Dr. Dobb's Journal - November 2007 - (Page 24)

d11zeid_p6db 9/10/07 10:59 AM Page 24 Core Technology GRID-ENABLING RESOURCE-INTENSIVE APPLICATIONS continued from page 22 • In the client-server scheme, client code requests data from the server, and input to the client is sent back to the server when the client makes any changes. This is the most common architecture. Architecture Client-Server Architecture Comparison Advantages Many tools available, especially in Java. Basis for other architectures. No need for central server. Robust (no single point of failure). • In the peer-to-peer scheme, there is no particular machine providing a service or managing the network resources. All responsibilities are uniformly distributed among the machines on the network. • Distributed objects are software modules that exist on multiple computers Disadvantages Overhead of managing socket connections can be tedious. If server fails, clients are useless. Harder to manage from a central location. Generally refers to sharing of files, but not hardware resources. Requires middleware or developer must manage synchronization and client-server overhead. Needs homogeneous hardware. High Administrative overhead for processes and network. Compared to a normal cluster, a grid has more middleware issues to deal with. Remote Method Invocation (RMI) Strategy Sockets throughout the network but are designed to work together. A program running on one machine sends a message to an object on a remote machine to perform some processing. The results are sent back to the calling machine. There are no distinctions between Advantages The “Standard” for network application communication. Disadvantages Data is sent as bytes and must be reconstructed into something useful at the receiving end of the connection. Complex Interface Description Language needed to allow various platforms to call the RPC. Brief Comparison of Communication Strategies Peer-to-Peer Remote Procedure Calls (RPC) Distributed Objects Very scalable. Increased flexibility. Sends complete data structures instead of just bytes. Offers reliability in the form of an acknowledgment message that the data has been successfully transferred. Eliminates socket management overhead. Passes objects instead of straight bytes or data structures. Eliminates socket management overhead. Access to rich Java library tools. Clustering High Performance. Excellent load balancing. Java only (unless JNI is used). Need JVM 1.5 or newer to use latest features. Grid Computing Can use heterogeneous hardware. Can be used across large networks such as the internet. No central administration. Table 1: Architecture comparison. Table 2: Communication strategies. 24 Dr. Dobb’s Journal l www.ddj.com l November 2007 http://www.sparxsystems.com http://www.sparxsystems.com http://www.ddj.com

Table of Contents Feed for the Digital Edition of Dr. Dobb's Journal - November 2007

Contents Hmmmm Alia Vox Developer Diaries Developer’s Notebook Smart Compilers - But Smart Enough? Conversations Grid-Enabling Resource-Intensive Applications Distributed Computing: Windows and Linux Adobe AIR: Desktop/Web Convergence Transparency on Demand Reusable Associations Effective Concurrency The Agile Edge Swaine’s Flames

Dr. Dobb's Journal - November 2007

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.