Zhen He

Home

Biography

Publications

Research Grants

PhD thesis

Teaching

Spark related topics

Zhen He
Department of Computer Science
Australian National University

Integrated Buffer Management for Object-Oriented Database Systems

Despite the dominance of relational database management systems (RDBMS) in the database market, object-oriented database management systems (OODBMS) continue to play an important role in complex data management. Complex data are often found in telecommunications, business, engineering and web based applications. The most common style of accessing complex data is navigation. Navigational access can generate excessive disk IO because objects in the path of navigation maybe placed in different disk pages. Excessive disk IO is becoming increasingly undesirable because disk IO performance improves at only 5-8% per year whereas CPU performance doubles approximately every 18 months. Thus disk IO is likely to be a bottleneck in an increasing number of OODB applications. This thesis focus on reducing disk IO effects to improve OODBMS performance.

Effective main memory buffer management is the key to reducing the disk IO bottleneck in OODBMSs. There has been much existing work, namely in the areas of: static clustering; dynamic clustering; buffer replacement; and prefetching. All of these techniques can be used together in a complimentary manner. Most existing research has focused on finding the best solution for each area with little regard on how solutions from the different areas affect each other. We believe synergy exists between the areas, and that exploiting the synergy leads to the best overall solution. This thesis focus on demonstrating synergistic techniques are both feasible to implement and outperform their non-synergistic counterparts.

We made general modifications to existing techniques to demonstrate the superiority of synergistic buffer management. There were three guiding principles behind our modifications: synergy; generality; and simplicity. Synergy refers to modifications that exploit synergies between the different buffer management areas. Generality refers to modifications that can be applied to a large range of existing algorithms. Simplicity refers to making the modifications easy and straightforward to apply. Following these guiding principles we developed three new frameworks: opportunistic prioritised clustering framework (OPCF); cache conscious clustering framework (C3); and path and cache conscious prefetching framework (PCCP). Each framework addresses the synergy between two different buffer management areas. Using the frameworks we developed a total of seven new buffer management algorithms. These algorithms were found to outperform existing algorithms in a variety of situations.

This thesis takes a first exploratory look into how OODBMS buffer management techniques can be enhanced by synergistic modifications. The preliminary results show that there is much potential in this approach and suggests that perhaps the next big breakthrough in improving OODBMS main memory buffer performance lies in such techniques.