We've Moved!
Visit SDSU’s new digital collections website at https://digitalcollections.sdsu.edu
Description
The growing popularity of "cloud computing" resulted in shifting a large number of internet applications to the cloud. This paradigm shift demands for a more sophisticated data management services at the cloud. Hence a challenge is to build a consistent, available and scalable data management system capable of serving petabyte of data for millions of users. One of the factors that categorize the data management system to be consistent, available and scalable is the data storage structure used to store the data. In recent years, there has been an increase in temporal and spatial data in database applications and the interval type is adopted by more and more database systems. Interval type being one of the simplest complex data type, the database implementers introduced corresponding management functions by allowing users to define the required complex types with appropriate operations. The background work for this thesis consists of a mobile social networking device capable of transferring transient, location based data to a central repository for long term storage and real time semantic analysis. With a potential for hundreds of daily transactions per individual, the need for efficient storage becomes imperative. This work presents the design and implementation of a novel data structure called a Relational Interval Tree, or RI-Tree, used to store transient, location based data captured by mobile social networking devices. A relational database schema is presented that defines the tables, fields, stored procedures and other elements of a RI-Tree. The transient data being the transactions between the two devices are of interval type, the RI-Tree allows for the efficient searching of device transactions that span a given time interval. A discussion on how an RI-Tree guarantees optimal complexity for storage space and I/O operations when updating or querying larger sets of device transactions is presented, with a focus on insertion and search time, storage, and operation complexity. Transient data captured from the social networking device are also sent to a cloud hosting a web portal, so that device users shall use to configure their personal, professional and basic information. The Web Portal facilitates the user to watch the profiles of people with whom their information is exchanged. Profile data exchanged by two or more device wearers that come into contact with one another is encapsulated in XML Schema compliant documents.