Paul Larson's Home Page

Per-Åke (Paul) Larson

Microsoft Corporation
One Microsoft Way
Redmond, WA 98052-6399
Email: palarson@microsoft.com
Phone: (425) 703-6260
Fax: (425) 936-7329

I am a senior researcher in the Database Group at Microsoft Research, part of Microsoft. My research is in the area of databases, in particular mid-tier database caching, query processing and query optimization.

MTCache Project

Mid-tier database caching attempts to improve system throughput and scalability by offloading part of the query workload from a backend database server to front-end database servers that contain replicas of some of the data. MTCache is a prototype mid-tier database cache for Microsoft SQL Server, exploiting its support for indexed views, transactional replication and distributed queries. The fact that some queries are offloaded to an MTCache server is transparent to applications. This makes it possible to change caching strategy without having to change applications. The cached data is updated asynchronously and thus may not be completely up to date. A user query can specify the level of data freshness and consistency required and MTCache guarantees that the requirements will be satisfied. The figure on the right shows results from one scale-out experiment - throughput increases linearly with each MTCache server added.

Team: Paul Larson, Jonathan Goldstein, Jingren Zhou

Further information about MTCache:

	Per-Åke Larson, Jonathan Goldstein, Jingren Zhou: MTCache: Transparent Mid-Tier Database Caching in SQL Server. ICDE 2004: 177-189
	Hongfei Guo, Per-Åke Larson, Raghu Ramakrishnan, Jonathan Goldstein: Relaxed Currency and Consistency: How to Say "Good Enough" in SQL. SIGMOD 2004: 815-826
	Per-Åke Larson, Jonathan Goldstein, Hongfei Guo, and Jingren Zhou, MTCache: A Transparent Database Cache for SQL Server, Data Engineering Bulletin 27(2): 35-40 (2004).

Recent Professional Activities

	VLDB 2005: Program chair of Industrial Applications and Experience track
	ICDE 2005: Program committee member
	SIGMOD 2004: Program committee member
	VLDB 2004: Program committee member

Recent Publications

A more complete list of publications can be found here.

Amiga Rahal, Qiang Zhu, Per-Åke Larson: Evolutionary techniques for updating query cost models in a dynamic multidatabase environment. VLDB Journal 13(2): 162-176 (2004)

Shu-Yuen D. Yao, Cyrus Shahabi, Per-Åke Larson, Hash-based labeling techniques for storage scaling, VLDB Journal, available in online version.

Per-Åke Larson: External Sorting: Run Formation Revisited. IEEE Trans. on Knowledge and Data Eng. 15(4): 961-972 (2003)

Per-Åke Larson: Data Reduction by Partial Preaggregation. ICDE 2002: 706-715

Amira Rahal, Qiang Zhu, Per-Åke Larson: Developing Evolutionary Cost Models for Query Optimization in a Dynamic Multidatabase Environment. CoopIS/DOA/ODBASE 2002: 1-18

Poess M., Smith B., Kollar L., Larson P., TPC-DS, Taking Decision Support Benchmarking to the Next Level, SIGMOD 2002. (pdf)

Goetz Graefe, Per-Åke Larson: B-Tree Indexes and CPU Caches. ICDE 2001: 349-358

Jonathan Goldstein, Per-Åke Larson: Optimizing Queries Using Materialized Views: A practical, scalable solution. SIGMOD Conference 2001.

Last modifed on 13-Dec-2004