System supports for data syndication service : high performance aggregation and archiving of frequently changed distributed data = 데이터 신디케이션 서비스를 위한 시스템 : 빈번하게 변경되는 분산 데이터의 고성능 통합 처리와 기록
high performance aggregation and archiving of frequently changed distributed data
Data Syndication Service (DSS) which aggregates information from different data sources such as sensors or Web sites, composes the information into a value-added and refined one, and disseminates them to its consumers is becoming popular. DSS is essential where there are a lot of data sources and the frequency of data updates are very high. In such domains, without DSS, people will experience difficulties in acquiring refined information from widely-distributed lots of data sources. For instance, without price comparison sites which are one of the most popular DSS, people who want to buy an item from online shops have to spend a few hours to find the shopping site which proposes the lowest price. In spite of the importance, DSS has not been much studied yet. In this thesis, we study the high performance architecture for DSS systems. First, we identify the functional requirements of DSS. Then, we discuss the high performance architectures of DSS system in Web and sensor network domains. According to the practical needs of each domain, content aggregation is focused in Web domain and content archiving is in sensor network domain. We also describe the prototype implementation as well as technical issues. We also measured the performance of prototype system to show the validation of our approach. With the proposed system, people who want to provide DSS provider can build up a new service without involving the details of complicated content aggregation procedures, and thus concentrate on developing the service logic. We demonstrate this by constructing a DSS Web site using the prototypes.