Scalable Adaptation of Web Applications to Users' Behavior

In this paper we present a comparative study of performance of an adaptive e-banking Web application supporting personalization either on a client or on a server side. Currently, modern applications being developed support various kinds of personalization. One of its types is changing behavior and appearance in response to actions taken by a user. Not only pre-defined rules but also new patterns discovered for different levels of events should be applied. Scaling such "interactive" applications to a large number of users is challenging. First, the stream of events generated by users' actions may be huge, and second, processing of the adaptation rules per single user requires computing resources that multiply with the number of users. This paper reports on the efficiency of the method enabling a client-side adaptation after moving adaptation logics from a server to a client.