From: Michael Snyder (msnyder@revolutioncomputer.com)
Date: Tue Feb 17 2004 - 17:53:00 GMT-3
While not directly related.
Be very careful with CAR. The excess burst size can put your bucket
into debt for seconds at a time. Nothing like throwing some high number
into the excess burst value to be safe, and having the interface
effectively flap every few seconds.
-----Original Message-----
From: braet_kamiel@nl.ibm.com [mailto:braet_kamiel@nl.ibm.com]
Sent: Tuesday, February 17, 2004 12:20 PM
To: ccielab@groupstudy.com
Subject: Re: RE: FRTS question
The explanation as described below applies to both FRTS and GTS:
These Bc, Be and Tc variables have the following relations.
- Be and Bc determine the size of the token bucket.
- Tc and Bc determine the amount of tokens that is added to the bucket
as a function of time.
The token bucket and the tokens can be defined as follows:
- The token bucket can be seen as a buffer for tokens, the size of the
bucket defines how many tokens can be held.
- The amount of tokens in the bucket represent the amount of data that
can be sent at any given time.
That's all you really have to know. But based on this you can come up
with the following conclusions.
- Bc and Tc determine together the long-term MAXIMUM average data
stream.
- If the bucket fills to capacity Bc + Be, newly arriving tokens are
discarded. Therefore the bucket specifies the MAXIMUM amount of tokens
in the bucket, Bc and Be defines the maximum amount of data that can be
sent at any given time.
- Since Be among others defines the size of the bucket, Be defines how
many UNUSED tokens can be saved (or buffered) to use later on, and does
NOT have anything to do with Tc.
>From this you can say:
- The size of the bucket determines amount VARIATION a dataflow can
have, while still being able to utilize the long-term maximum average
data stream.
This archive was generated by hypermail 2.1.4 : Fri Mar 05 2004 - 07:13:50 GMT-3