Information Technology Reference
In-Depth Information
1.1
Data transmission in ubiquitous systems
This paragraph will explain the techniques used to enable the transmission of
data in distributed environments for ubiquitous computing. The interaction
plays a crucial role for ubiquitous devices: the reduced potential of the
individual computer is balanced by the cooperation between them. Note
that this last point is a 'conditio sine qua non' for the ubiquitous system.
Ubiquitous devices exchange among themselves a certain flow of infor-
mation, and this exchange depends on the quality of service provided by the
system. The information must be processed, routed and exchanged as quickly
as possible (each node has the data needed by a number of other nodes). This
goal can be achieved by exploiting the fact that each node of a ubiquitous
system is actually a computer but within the limits imposed by its structure.
Summing-up, since the tasks of the ubiquitous node are not so complex, it
will be sufficient to provide it with resources that, though not being
'excellent', are perfectly suited to the objective (or objectives) that it must
pursue [4].
1.2
Objectives
The basic problem in the transmission of technical data in a ubiquitous
environment is to determine an efficient plan (scheduling) to determine the
order in which to meet in time the required data as soon as they are
presented to a node. Until now, one of the most used policies for this regard
was the one based on the RxW algorithm. This algorithm calculates for
each data the object requested to the node, the product of computer-users
(R) who requires it and for the maximum allowed waiting time (W), the
time elapsed since the data was required for the first time. A system based
on a RxW policy then sends a data pack for two reasons: because it is
highly required (high R value) or because some node has required it for a
long time. This algorithm responds in a balanced way both to the large
demands of a certain feature (hot requests) and to those who are not very
(cold calls) so requested. Other proposed algorithms were valid application
for environments characterized by applications with a variable size of data
requirement, such as the MAX algorithm, proposed by Acharya and
Muthukrishnan, whose characteristic was to first meet the demands for
larger data [5].
Finally, we mention in this section the fact that the increasingly urgent
need to reduce the response time of the node which require information has
led to the use of particular components, such as the cache, to speed up the
supply of data. As a consequence, the problem of implementing algorithms
that support the introduction of such hardware arises.
