Key Techniques Study On Service Performance And Resource Scheduling Of Smart TV Operating System

With the development of the integration of telecommunications networks, cable TV networks and the Internet , various integration services and terminals emerge continually. Smart TV becomes the new research focus in the integration terminal fields, thus attracting the attention of researchers, service providers and television manufacturers. However, due to the limited nature of hardware resources in embedded system, Smart TV cannot fulfill resource requirements of every concurrent service. In order to guarantee the qualities of concurrent services and achieve better overall performance, it is crucial to optimize resource management and scheduling in Smart TV. Performance optimization can be done in two layers:System layer. It is necessary to optimize the management and scheduling of the resources and make appropriate allocations among tasks.Application layer. It is necessary to consider the characteristics of the terminal to optimize the application logic design, thus improving application capabilities and efficiencies of the resources usage.Under the support of the National Key Technology R&D Program new generation of radio and television service system supporting cross-regional and multi-service-providers , this dissertation will discuss several key issues related to Smart TV performance optimization based on the two-layers division above, including the design of system resource management framework, resource allocation algorithm, multi-media file transfer protocol and streaming media rate smoothing etc., in order to improve the capabilities of Smart TV services which result in better contributions and innovations are as follows:1) Recognizing the lack of effective means to guarantee QoS of Smart TV services under restricted resources, a QoS model, which establishes the mapping between QoS levels and resource requirements of services, is proposed. Based on the QoS model, a QoS guarantee resource management framework is built, which distinguishes different types of resources, allocates and schedules resources in order t QoS. The framework uses Linux DLL to realize dynamic loading and replacement of algorithms, and has a good expansibility. Finally, a priority-based resource scheduling algorithm is presented to guarantee the QoS of concurrent services. 2) A utility based resource allocation model with the objectives of system utility maximization and resource consumption minimization is established for Smart TV operating system resource allocation. This model is planned as multiple-choice and multiple-dimension knapsack problem, and then a heuristic algorithm RA_GHEU is proposed. This algorithm prefers to choose allocation schemes with high utility per resource unit, and seeks for optimization from both system utility aspect and resource consumption aspect simultaneously. Compared with back-tracing algorithm with a time complexity of O(Nn), RA_GHEU whose time complexity is O(n2), can obtain near-optimal solution in an extreme short time, and is suitable for smart TV operating system real-time resource allocation.3) A transport protocol SMFTP for the large multi-media file transfer on the Smart TV is proposed. Built on top of UDP, SMFTP uses a uniform retransmission strategy to handle error recovery, thus avoiding retransmissions during the process of data transfer and reducing the cost of timer and pointer addressing on the receiver side. The sender uses DAIMD (Additive Increase Multiplicative Decrease with Decreasing Increases) rate control algorithm to adjust sending rate in time according to periodic feedbacks from the receiver, which improves responsiveness to changes in network and terminal resources. SMFTP can ensure the efficiency of transmission with a low resource requirement to the receiver. Experimental results show that in the network environment with bandwidth of 100Mbps, compared with the FTP and UDT protocols, SMFTP file transfer time was reduced by 17% and 3%, respectively, CPU usage 42.3% and 50.3% , and memory occupation ratios 4% and 15%.4) For the overflow problem in VBR traffic burst on Smart TV, a bandwidth smoothing algorithm RBCBS is proposed for transmitting VBR streams smoothly subjected to maximum data transfer rate and buffer size of client, in order to guarantee the QoS of streaming media service with limited resources of Smart TV. The algorithm tries to transmit data as fast as possible with the constraints of transfer rate and buffer size, so as to take full use of the available buffer resource. The algorithm s time complexity is O(N). Compared with traditional RCBS (Rate-Constrained Bandwidth Smoothing) algorithm, the stream traffic of RBCBS reduces rate change numbers by 11.9% and ratio of rate change by 12.6%, so the data transmission is more stable and the pressure on the terminal is reduced.