Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

frame-relay traffic shapping question

Dear all.

i have a question related to the frame-relay traffic shapping question.cir is the commited rate. bc is the traffic which is sent during every ime interval. be is the excess burst.

example. i have line 128kbps to service provider. 64 is my cir so in the map class frame-relay cir 64000 and bc is 8000 which means 8000 bits will go every interval which is by default about be. excess burst will go only in the first interval or every interval ? if i say frame be 32000. will 4000 additional bits go in every interval ?

please advice .


Re: frame-relay traffic shapping question

Here's my summary of shaping operation:

* The token bucket used for shaping has a size of Bc+Be (if Be is zero, this equals Bc)

* The token bucket starts off full i.e. with a capacity of Bc+Be

* At the start of every Tc interval, Bc worth of tokens are added to the bucket. If the number of added tokens would result in the capacity of the bucket exceeding Bc+Be, these excess tokens are discarded.

* After the tokens are added to the bucket, the shaper checks if there are any packets that are waiting to be transmitted. If there are, an equivalent number of bytes of packets as the token bucket contents are transmitted. If there are more packets than tokens, these packets will have to wait until the next Tc.

* The capacity of the token bucket will approach Bc+Be only of the traffic has been quite for a little while.

* The maximum amount of traffic that can be sent in any interval is Bc+Be, since that is the size of the token bucket. If the bucket has filled up to Bc+Be, this implies that the connection has been quiet for a little while so that we can allow it a little burst.

* The long-term average of traffic will never exceed CIR, since the token bucket is filled at a CIR rate.

Now, applying this to your example, the maximum size of the token bucket will be Bc + Be = 40000 bits. The bucket will fill up to this level only if there are no packets in the queue for a little while, allowing the bucket to fill up. Then, say that that a 1500 byte (12000 bits) packet comes along. It will be transmitted and the token bucket size reduced to 28000 bits. If there are more packets, they will be transmitted as long as the token bucket has sufficient packets. If there are no more packets, the token bucket will be filled with another 8000 bits (Bc) at the next measurement interval, giving a total of 36000 bits. Whether or not all the Be + Bc bits will be used up in the first interval depends on how many packets are waiting in the queue. If there are sufficient packets, they can all get used up. If not, they get carried forward to the next interval.

Pls do remember to rate posts.