Student Name:
Title of Talk:
Speaker:
Date:
Bojja, Kiran Kumar.
Pervasive Computing
Sachin Singh
09-11-2003.
Summary of Talk:
The most successful technologies are those that disappear in to every day life. Traffic signaling
system is one example. Pervasive computing is different from mobile computing and distributed
computing in terms of robustness and maturity requirements. It provides ubiquitous computing
environment to clients. A smart space connects the two disjoint worlds, which is a building block
to this everywhere computing environment. These smart spaces, which can be different from each
other, will be integrated into global computing environment based on client thickness. To achieve
certain level of invisibility, the computing environment should be with information and
intelligence, which allows minimal user interaction. Thus balancing between degree of smartness
and security is one of the major issues in pervasive computing. Location awareness, Context
awareness will determine the degree of adaptability of computing environment to its client.
The pervasive computing system consists of one or more Autonomous Systems (AS). Each AS has
unique URI, unique AS head and consist of smart spaces. Each smart space consists of sensor
devices to communicate with the client. Each specification will have a detailed semantic
description on the semantic web including functionalities, technical details, etc. Each individual
will have personal profile with user information and security concerns. The pervasive profile gives
user authentication inputs that are used to provide authentication services. Smart spaces will
communicate with each other and also with AS Head. In global view, Each AS Head will
communicate with its counterpart to provide effective computing environment. The goal of
pervasive services is to provide right service at right time. The client needs to get the description of
service at his end and then he is ready to use the service. These semantic descriptions will have
unique URI on semantic web. In short sensors do basic computing and managers interpret the
semantics and drive sensors together to provide services within the AS.
Strengths:
 Pervasive systems provide the client with computing environment everywhere.
 As the environment persistently meets the client needs, the user can interact at subconscious
level. This makes the environment seamless.
 The biggest strength of pervasive computing systems is their adaptability to different contexts,
locations and different roles of clients.
 Computing resources will be used to improve the user experience.
Weaknesses:
 The requirement for sophisticated smart spaces is difficult to provide because of the sever
bandwidth and energy constraints on mobile devices.
 Huge differences in smart spaces make it difficult to integrate them in to a global computing
environment. Protocols need to be written to facilitate reliable communication. Clint thickness
is the constraint here.
 The clients need to be sophisticated. This will increase the cost and complexity.
Interesting points:
 User-centric rather than technology-centric.
 Utilizing computing resources efficiently to improve user experience
 Balancing between security policies and degree of smartness of the environment.
 Developing protocols to facilitate the communication between different autonomus systems or
smart spaces.
Critical questions:
 How does one can determine the accuracy of knowledge in determining the user intent? How
does this information transfer between the different layers in pervasive computing systems?
Trust issues need to be specified.
 How do we determine the standards for client thickness to avoid device failures during heavy
load periods?
 How can we reduce the energy demand on mobile devices?
 In providing context awareness to the computing environment, How does the context
information have to be updated?