When Episode 3 came out last month I was disappointed in the complete inability of cinema company websites to handle the increase in traffic.
In general I think businesses have a lack of understanding regarding the burstiness nature of the web and it is our job to make them aware of it.
So come on people get your act together, cache some data and improve the user experience!!
Lets pick up the pace from the appserver end of things so users who move to ADSL2 get better response times, not just a bigger pipe.
While you are at it gzip some of your large pages as well.
One Response
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.
The problem with dealing with spikes is the fact that you need to over-resource for them. It’s usually not cost-effective to have sufficent computing power to handle the load.
What you _really_ need is an efficient way to scale your system dynamically, _plus_ a pool of resources shared with other systems that have different usage patterns. This shared pool can then be drawn on to deal with spikes as needed. Sun, for example, is offering such a service to certain sites, and IBM can do similar stuff (which is why IBM handles the major sporting events, for example). However, this is all highly non-trivial.