Sunday
Feb172008
Web Accelerators - snake oil or miracle remedy?
Sunday, February 17, 2008 at 11:32PM
Perhaps this question is borderline off-topic but since high scalability solutions often have a global aspect I will give it a try...
Have anybody had any experience with different techniques for speeding up their application to places that have a problem with poor ping response time? Ideally I would love to be running only one data center world-wide but one day I know that our sales department will sign up a customer with an unacceptable response time... Could installing a web-accelerator in front of our application extend the reach of our current data center or will we just add complexity and another source of potential errors?
Have anybody had any experience with different techniques for speeding up their application to places that have a problem with poor ping response time? Ideally I would love to be running only one data center world-wide but one day I know that our sales department will sign up a customer with an unacceptable response time... Could installing a web-accelerator in front of our application extend the reach of our current data center or will we just add complexity and another source of potential errors?
Reader Comments (5)
It depends on what they're calling a 'web accelerator'.
Is it just a caching server like Squid?
Sorry,
realized that I was a bit thin on the technical details; I am specifically wanting to hear "war stories" about server side web-accelerator techniques that targets the "middle mile" (not client or server performance) in networks with reasonably good bandwith but high latency. One of the techniques I know of is overcoming the browsers typical TCP/IP connection limitation of 2-4 connections by manipulating your html into serving html elements (gifs, css, javacripts) from different sub-domains.
The answer depends on what you're trying to do.
If you're trying to speed up access to web-based applications for general users accessing your website across the internet then single-ended web accelerators (from the likes of F5, A10) are great products. The primary focus of these products is to offload server functions (TCP termination, compression, caching) thereby improving overall performance.
If you're an enterprise looking to speed up web-based application performance across the WAN then you open yourself up to symmetric solutions (from the likes of Silver Peak and Riverbed). Symmetric solutions require the installation of an appliance at both "ends" of the WAN link. The primary focus of symmetric solutions is to offload the WAN rather than the server. As such, these solutions have features that include advanced data-reduction techniques, TCP acceleration, packet-loss correction, out-of-order-packet correction, and encryption. The goal of these products is to remove the WAN as a bottleneck in application delivery and data replication environments. WAN acceleration solutions offer acceleration for all applications not just http(s).
Oddly enough I just wrote my first article about this type of stuff. I'd appreciate it if any of you could read it and provide feedback. The article is at http://www.winnersdontlose.com/2008/02/application-acceleration-strat.html.
Hope this helps,
Tony
Interesting article Tony, thanks. Could you explain this bit a little more?
>Akamai's service in this arena is called Web Application Accelerator (WAA) and >utilizes their SureRoute technology to pass end user traffic back to origin and vice >versa. In theory, WAA bypasses the BGP that most providers user by maintaining >their own (more optimized) route tables, thereby decreasing the latency an end >user experiences.
How much faster in practice would their routing layer be?