[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference azur::mcc

Title:DECmcc user notes file. Does not replace IPMT.
Notice:Use IPMT for problems. Newsletter location in note 6187
Moderator:TAEC::BEROUD
Created:Mon Aug 21 1989
Last Modified:Wed Jun 04 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6497
Total number of notes:27359

2060.0. "question on Statistics computation" by COL01::LUNT () Fri Jan 10 1992 05:01

Hello,


	I have  just finished pouring through all the notes on Statistics and
have found no answer to my question. The question is:

If I am issuing the following command:

Show node4 BYLV24 line isa-0 all statistics, for every 00:10:00 until 

	15-jan-1992:15:00 duration 00:09:58

and the manual says that the following formula is being use
to calculate the multiple collision percentage:

mcfs *100% =CSMA/CD contention with multiple collision
----
 fs

a. Does MCC then take a sample at 00:00:00 minutes and then again at
   00:09:58 minutes subtract the numbers (9:58min sample - 0min sample) and
   then compute the statistic? Or does MCC poll, say every minute and compute
   the statistic for every minute and then average it over the 9:58 minutes?
   (By the way this was asked in note 1304 but never answered)

b. Why do you only use Frames Sent(fs) and not a combination of frames Sent(fs)
   and Multicast Frames sent(mfs) i.e.  mcfs/(fs+mfs)? Or does the multiple
   collision frames counter only count multiple collisions for non-multicast
   frames?

Thanks in adance,

Julie Ann Lunt
Networks and Communications, Cologne Germany

T.RTitleUserPersonal
Name
DateLines
2060.1TOOK::ANWARUDDINAnwarTue Jan 14 1992 10:4758
Hello,


	I have  just finished pouring through all the notes on Statistics and
have found no answer to my question. The question is:

If I am issuing the following command:

Show node4 BYLV24 line isa-0 all statistics, for every 00:10:00 until 

	15-jan-1992:15:00 duration 00:09:58

and the manual says that the following formula is being use
to calculate the multiple collision percentage:

mcfs *100% =CSMA/CD contention with multiple collision
----
 fs

>>>>>>>> 
>>>>>>>> The attribute "CSMA/CD Contention..." is not defined for a 
>>>>>>>> Phase4 entity. So I presume you mean:
>>>>>>>> SHOW NODE bylv24 CSMACD STATION isa-0 ALL STATISTICS.
>>>>>>>> 

a. Does MCC then take a sample at 00:00:00 minutes and then again at
   00:09:58 minutes subtract the numbers (9:58min sample - 0min sample) and
   then compute the statistic? Or does MCC poll, say every minute and compute
   the statistic for every minute and then average it over the 9:58 minutes?
   (By the way this was asked in note 1304 but never answered)

>>>>>>>>> Depends on whether the data is "current", that is, if it is
>>>>>>>>> obtained directly from the entity or "past", that is, gotten
>>>>>>>>> from the repository. In the former case two samples are
>>>>>>>>> gotten (one at the beginning of the period and the other at the
>>>>>>>>> end of the period) and statistics computed. In the later case
>>>>>>>>> there can be more than two samples during the period of interest
>>>>>>>>> and these are used to calculate the change in counter values
>>>>>>>>> over the requested duration and then statistics are computed
>>>>>>>>> for the entire duration, not for every pair of samples and then
>>>>>>>>> averaged.

b. Why do you only use Frames Sent(fs) and not a combination of frames Sent(fs)
   and Multicast Frames sent(mfs) i.e.  mcfs/(fs+mfs)? Or does the multiple
   collision frames counter only count multiple collisions for non-multicast
   frames?

>>>>>>>>> "fs" is the total number of frames(including multicast) successfully 
>>>>>>>>> transmitted. This also includes frames transmitted after collisions 
>>>>>>>>> on previous attempts.

Thanks in adance,

Julie Ann Lunt
Networks and Communications, Cologne Germany