The Research Of Real-time Prefetching For Rail Transit

Wireless prefetching technology has received extensive attention because it significantly reduces content request delay and network congestion.Most of the existing prefetching schemes consider weak mobility,irregular network topology and random roaming scenarios,which are not consistent with the application scenarios of rail transit.In this article,we studied the prefetching problem in the rail transit scenario under the CRAN architecture.For its characteristics of linear topology,fixed movement route,relatively stable personnel and the limited computing power environment in rail transit scenarios,our goal is to minimize the average access delay,and the design is based on a real-time prefetching strategy which is triggered BBU-by-BBU(base-band unit).Firstly,based on the historical access information,current access content and content request rate of each passenger in the vehicle,all the content that passengers may access during the next BBU and their corresponding access probability are computed based on the Markov chain data access model.Secondly,based on the size,access probability,and cache size of BBU,the actual prefetched contents are determined and prefetched with the goal of minimizing the average access delay.We formalized the problem and proposed a heuristic solution.Further,we studied the special case in which all passengers have the same content request rate,and proved that the problem is the classic 0-1 knapsack problem,so that it can be solved by the classic dynamic programming algorithm with complexity().For the special case where all files have the same size,a faster algorithm with complexity ()is presented.Finally,we conducted simulation experiments on the proposed algorithm.The experimental results show that when the cache space size accounts for 20% of the total file set,the file prefetching hit rate reaches 60%;when the cache space size accounts for50% of the total file set,the prefetching hit rate of the file reaches 80%.The above simulation results reveals the effectiveness of the real-time prefetching algorithm in improving the average access delay.