Skip to Main Content
This letter elaborates on a scheme that combines batch patching at an origin server and prefix/interval caching at an edge server receiving the clients' requests. We derive a cost function that factors in the aggregate backbone rate, the cache occupancy, and the disk bandwidth utilization. We define the optimal batched patch caching strategy as a function of the client request rate. Finally, we show how various strategies including full caching, no caching and pure prefix caching with no patching are optimal derivations of our scheme under different request rates. We demonstrate the benefits of our scheme compared to classical streaming strategies.