Mobile multimedia content delivery is a recent development that merges multimedia technology with the freedom and benefits offered by wireless network access to mobile/nomadic users. The technological push for this convergence stems from the immense user demand for mobile access having the same level of service "on the move" as that found on desktops. Yet, the convergence does not come without problems. On the one hand, the mobile environment has fluctuating levels of network service provision, periods of disconnection and dynamic environmental changes. On the other hand the applications are often executed on devices with limited resources available such as memory, battery power and CPU speed. In addition, if we also consider that media streaming requires bandwidth, storage and continuity, this convergence is a hard task to accomplish. In this paper, we propose a novel caching architecture for content delivery to mobile devices in an attempt to provide efficient mobile user support for interaction with newly discovered services and uninterruptible/continuous service provision.