• Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint

16.3. Conclusion

Just as there are different types of FlashCom applications, there are different approaches to building multi-instance applications that are scalable and fault-tolerant. Media-on-demand, live one-way broadcasting, and n-way communication applications all provide different challenges. Applications that require real-time instance coordination are more challenging to design and build than those that do not. Component-based applications are a good way to partition multi-instance applications. The components introduced in Chapter 13 through Chapter 15 have client- and server-side parts that work together. In a multi-instance application, a good practice is to further divide components so that their server-side parts know how to communicate with one another. In a simple broadcast tree, a stream list component might have four parts—one each for a source client, root server, leaf server, and destination client. Writing each part of the component separately simplifies multi-instance development.

This chapter has dealt with building multi-instance and multiserver applications that don't bog down as the number of client connections increases. Unfortunately, adding instances and servers won't help application performance over slow network connections. The next chapter describes what you can do to provide a good application experience despite network limitations.


PREVIEW

                                                                          

Not a subscriber?

Start A Free Trial


  
  • Creative Edge
  • Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint