The performance that your portal must deliver directly affects your deployment requirements. Scalability, capacity, and high availability are some of the standards you need to consider.
Develop your performance and capacity requirements by answering the following questions:
What performance requirements exist?
What high availability requirements exist?
What response times are acceptable? How do the response times of your stand-alone systems compare with response time requirements of your portal?
If you size your portal infrastructure for good response times during regular hours, can you tolerate a possible degradation in performance during peak load times?
How many concurrent sessions, or connected users, are likely during peak use? (Count only users who are active. Do not include users who are, for example, away on vacation, on leave, or sleeping.)
What is the above-normal peak time? How does this information affect your peak concurrent user estimate?
What sort of user activity occurs during peak periods? Logins or reloads?
How long do you expect the typical user to be connected, or have a valid portal session open? What use statistics do you have for existing applications? Do you have Web traffic analysis figures for an existing portal?