Welcome to D
SIGMOD 2003
<<< = SIGMOD'03 Pa>>>
PODS 2003
SIGMOD-RECOR
ADBIS
CIDR 2003
CIKM 2003
DASFAA 2003
Data Enginee
DEBS
DMKD 2003
DOLAP 2003
DPDJ 2003
ER
GIS 2003
Hypertext 20
ICDE 2003
ICDM 2003
ICDT 2003
JCDL 2003
KRDB 2003
MIR 2003
MIS 2003
MMDB 2003
RIDE 2003
SBBD 2003
SIGIR 2003
SIGIR-FORUM
SIGKDD 2003
SIGKDD-EXP
SSDBM 2003
TIME 2003
TODS
VLDB 2003
VLDB Journal
WIDM 2003

DBCache: Middle-tier Database Caching for Highly Scalable e-Business Architectures


Christof Bornhövd, Mehmet Altinel, Sailesh Krishnamurthy, C. Mohan, Hamid Pirahesh, and Berthold Reinwald

  View Paper (PDF)  

Return to Caching


Abstract

Multi-tier infrastructures have become common practice for implementing high volume web sites. Such infrastructures typically contain TCP load balancers, HTTP servers, application servers, transaction-processing monitors, and databases. Caching has been widely used at different layers of the infrastructure stack to improve scalability and response time of e-business applications. The majority of existing caching mechanisms target only static HTML pages or page fragments. However, as web applications become more dynamic through increased personalization, these caching techniques turn out to be less useful. Consequently, as more application requests result in increased querying and updating of backend database servers, scalability limits are often reached. Middle-tier database caching is a promising approach to overcome the above problem. Our research prototype is based on DB2 UDB. We introduce a new database object called Cache Table that enables persistent caching of entire tables, or subsets of tables. Cache tables come in two flavors. The content of a cache table is either defined declaratively and populated in advance at setup time (declarative cache tables), or the content is determined dynamically and populated on demand based on incoming queries (dynamic cache tables). In this demonstration, we focus on dynamic cache tables and present the following features: query plans for dynamic cache tables; cache constraints; cache population and maintenance; cache-aware routing. We demonstrate these DBCache features using IBM's Trade2 J2EE benchmark, which models an online brokerage firm. Trade2 consists of Java classes, Java Servlets, Java Server Pages, and Enterprise Java Beans.

BIBTEX


@inproceedings       {DBLP:conf/sigmod/BornhovdAKMPR03,
  author    = {Christof Bornhövd and
                Mehmet Altinel and
                Sailesh Krishnamurthy and
                C. Mohan and
                Hamid Pirahesh and
                Berthold Reinwald},
   booktitle = {SIGMOD Conference},
   title     = {DBCache: Middle-tier Database Caching for Highly Scalable e-Business Architectures.},
   pages     = {662},
   year      = {2003},
   url       = {db/conf/sigmod/sigmod2003.html#BornhovdAKMPR03},
   ee        = {http://www.acm.org/sigmod/sigmod03/eproceedings/papers/dem06.pdf},
   crossref  = {conf/sigmod/2003},
   bibsource = {DBLP, http://dblp.uni-trier.de} 
}



©2004 Association for Computing Machinery