| Summary: | With the stupendous rate of the development in network fields, more and more network based software comes into our eyesight. This thesis mainly focuses on how to build a data integration and synchronization system in order to support a distributed Web OS (Operation System) working properly with huge database on different servers. Jupiter Visualiser by Polystar OSIX is a process and analysis tool that generates customer and business intelligence information for telecom operators and puts it at their fingertips. Jupiter Visualiser does this by collecting raw network information via real-time probes and combining it with reference data from third-party sources. This thesis describes the progress of designing and implementing a sub project which is based on Jupiter Visualiser. With the purpose to achieve high availability and low coupling in fetching data from massive storage, Sub project VSM is born as middleware to handle over this work instead of making queries between Jupiter Visualiser and MySQL directly. On the other hand, efficiency optimization and coding structuring are also important facts within discussion in second half. Some measurements are made to test the strength and fault tolerance of VSM and Jupiter Visualiser code. In order to have a web OS with both high quality and efficiency, much work have been done as well such as algorithm optimization, interface simplification and buffering cache design and so on. With the help of Jupiter Visualiser team OSS team and Product Manager, a complicated work was done successfully within several months. Although there would be many things left to improve and modify, this thesis still provide a practical satisfying way to integrate web OS with massive level database.
|
|---|
