Which kinds of OS mechanisms should be provided for database management?
Proceedings of the International Workshop on Experiences with Distributed Systems
Fbufs: a high-bandwidth cross-domain transfer facility
SOSP '93 Proceedings of the fourteenth ACM symposium on Operating systems principles
IO-Lite: a unified I/O buffering and caching system
ACM Transactions on Computer Systems (TOCS)
Operating system support for database management
Communications of the ACM
ACM Transactions on Database Systems (TODS)
Operating System Extensions for the Teradata Parallel VLDB
Proceedings of the 27th International Conference on Very Large Data Bases
Notes on Data Base Operating Systems
Operating Systems, An Advanced Course
Virtual memory management for database systems
ACM SIGOPS Operating Systems Review
I/O Phase Characterization of TPC-H Query Operations
IPDS '00 Proceedings of the 4th International Computer Performance and Dependability Symposium
An Efficient Zero-Copy I/O Framework for UNIX
An Efficient Zero-Copy I/O Framework for UNIX
Characteristics of production database workloads and the TPC benchmarks
IBM Systems Journal - End-to-end security
Mining block correlations to improve storage performance
ACM Transactions on Storage (TOS)
C-Miner: Mining Block Correlations in Storage Systems
FAST '04 Proceedings of the 3rd USENIX Conference on File and Storage Technologies
C-Miner: mining block correlations in storage systems
FAST'04 Proceedings of the 3rd USENIX conference on File and storage technologies
| Hi-index | 0.00 |
A range of database services are being offered onclusters of workstations today to meet the demandingneeds of applications with voluminous datasets,high computational and I/O requirements and alarge number of users. The underlying database engineruns on cost-effective off-the-shelf hardwareand software components that may not really betailored/tuned for these applications. At the sametime, many of these databases have legacy codesthat may not be easy to modulate based on theevolving capabilities and limitations of clusters. Anindepth understanding of the interaction betweenthese database engines and the underlying operatingsystem (OS) can identify a set of characteristicsthat would be extremely valuable for future researchon systems support for these environments.To our knowledge, there is no prior work that hasembarked on such a characterization for a clustereddatabase server.Using IBM DB2 Universal Database (UDB) ExtendedEnterprise Edition (EEE) V7.2 Trial versionand TPC-H like decision support queries, this paperstudies numerous issues by evaluating performanceon an off-the-shelf Pentium/Linux clusterconnected by Myrinet. These include detailed performanceprofiles of all kernel activities, as well asqualitative and quantitative insights on the interactionbetween the database engine and the operatingsystem.