Using iPEP, when a website becomes overloaded the surplus workload is automatically diverted to an overload server, which manages the situation according to company policy. This can range from diverting users to a cascade of servers providing a range of service levels, through to totally blocking access above a certain level of demand. The choice is solely a matter of company policy.
In practice, therefore, the logical website can be made up of a number of different software servers, each executing an element of the website policy. This will normally involve providing visitors with different levels of service or changing their priority level, depending on what actions they are taking. For example, all e-retail sites have a large number of potential customers browsing the site at any one time. iPEP will identify each user when they first access the website, and allow their interaction to start with a period of time - typically a few minutes - at the highest priority level. Their priority rating will then depend upon subsequent activity. So in a typical e-retail website every visitor starts as a "customer" with a similar potential to be downgraded into a `window shopper'.