Since the
optimization only applies for requests with the same deadline, its efficiency depends on how often it can be applied ( i.e., how many requests have the same or a similar deadline). To
increase this probability, the following tricky technique can be used: all reequests have release times that are multiples of the period p. Hence, all requests have deadlines that are
multiples of the period p. Therefore, the requests can be grouped together and be served accordingly. For requests with different data rate requirements, in additon to SCAN-EDF,
the employment of a periodic fill policy is proposed to let all requests have the same deadline. With this policy, all requests are served in cycles. In every cycle, each request gets
an amount of service time that is proportional to its required data rate. The cycle length is eqqual to the sum of the service times of all requests. Thus, in every cycle, all requests can be
given a deadline at the end of the cycle.
