Rivisiting Token/Bucket Algorithms in New Applications

Abstract

We consider a somehow peculiar Token/Bucket problem which at first sight looks confusing and difficult to solve. The winning approach to solve the problem consists in going back to the simple and traditional methods to solve computer science problems like the one taught to us by Knuth. Somehow the main trick is to be able to specify clearly what needs to be achieved, and then the solution, even if complex, appears almost by itself.