Currently, the primary legal interpretation approving cloud DVR in cablevision decision relies on individual copies for each video recorded. This means every recorded video in the cloud needs its own copy in the data center. With totally 20 billions of videos available to around 50 millions subscribers, how to manage and store the big data efficiently is required for the datacenter. In this paper, an asynchronous service system with categorizing users by cloud DVR usage is proposed for this purpose. In this system, cloud DVRs with similar requests are constructed in one cluster. The cloud recording service on this cluster becomes idle after a while of no requests. New arriving requests are buffered when the service wakes up. Cloud DVRs accept two requests: cable video recording and Internet video recording. Besides proposing the architecture, we also present the challenges that have to be solved before this strategies could be applied in industries.