics
Outbound Utilization = 9.22 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:24:09 Statistics
Outbound Utilization = 8.96 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:26:09 Statistics
Outbound Utilization = 7.44 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:28:09 Statistics
Outbound Utilization = 8.5 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:30:09 Statistics
Outbound Utilization = 7.57 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:32:09 Statistics
Outbound Utilization = 8.45 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:34:09 Statistics
Outbound Utilization = 8.24 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:36:09 Statistics
Outbound Utilization = 8.6 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:38:09 Statistics
Outbound Utilization = 7.33 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:40:09 Statistics
Outbound Utilization = 11.73 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:42:09 Statistics
Outbound Utilization = 12.26 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:44:09 Statistics
Outbound Utilization = 31.72 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:46:09 Statistics
Outbound Utilization = 50.93 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:48:09 Statistics
Outbound Utilization = 73.28 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:50:09 Statistics
Outbound Utilization = 91.95 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:52:09 Statistics
Outbound Utilization = 11.99 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:54:09 Statistics
Outbound Utilization = 12.87 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:56:09 Statistics
Outbound Utilization = 10.45 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 12:58:09 Statistics
Outbound Utilization = 6.69 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:00:09 Statistics
Outbound Utilization = 10.6 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:02:09 Statistics
Outbound Utilization = 12.13 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:04:09 Statistics
Outbound Utilization = 9.71 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:06:09 Statistics
Outbound Utilization = 9.93 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:08:09 Statistics
Outbound Utilization = 11.02 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:10:09 Statistics
Outbound Utilization = 7.59 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:12:09 Statistics
Outbound Utilization = 14.83 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:14:09 Statistics
Outbound Utilization = 32.31 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:16:09 Statistics
Outbound Utilization = 49.86 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:18:09 Statistics
Outbound Utilization = 62.03 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:20:09 Statistics
Outbound Utilization = 77.16 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:22:09 Statistics
Outbound Utilization = 10.3 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:24:09 Statistics
Outbound Utilization = 9.56 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:25:09 Statistics
Cannot read attributes.
MCC Error = 0, T1.0.0,
%MCC-S-SPECIALIZED_EXC, success with
management module-specific exception
reply,
Show,
Node4 rtgar3 Circuit rtrom ,
Show,
Node4 rtgar3 Circuit rtrom ,
Node not currently accessible.,
PersistenceClass = Temporary Problem
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:28:09 Statistics
Outbound Utilization = 0.56 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:30:09 Statistics
Outbound Utilization = 0.35 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:32:09 Statistics
Outbound Utilization = 0.21 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:34:09 Statistics
Outbound Utilization = 1.57 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:36:09 Statistics
Outbound Utilization = 1.42 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:38:09 Statistics
Outbound Utilization = 1.51 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:40:09 Statistics
Outbound Utilization = 1.55 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:42:09 Statistics
Outbound Utilization = 3.69 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:44:09 Statistics
Outbound Utilization = 17.3 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:46:09 Statistics
Outbound Utilization = 30.23 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:48:09 Statistics
Outbound Utilization = 43.42 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:50:09 Statistics
Outbound Utilization = 57.27 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:52:09 Statistics
Outbound Utilization = 7.11 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:54:09 Statistics
Outbound Utilization = 7.83 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:56:09 Statistics
Outbound Utilization = 6.51 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 13:58:09 Statistics
Outbound Utilization = 7.81 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 14:00:09 Statistics
Outbound Utilization = 4.86 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 14:02:09 Statistics
Outbound Utilization = 5.95 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 14:04:09 Statistics
Outbound Utilization = 7.24 Percent
Node4 rtgar3 Circuit rtrom
AT 22-APR-1991 14:06:09 Statistics
Outbound Utilization = 7.85 Percent
-------------------------------------------------------
The same result on an other router and another circuit:
_______________________________________________________
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:30:42 Statistics
Inbound Utilization = 5.56 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:32:41 Statistics
Inbound Utilization = 87.6 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:34:41 Statistics
Inbound Utilization = 7.28 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:36:41 Statistics
Inbound Utilization = 10.06 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:38:41 Statistics
Inbound Utilization = 5.63 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:40:41 Statistics
Inbound Utilization = 2.55 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:42:41 Statistics
Inbound Utilization = 18.59 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:44:41 Statistics
Inbound Utilization = 36.35 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:46:41 Statistics
Inbound Utilization = 54.17 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:48:41 Statistics
Inbound Utilization = 74.2 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:50:41 Statistics
Inbound Utilization = 103.99 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:52:41 Statistics
Inbound Utilization = 90.22 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:54:41 Statistics
Inbound Utilization = 90.71 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:56:41 Statistics
Inbound Utilization = 93.4 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 11:58:41 Statistics
Inbound Utilization = 80.37 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:00:41 Statistics
Inbound Utilization = 95.14 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:02:41 Statistics
Inbound Utilization = 94.06 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:04:41 Statistics
Inbound Utilization = 95.87 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:06:41 Statistics
Inbound Utilization = 93.46 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:08:41 Statistics
Inbound Utilization = 91.31 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:10:41 Statistics
Inbound Utilization = 15.19 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:12:41 Statistics
Inbound Utilization = 199.27 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:14:41 Statistics
Inbound Utilization = 381.59 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:16:41 Statistics
Inbound Utilization = 562.4 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:18:41 Statistics
Inbound Utilization = 746.27 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:20:41 Statistics
Inbound Utilization = 892.99 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:22:41 Statistics
Inbound Utilization = 96.86 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:24:41 Statistics
Inbound Utilization = 94 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:26:41 Statistics
Inbound Utilization = 76.55 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:28:41 Statistics
Inbound Utilization = 93.48 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:30:41 Statistics
Inbound Utilization = 92.52 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:32:41 Statistics
Inbound Utilization = 87.55 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:34:41 Statistics
Inbound Utilization = 94.86 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:36:41 Statistics
Inbound Utilization = 94.06 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:38:41 Statistics
Inbound Utilization = 88.12 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:40:41 Statistics
Inbound Utilization = 20.91 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:42:41 Statistics
Inbound Utilization = 210.82 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:44:41 Statistics
Inbound Utilization = 396.5 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:46:41 Statistics
Inbound Utilization = 583.34 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:48:41 Statistics
Inbound Utilization = 774.47 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:50:41 Statistics
Inbound Utilization = 915.31 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:52:41 Statistics
Inbound Utilization = 95.3 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:54:41 Statistics
Inbound Utilization = 92.11 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:56:41 Statistics
Inbound Utilization = 94.31 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 12:58:41 Statistics
Inbound Utilization = 69.88 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:00:41 Statistics
Inbound Utilization = 93.81 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:02:41 Statistics
Inbound Utilization = 93.76 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:04:41 Statistics
Inbound Utilization = 93.22 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:06:41 Statistics
Inbound Utilization = 93.95 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:08:41 Statistics
Inbound Utilization = 94.36 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:10:41 Statistics
Inbound Utilization = 27.99 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:12:41 Statistics
Inbound Utilization = 200.84 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:14:41 Statistics
Inbound Utilization = 371.84 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:16:41 Statistics
Inbound Utilization = 555.53 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:18:41 Statistics
Inbound Utilization = 739.37 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:20:41 Statistics
Inbound Utilization = 934.62 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:22:41 Statistics
Inbound Utilization = 94.58 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:24:41 Statistics
Inbound Utilization = 95.97 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:26:41 Statistics
Inbound Utilization = 1.72 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:28:41 Statistics
Inbound Utilization = 4.23 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:30:41 Statistics
Inbound Utilization = 5.54 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:32:41 Statistics
Inbound Utilization = 26.12 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:34:41 Statistics
Inbound Utilization = 7.48 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:36:41 Statistics
Inbound Utilization = 58.89 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:38:41 Statistics
Inbound Utilization = 46.09 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:40:41 Statistics
Inbound Utilization = 2.58 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:42:41 Statistics
Inbound Utilization = 113.94 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:44:41 Statistics
Inbound Utilization = 304.99 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:46:41 Statistics
Inbound Utilization = 492.15 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:48:41 Statistics
Inbound Utilization = 671.25 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:50:41 Statistics
Inbound Utilization = 854.42 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:52:41 Statistics
Inbound Utilization = 93.81 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:54:41 Statistics
Inbound Utilization = 90.33 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:56:41 Statistics
Inbound Utilization = 89.86 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 13:58:41 Statistics
Inbound Utilization = 91.49 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 14:00:41 Statistics
Inbound Utilization = 92.48 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 14:02:41 Statistics
Inbound Utilization = 94.24 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 14:04:41 Statistics
Inbound Utilization = 94.45 Percent
Node4 rt6bo Circuit fnal
AT 22-APR-1991 14:06:41 Statistics
Inbound Utilization = 94.96 Percent
|
| 867.9 | timer coincidence??? | JETSAM::WOODCOCK | | Wed Apr 24 1991 16:58 | 18 |
| Anwar,
Please note this user has a circuit counter timer of 1800 (30 min.).
Also note that approx. each 30 minutes for each of his log files there
is an upward trend of utilization. The second example more pronounced
because it actually is passing a greater % of traffic. Why this upward
trend lasts about 10 minutes doesn't seem to make sense but this could turn
out to be more than mere coincidence.
re:-.1 Try changing your circuit counter timer to 65000 and rerun your
test while noting any counter reset times. Then check to see if anything
unusual happens around the reset time. I also have a com procedure which
can calculate the % utilization (without doing any normalization of data)
for intervals very similar to what MCC uses. If you would like to use it
as a comparison test contact me via mail.
best regards,
brad...
|
| 867.10 | zeroed counter is a prob | JETSAM::WOODCOCK | | Fri May 31 1991 20:01 | 15 |
| Is there a status of this problem. I have also just reproduced this same
type of error while working on a different issue. This is not specific to
utilization as the problem occured while obtaining packet rates. The counters
had automatically zeroed during a command to calculate PR each 2 minutes.
The PR was between 200-300 pps for many minutes. After the counters zeroed
PR dropped to 13 pps then stepped up about 300 pps each interval until it
reached 1400 pps about 10 minutes later. At this time PR resumed normal
readings from 200-300 pps. No I don't believe this was normal traffic. MCC
appears to have a very serious problem in dealing with counters zeroeing.
If you need more definitive proof I'm reasonably confident I could recreate
it. Has this been QARed? I looked but didn't see one. The importance of this
module being accurate is CRITICAL.
best regards,
brad...
|
| 867.11 | REPORTS: DETAILED RESULTS | ROM01::NINNI | Don't worry...be happy | Fri Jun 14 1991 08:05 | 121 |
|
Hi
Finally we have a complete analysis of DECmcc Performance Analyzer; we
have applied this analysis to the customer site of INFN (Istituto
Nazionale di Fisica Nucleare), an italian strategic account, that is
evaluating the pheasibility to use DECmcc like its network management
platform. Of course the Perfomance Analyzer results are very important
for their purposes.
Their network is related with the other research customers at italian
level and european level (HEPNET network) and so if we convince INFN of
the value-added of our solution, it will be possible an extension of the
DECmcc usage to all the research network.
The network is a DECnet-IP network.
Thanks in advance for your help
Giuseppina
=========================================================================
Some analysis's results about the DECmcc Performance Analyzer
Functional Module at INFN customer site
The test was executed on a WAN link at 768 Kb/s speed; if A and B are
the nodes and C their common circuit (like in the picture)
-------- C=768 Kb/s ----------
| A |---------------| B |
-------- ----------
we collected the following datas, using a DCL procedure, in order to
compare the traffic information provided by MCC with the same provided
by the NCP circuit's counters:
1. with NCP, the counters of the C circuit from the node A;
2. with MCC, the INBOUND UTILIZATION of the C circuit from node A
(Statistics duration: 1 minute)
3. with NCP, the counters of the C circuit from the node A;
4. with NCP, the counters of the C circuit from the node B;
5. with MCC, the OUTBOUND UTILIZATION of C circuit from the node B;
(Statistics duration: 1 minute)
6. with NCP, the counters of the C circuit from the node B.
The datas were collected by a batch procedure during 18 hours.
The utilized procedure is:
=======================================================================
$ NCP:=="$SYS$SYSTEM:NCP"
$ again:
$ ncp tell rt4bo sho circ tdm-cern counter to -
dkb300:[examples_cnaf]statistics.dat
$ manage/ent
show node4 rt4bo circuit tdm-cern inbound utilization, for dura 0:1:0, -
to file dkb300:[examples_cnaf]tttt.1
exit
$ ncp tell rt4bo sho circ tdm-cern counter to -
dkb300:[examples_cnaf]statistics.dat
$ append dkb300:[examples_cnaf]tttt.1 dkb300:[examples_cnaf]statistics.dat -
/new_version
$ ncp tell cernit sho circ infn counter to -
dkb300:[examples_cnaf]statistics.dat
$ manage/ent
show node4 cernit circuit infn outbound utilization, for dura 0:1:0, -
to file dkb300:[examples_cnaf]tttt.2
exit
$ ncp tell cernit sho circ infn counter to -
dkb300:[examples_cnaf]statistics.dat
$ append dkb300:[examples_cnaf]tttt.2 dkb300:[examples_cnaf]statistics.dat -
/new_version
$ delete dkb300:[examples_cnaf]tttt.*;*
$ goto again
============================================================================
The INBOUND UTILIZATION of the node A and the OUTBOUND UTILIZATION of
the node B should have quite the same value, with some differences due
to the little difference of the collections' time.
The obtained results were elaborated in order to have both the graphics
displays of the the INBOUND/OUTBOUND UTILIZATION, done by MCC and the
graphics of the INBOUND/OUTBOUND UTILIZATION done with the circuit
counters of NCP.
We used the formula, which is in the Performance Analyzer Function Module Use
Manual , at pag. A-6.
The graphs display the progressive number of the collections like X-axis
and the obtained percentuage like Y-axis.
The counters of C circuit on the node A are zeroed every 30th minutes,
while the counters of the same circuit C on the node B are zeroed every
4 hours; both the zeroings are done by external procedures (NCP ZERO
circuit-id), without the utilization of the "circuit counter timer"
parameter.
We examined the obtained graphs, (with the determination of the zeroing
instants of the circuit counter and their relevations on the X-axis) and we
observed the following results:
1)A period of quite 10 minutes exists (immediately after the
counters' zeroing) during which MCC produces wrong information. Out of
these periods, the values provided by MCC are quite the same of those
otained from the NCP circuit counters.
If you need the graphs, we can send to you. Anyway you can obtaine them
easily
Bologna, 14/5/1991
Davide Salomoni (INFN contact)
Giuseppina Ninni (DEC EIS contact)
|
| 867.12 | Being investigated | TOOK::ANWARUDDIN | Anwar | Fri Jun 28 1991 19:10 | 3 |
| The problem of the zeroing of counters and its effect on statistics
is being investigated. We hope to fix this problem in the next release
of DECmcc.
|
| 867.13 | observation | JETSAM::WOODCOCK | | Tue Jul 30 1991 17:37 | 33 |
|
> The problem of the zeroing of counters and its effect on statistics
> is being investigated. We hope to fix this problem in the next release
> of DECmcc.
Hi there,
I just happened to be burning up calculator batteries to try to determine
the scope of this inaccuracy and figured I'd just post an observation. I'm
polling counters hourly and zeroed the counters at a couple different times
with relation to the polling. The first time the counters were zeroed about
49 minutes into the poll cycle. MCC stats were reported very low compared
to reality for the interval when zeroed occured. What appears to be a
possibility of happening is that MCC took the counters (which only had 11
minutes worth of data) and divided by the user specified duration (60 min.).
I took a ratio of 49/60 and applied this to the stats and it looked to be
in the ballpark. The interval after this zeroing also proved to be higher
than what should have been. Not sure why but maybe it took the extra 11 minutes
worth of data and plugged it into the formula along with the normal 60 minutes.
The second test I zeroed the counters within a couple minutes after the
polling cycle. This proved to be pretty close to accurate (within 5%) even
at the zeroed interval and the one following.
It appears that the closer the zeroing occurs *after* the poll cycle the
more accurate the stats are. Because there is no _reasonable_ way of keeping
polling and zeroing syncronized we will always see a level of inacuracy. At
this time with a maximum COUNTER TIMER settable by the routers of about 65000
(18 hrs) and a poll rate of 1 hour we'll receive two bogus reports each 18
hours. 11% inaccuracy minimum at these rates....not good :-(.
best regards,
brad...
|
| 867.14 | Strange, ... | MAYDAY::ANDRADE | The sentinel (.)(.) | Fri Aug 30 1991 06:51 | 10 |
| Doesn't MCC ever look at the seconds since last zeroed counters.
MCC polling would be that much more robust, if it used this data.
If the counters are zeroed then it should know enough not count the
missing time. On the other hand, if its MCC that misses a few polls,
then it should be able to fill in for the missed polls. I mean 65536
seconds equals 18.2 hours, so you could miss a max of 18 hourly polls
and still get the averages out of the data correctly.
Gil
|