Lev Novik

Adapting microsoft SQL server for cloud computing

Cloud SQL Server is a relational database system designed to scale-out to cloud computing workloads. It uses Microsoft SQL Server as its core. To scale out, it uses a partitioned database on a shared-nothing system architecture. Transactions are constrained to execute on one partition, to avoid the need for two-phase commit. The database is replicated for high availability using a custom primary-copy replication scheme. It currently serves as the storage engine for Microsofts Exchange Hosted Archive and SQL Azure.

P2P replica synchronization with vector sets

This paper describes an enhanced replica synchronization mechanism built in Microsofts WinFS replica management system. The system reconciles autonomously-operating replicas in a completely peer-to-peer manner, without employing a central master or locking. The main challenge is for two replicas to exchange meta-information efficiently about (potentially numerous) data objects in order to discover what updates they are missing, and detect conflicts. The paper introduces a novel bundling mechanisms called VS, that groups together multiple objects and represents their state in a single version-vector. VS provides improved storage and communication overheads over previously known optimistic replication schemes, in the following sense. Under normal, low-fault situations, it maintains and communicates as little as a single version vector in order to represent precedence ordering of the entire set of data objects. Moreover, under settings of severe communication disruptions, VS degenerates to no worse than a single vector per object. This dramatically improves the complexities described in a preliminary write-up of the WinFS replication scheme. The VS mechanism has potentially wide applicability as a mechanism for compactly handling synchronization of arbitrarily overlapping groups of objects.

Annotating database schemas to help enterprise search

In large enterprises, data discovery is a common problem faced by users who need to find relevant information in relational databases. In this scenario, schema annotation is a useful tool to enrich a database schema with descriptive keywords. In this paper, we demonstrate Barcelos, a system that automatically annotates corporate databases. Unlike existing annotation approaches that use Web oriented knowledge bases, Barcelos mines enterprise spreadsheets to find candidate annotations. Our experimental evaluation shows that Barcelos produces high quality annotations; the top-5 have an average precision of 87%.