III: Small: Automatic Incremental Design for Next-Generation Database Systems

点击翻译按钮获取中文摘要

Data management systems are undergoing a sea change. Specializedengines with specialized physical storage structures are emerging toaddress the extreme data volume and performance requirements of modernapplications. At the same time, the complexity of these systems, theapplications in which they are used and the platforms on which theyare built is increasing. Thus, there is a growing need for automaticdatabase design for these emerging database systems. Unfortunately,previous work in automatic design cannot be used directly. While theconceptual framework of previous research is useful, the specificsmust be reworked in order to adequately take advantage of theopportunities and address the challenges that these new structurespresent. Furthermore, most design tools only create complete designs.There is also a need for automatic designers that can produce a newdesign that is sensitive to the cost of migrating an old design to anew one. Such an incremental designer would often generate asub-optimal design if that design will produce 80% of the benefit with20% of the work.The PIs propose to investigate this incremental automatic designparadigm for newly emerging database systems. Specifically, theproject addresses three data management platforms: column stores forOLAP, main memory, cluster-based systems for OLTP, and an extension torow stores for exploiting correlations in data attributes.The PIs expect to extend the work on current design tools bydemonstrating workable incremental designers as described above. Thereis a strong need for such tools, and if this research is successful,it should enable successful deployment of many new-style datamanagers. Moreover, incremental design ideas based on soundcost-benefit analysis are applicable to other data-intensive computingenvironments and constitute an important direction towards trulyautonomic computing. Further information on the project can be foundon the project web page: http://database.cs.brown.edu/projects/auto/