Quality of Information Services

One of the main problems when using a database system is that the user does not know when the answer comes. This is acceptable if operations are carried out asynchronously, like a bank transaction (e.g., a money transfer) or the delivery
of a business report which is known to take several days. On the Web, however, users typically get impatient after a few seconds and visit another site (potentially, the Web site of a competitor) in order to try the request there.

The fundamental problem is that all commercial information systems have been designed using a best-effort strategy: a request is processed as early as possible, given the current load of the system and possibly taking priorities of requests into account. Clearly, this strategy is sufficient for many applications and it is good if the database is given enough resources so that all delays are tolarable. However, as mentioned above, it is easy to envision environments in which this strategy is not satisfactory. The goal of this project is to devise new serviceoriented database architectures. Such architectures allow the database to adapt to non-functional user requirements. If a user needs an answer quickly (and is willing to pay for it), the database will deliver the answer to that user in the specified amount of time or indicate to the user that it is not able to do so. Furthermore, the database makes the whole process more transparent to the user and indicates how much progress has been made on the request and possibly allows the user to re-prioritize the execution of a request.