J2EE clustering 1
Overview
If you want to build a scalable and highly reliable website, you need to understand the cluster technology (clustering ). in this article, Abraham Kang describes how to implement a J2EE cluster and lists Bluestone Total-e-server, Sybase Enterprise Application Server, what are the technical differences between SilverStream Application Server and WebLogic Application Server in the cluster. based on this knowledge, you can design your own effective and efficient J2EE applications.
Abraham Kang
Enterprises are increasingly choosing Java 2 and Enterprise Edition (J2EE) to develop their job-based online applications. in the J2EE framework, the cluster technology ensures the minimum downtime and maximum scalability. A cluster is a set of application servers that transparently run J2EE application services, as if they are a whole. additional machines must be provided in the cluster. to minimize the number of healthy services, each component must be redundant.
In this article, we will obtain the basic knowledge about clusters, cluster methods, and important cluster services. because the industry cluster technology is very different, we will compare the advantages and disadvantages of each technology. further, we will discuss the features to be implemented by the cluster-related application server.
To contact the actual Application, let's take a look at how the cluster is implemented by HP Bluestone Total-e-Server 7.2.1, Sybase Enterprise Application Server 3.6, SilverStream Application Server 3.7, and BEA WebLogic Server 6.0.
In the second part of this article, we will discuss cluster programming and error recovery strategies, and test how the four application server products we mentioned scale and restore errors.
Cluster definition
<Div> The J2EE application server supplier defines a cluster as a group of computers working together to provide transparent enterprise-level services (supporting JNDI, EJB, JSP, HttpSession, and component error recovery ). they are deliberately vague, because their implementations are different. some developers place a dispatcher on the front end of a group of independent machines. The dispatcher accepts user requests and then uses the HTTP redirect header to forward requests to a specific server in the cluster. another part of developers implement a tightly integrated machine combination, where each machine can fully perceive the existence of other machines around it, together with the objects on them.
In addition to machine integration, the cluster also includes redundancy and error recovery:
Server load balancer: enter a single entry point of the cluster, the traffic indicator of the site or application server
Gateway Router: egress of the internal network
Multi-layer switch: packet filtering or frame filtering to ensure that each machine only receives information related to itself
Firewall: Port-level filtering to prevent hacker from entering the cluster or internal network
SAN (Storage Area Networking) controller: connects application servers, Web servers, and databases to the backend Storage media, manages the hard disk to which data should be written, and backs up errors
