Telecom Industry Software System Optimization

I recently read a lot of Internet site optimization solutionsArticleI feel very helpful. I would like to summarize the experiences of software optimization in the telecom industry over the past few years.

Software in the telecom industry is very different from the general internet. Although the current architecture adopts the B/S structure, the software interfaces are web interfaces, but the two are actually quite different.
Software users in the telecom industry are usually internal employees, partners, and business users. The total number of transactions is generally small, but there are many read and write operations, which are transaction-intensive access. Unlike normal websitesProgram, Many are read-only access.
Industry software users and their business needs high security. Different from the security requirements of general websites, the security requirements mainly reflect the isolation from core business systems.
The investment in software in the industry is generally guaranteed, so the choice of software and hardware solutions tends to be the commercial software solutions of large companies rather than open source software solutions. In recent years, with the changing cost requirements, a large number of open-source solutions have also been adopted in non-core business systems.

 

Common optimization measures include (but are not limited ):
1. The web server (Application Server) is separated from the database server.
2. Increase the number of application servers and use Apache and other software Load Balancing Solutions
3. Use cache, including page cache and data cache
4. Using F5 and other hardware Load Balancing solutions, you can adjust the application to scale horizontally.
5. Add a Database Cluster
6. Database Application separation can be performed based on the Application nature or the read/write frequency.

 

Compared with the optimization measures of Internet sites, the optimization ideas are the same, but the specific problems are somewhat different:
Website programs usually use a large number of dynamic and static page separation and separate optimization, industry software to specific analysis, not necessarily need to use this technology.
Large-scale Distributed applications, such as Google, adopt the Distributed File System GFS, parallel computing mapreduce, and distributed storage technology bigtable, which are already very different from the traditional technical architecture. As an industrial application, it is generally not necessary, and there is no technical strength to implement such a system (although there are open-source projects, such as hudoop ).