During a JavaOne Hands-On Labs presentation, Rangarajan and Ramnarayan will discuss the concept of data grids in clustered J2EEâ¢ deployments and the role played by new Java standards like JCache (JSR 107) and JDK 1.5 in the implementation of a data grid.
Over the last several years, grid computing has been primarily utilized in heterogeneous, decentralized groups of computers to tackle large-scale analytical problems. But now the field has rapidly expanded to include locating, managing, and monitoring disparate compute resources, as well as managing and providing virtualized access to the underlying data.
Providing virtualized data access to an organizationâs data repositories is referred to as a data grid. Data grids provide transparent access to data residing in multiple disparate data repositories like a relational database, a legacy application, or a web service. A reliable data grid solution has to provide guaranteed data access response time and throughput, data distribution, and synchronization strategies. Such a data grid solution also needs to be relatively non-intrusive for adoption in existing large-scale Javaâ¢ 2 Platform, Enterprise Edition (J2EEâ¢) deployments.
Data grids are being increasingly adopted as new platforms, such as blade servers, enable system administrators to dynamically provision hardware resources based on demand. When applications become dependent upon back-end data repositories, significant data latency results and consequently the system becomes unable to scale to increased loads. A data grid or data back-plane co-located with the J2EE application server will shield the back-end from a barrage of client requests and enable a scalable distributed architecture.
The JavaOne Conference will take place June 27 â 30 in San Francisco at The Moscone Center. The presentation, âData Grids for J2EEâ¢ Platform Cluster Deployment," occurs on Thursday June 30 at 3:45 pm, Hall E 134. GemStone is located at booth #428.