Distributed Clients Profile Management System For Pervasive Systems

The heterogeneity of device capabilities, user’s context and communication network thatrnexists in pervasive systems has accentuated the need for a content adaptation. Thernpurpose of the content adaptation is to convert the original content before delivery so thatrnthe new content would fit to the capability of the client’s device being used, the availablernbandwidth of the communication network, and the specific need of the user. In order tornperform such an adaptation, the adaptation system is required to maintain adequaterninformation about the clients’ profile, the network profile, the content profile and others.rnThis thesis work deals with the management of the clients’ profile that includes devicernand user profile. Specifically, it tries to set up an appropriate way of managing clients’rnprofile in pervasive environment in relation to content adaptation. To this end, arndistributed clients’ profile management scheme is proposed in the current work. In therndesign of the management scheme, two issues have been considered: (a) client’s profilerncharacteristics such as existence of confidential information in the client’s profile,rntemporal nature of some of the profiles, etc. (b) constraints that exist in the pervasivernsystems such as heterogeneity of the client devices in terms of storage, processing, andrndisplay capabilities, mobility of the clients, etc. Based on the profile characteristics andrnthe constraints of pervasive systems, we proposed to fragment the client’s profiles intorntwo: those that can be kept on the client devices and those that can be kept on the serverrnside. The main consideration of this fragmentation is the security issue of confidentialrnprofiles and the storage size constraint of the client devices. In the proposed architecture,rnthe server side profiles are, in turn, distributed on the local servers to meet thernrequirements such as availability of profile information in partial failure of the system,rnperformance, scalability, etc.rnThe proposed system consists of different components that interact with each otherrnthrough web interfaces for the management of distributed client’s profile in a transparentrnway. It also provides interfaces to external components such as automatic profilernextractors so that the dynamic profiles will be built in the course of service provision.

Get Full Work

Report copyright infringement or plagiarism