Conclusion

The caching options provided by ATL Server are powerful and versatile. They make the task of caching data very simple, and they also allow you a great deal of flexibility when you need to customize the cache.

In this chapter you examined the basic usage of the ATL Server caching support and you saw how you can customize it to provide greater type-safety and ease of use. You also learned how you can tune the caches by setting limits on their allocation size and entry counts. In addition, you saw how you can integrate the caches with Perfmon to provide performance information such as cache hits and misses.

However, theres also a great deal of advanced customization you can do with the caches that we didnt discuss in this chapter. For example, the DLL cache and the file cache have cache peer support. A peer class is one thats called back when entries are added and removed from the cache. These classes can provide even more powerful customization support by acting as a kind of trigger for these cache operations. For more information on peer classes, consult the MSDN documentation.

One of the main reasons for choosing ATL Server as a technology is to create a high-performance application. ATL Server allows you to get as close to the underlying technologies as you desire , and though this ensures that maximum performance is attainable, it also means that the development team is responsible for ensuring that its application performs well. In the next chapter youll take a look at some of the application tuning tips and tricks that the ATL Server team learned while working with both internal and external groups