We consider an infinite-capacity storage system. The cumulative input
to the system is assumed to be either (a) a non-decreasing Levy proces
s or (b) an integrated continuous-time Markov chain. Reward accumulate
s at a rate depending on the instantaneous release rate. The objective
is to choose the release rule in such a way as to maximize the expect
ed total discounted return. In this note we show how to determine the
expected discounted return when the release rate is either constant or
a linear function of the content.