Sundar Yamunachari wrote: > Possible solutions: > > 1. Use CherryPy web server to serve both AI install files and AI manifests. > Advantages: > - Flexible for adding the features we need. > - Light-weight (Use only the code we need) > > Disadvantages: > - We have to maintain the code > - We have to create packages for other operating systems (Solaris > 10 and Linux) > - Is it scalable?
cherrypy is not suitable for serving large amounts of data and its threading model is fairly primitive (a fixed pool of threads that cannot grow or shrink, though this is set to change in the near future). As such, it isn't very scalable if you're talking about thousands of simultaneous requests, etc. However, the major advantage to cherrypy approach is that you can reverse-proxy the aspects that need higher performance with Apache, pound, or various other web servers. Additionally, cherrypy applications can be very portable and can work on many platforms while requiring very little administrative work for setup. Finally, cherrypy provides a rich web interface that makes it easy to rapidly alter your application. Cheers, -- Shawn Walker
