基於分散式雜湊表的視頻資源多屬性查詢策略

Abstract

基於分散式雜湊表的資源組織機制具有較強的可擴展性和較快的查詢速度，但它在進行多屬性查詢時，會產生較大的查詢開銷和較低的查詢速度。為此，本文提出了基於分散式雜湊表的環狀域機制，採用兩層資源組織結構以支援高效的多屬性查詢，在上層利用分散式雜湊表對單個關鍵字定位具有Log2N時間複雜度的優點，以及視頻資源的分類資訊較為穩定的特點，採用改進的分散式雜湊表將多個視頻資源類以環狀方式有效地組織起來；在下層根據視頻資源的詳細描述資訊採用擴展域的方式進行組織。在此基礎上提出了最小屬性組優先的多屬性查詢策略。通過模擬實驗表明此查詢策略在3個屬性查詢的平均查詢時間複雜度上比關鍵字融合策略縮短12%，比逆雜湊表策略縮短33%。Video resource organization mechanism is the foundation of providing high-quality video service in network. The resource organization mechanism based on distributed hash table that provides high scalability and fast query speed is very common currently, but it results in larger query cost and lower speed in multi-attributes query. This paper proposes a hierarchy ring-region resource organization mechanism based on distributed hash table, in which we have adopted a two-layer structure to meet the requirements of multi-attributes query. In the upper layer, because single attribute locating in distributed hash table has time complexity of Log2N, and the classified information of video resources is stable, we adopted advanced distributed hash table to organize video resources. In the lower layer, we adopted extend regions to organize video resources flexibly in the resource class according to the detailed description information. Based on this structure, this paper presents a minimum attribute group first multi-attribute query algorithm. The simulation results demonstrate that our algorithm saves 12% average query time complexity than the Keyword Fusion algorithm and 33% than the Inverted Hash Table algorithm in three attributes query.