Service Discovery Architecture and Protocol Design for Pervasive Computing - Advanced Design Approaches to Emerging Software Systems - page 85

INTRODUCTION processes and the basic steps that each device or

software component is taken.

In pervasive computing environments, people are
surrounded by a variety of computing devices.
Those devices communicate with each other and
provide network services and information without
people's active attention (Weiser, 1991). Presently,
PCs, smartphones, MP3 players, and laptops sur-
round us. In the near future, additional networked
computers, ranging from sensors, RFID tags to
extremely dynamic and heterogeneous devices
will provide a variety of services. It becomes
overwhelming to manage these devices, configure
different kinds of applications, and dynamically
find the available computing services in such
pervasive computing environments.
Service discovery protocols enable comput-
ers to be easier to use. They facilitate interaction
between computers, with an aim to approach zero
administration overhead and therefore free users
from tedious and redundant administrative and
configuration work. Therefore, service discovery
research is critical to the success of pervasive
computing (Kindberg & Fox, 2002).
The objective of this chapter is to discuss
the key design issues and solutions for service
discovery protocols in pervasive computing
environments. In the last 15 years, many service
discovery protocols were designed by industry,
academia, and international standards develop-
ment organizations. The protocols emphasize on
aspects of the service discovery. We analyze the
design of the major components, their interactions,
service selection, performance optimization, and
security and privacy issues.

Service Discovery Models

There are three service discovery models. A trivial
service discovery model is one in which a client
(computing device) knows a service (network ser-
vice) in advance, or the client has already cached
the service's information, so that the client does
a local lookup before contacting the service. The
second model is the client-service model as shown
in Figure 1 (a). The model performs best in simple
environments such as home environments. Clients
inquire about all services. If a service matches the
client's enquiry, it replies back. Then, the client
communicates with and accesses the service. To
support thousands of computing services, such
as the services in public environments, we may
optionally use directories to store all the service
information. This third model is call client-service-
directory model. A client queries a directory for
service information and then contacts services. We
discuss the different perspectives of the clients,
services, and directories.

Client View

In most cases, a client is a program that runs on
behalf of a user and interacts with the user. It
usually takes the following steps.

• A client queries directories for services. A

client either browses services or looks for
a specific service.

• Alternatively, without going through direc-

tories, a client directly queries all the ser-
vices. All the services that meet the query
requirement reply back to the client.

BACKGROUND

Before we present representative service discovery
protocols, we describe three general models. The
models focus on the functionalities of comput-
ing devices or software in the service discovery

• Then the client program or the user selects

a service to use.

• Finally, the client uses the service.