Replication schemes for opportunistic networks with impatient users

As the number of mobile users increases, and cell phones become more powerful, delivering multimedia content to them using a centralized infrastructure becomes both expensive and inadequate. Here we study an alternative solution, which leverages local dedicated caches on these devices to opportunistically fulfill other user requests, in a peer-to-peer manner. The problem we study is that of choosing the allocation of content items to these caches, to fulfill the demand of all users in a timely manner. We show that the allocation´s efficiency itself is determined by a previously overlooked factor - the impatience of content requesters - which leads to completely different allocations being optimal as content requesters impatience follows different delay-utility functions. Moreover, although no global cache state can be maintained in such opportunistic environment, we show for a homogeneous network that the optimal cache allocation can be attained by simple reactive replication algorithms that use only local knowledge.