This paper proposes a novel network architecture for optimal and balanced provision of multimedia services, exploiting a resource prediction system. This architecture enables for the long-term prediction of multimedia services future demands, based on the history of previous network resources usage. The proposed research approach provides the opportunity for the optimal distribution of streaming data, among Content Delivery Networks, cloud-based providers and Home Media Gateways. The short-term prediction that is performed, enables for making the proper decisions by the system, according to specific network metrics, towards achieving higher Quality of Service and Quality of Experience for the end users. The validity of the proposed system is verified through several sets of extended experimental simulation tests, carried out under controlled simulation conditions.