[e2e] Token/slotted access based mechanisms
craig at aland.bbn.com
Wed Oct 15 10:36:45 PDT 2003
I have this vague recollection that Claudio Topolcic, Winston Edmonds and
Karen Seo had to solve this problem for the wideband satellite network (which
was used for real-time conferencing) in the 1980s. There was a paper on
it at ACM SIGCOMM 1988 -- don't know if it talked about the slot allocation
In message <032201c39335$c0596660$fb01a8c0 at aparabgl.com>, "Alok Dube" writes:
>I am not sure if this is the right mailing list to ask this question
>(primarily I dont see any mailing list where these topics are discussed and
>this seemed the closest), but incase someone can point me to the correct
>mailing list/documents to lookup, it would be great.
>Consider a slotted access/token based access mechanism.
>Multiple nodes connect to the token based access mechanism.
>The total number of tokens/slots is finite..say X
>A node is given access to the slots based on say, its bandwidth requirements
>(i get 1 slot/tokens or 2 slots/tokens etc).
>Essentially X decides the sampling rate from the slot/token based "sampler"
>into the network/across the network.
>now is there any way that one can allocate "evenly spaced" slots to the
>It is more to the tune of, if a node needs 2 slots, it should get slots
>which are spaced apart by a distance of X/2, if a node needs 3 slots it
>should get slots spaced apart by X/3 etc... the "X" is variable and it can
>be decided in such a manner that all requirements fit into the need.( I
>would call X the envelope size).
>Is there any work done somewhere in this direction?
More information about the end2end-interest