May 15 2006

We have made a small real-case scenario benchmark to simulate the realtime compression of dynamic generated web pages. You can download it and test it by your own.

The benchmark consists of a simple page ASP.NET that produces an output of 1500bytes, 12K and 24K respectively. Then the client executes concurrent threads that ask recursively these pages.

We have ran the benchmark on an entry level AthlonXP 3000+ CPU with Windows 2003 Server and IIS6, the client ran on the same host.

The results are compared to one native implementation, and three commercial solutions that we keep without name or order for respect purposes. All the products ran with default setup and zlib level 1 compression deflate, didn't save log and didn't treat debug, this configuration is the optimal to achieve the maximum possible speed with every product.

1) solution n.1 doesn't produce realtime compression of the dynamic ASP.NET page
2) solution n.2 doesn't produce realtime compression of the dynamic ASP.NET page
3) solution n.3 doesn't produce realtime compression of the dynamic ASP.NET page
4) solution n.4 offers realtime compression of the dynamic pages with ~500 pages for second
5) Compress Content offers ~1.600+ compressed pages per second with more than 300% speedup
6) another test was done calling the same webpages without random querystring and using the cache offered by the solution n.4, without any noticeable speed difference
7) Compress Content is so fast keeping the HTTP connection open, in contrast to some other solutions sending a Chunked Encoding answer and closing the client connection (lowering the overall server performance)

The same benchmark, done with static files, produced an outstanding result of ~5.000 realtime compressed pages per second, without caching techniques.

Compress Content is ideal for websites running dynamic pages such as forums, webboards, blogs, database driven applications. Today most of the internet web sites are driven by dynamic applications. With low cost, high performance and 100% reliability you can save big money by reducing the bandwidth (typically by 300%) and the hardware demands, and enhancing the user experience with the short web pages loading time.