Analysis on Architecture and performance of Web Report Software

Web Report Software generally has three architectures. One is a pure Java report, one is a control method, and the other is a report software in the form of an independent server.

 

1. Control Mode: In this mode, an ActiveX control is installed on the browser. on the server side, only data is retrieved. Report calculation and generation are performed in the control. The reason for using this method is that its report function is often implemented using Delphi, and many products use the same third-party components.

The disadvantage of the control method is:

1) Deployment: a single control can be used to complete most of the functions implemented by other report servers. The more powerful the function, the larger the size of the control. This is a huge burden on deployment. In addition, when the report system is upgraded, you must reinstall the control. The B/S structure of the Web Report software has no advantages in deployment.

2) Performance: for PC-Side Computing and server-side computing, small reports may not be much different, but for large reports, it is clear that the difference is very large. In addition, if a Web report system is used by multiple users, the scheduled computing and cache modes that can be used by the server cannot play a role.

Therefore, the control method is generally only applicable to simple and small reports. The simplicity here is not only a simple report style, but also a simple computing workload and report scale.

 

2. independent server: install an independent (logical) Report Server in this way. It avoids the disadvantages of complicated deployment and inconvenient upgrade of Web Report software using controls. It can effectively manage reports through the server and handle performance problems well. However, it is difficult to deploy, and many such Web Report software products are difficult to achieve cross-platform implementation. Especially when it involves issues such as clusters and connection pools, it cannot provide good scalability.

 

3. Java report software: first, it can easily implement cross-platform deployment of web reports. For ASP ,. net and other non-Java applications, it can provide Web Report services in the form of a report server. At this time, this Web report tool works in a way similar to products such as crystal report, it has the same advantages and disadvantages as an independent server (the independent server here refers to the logic, and can still be physically installed together with other servers ). For Java applications, the report software can be embedded into the J2EE application architecture in the form of class packages, forming a seamless connection with the application system. At this time, its advantages are even more obvious:

1) All Java application server performance improvement means can play a full role in Web Report software, such as servers with stronger computing capabilities (which does not play a major role in control-based servers, independent report servers are difficult to deploy on Unix servers), Server Load balancer, and connection pools.

2) Web Report software can be more closely connected with applications and used more flexibly by applications.

 

Some people say that you should not always talk about Unix. In actual applications, many application servers use PC servers. This is true, but (not to mention that many Web reporting tools do not support Linux well) do not forget that the reason why the application server layer uses PC servers is that it is suitable for horizontal scaling, that is, cluster and Server Load balancer technology. This is too easy for Java report software. It can be deployed directly with the application. It can be said that it is born, but it is very troublesome for independent servers.