Broadcasting a Means to Disseminate Public Data in a Wireless Environment—Issues and Solutions - Advances in Computers - page 67

7 . 3 S e c t i o n C o n c l u s i o n
One of the problems associated with broadcasting information on parallel air
channels is the possibility of conflicts between accessing objects on different chan-
nels. Since the mobile unit can tune into only one channel at a time, conflicting
objects have to be retrieved on a subsequent broadcast. In addition, switching be-
tween channels imposes an extra cost. During the channel switch time, the mo-
bile unit is unable to retrieve any data from the broadcast. Conflicts will directly
influence the access latency and hence, the overall execution time. The issue of
conflicts was introduced in Section 3.5 . In an attempt to reduce the access time
and power consumption, as noted in Section 3.6 heuristics were used to develop
a scheduling access patterns that reduces the number of passes over the parallel air
channels. Analysis and effectiveness of such a policy was also the subject of this
section.

8. Conflict Resolution and Generation of Access Patterns

Beyond Heuristics

Our discussion in Section 7 mainly focused on problem-specific heuristics. Next
Object is a heuristic algorithm that always retrieves the next available object on a
broadcast. Row Scan simply reads all the objects from one channel in each pass.
Thesocalled bottom up , and top down tree-based algorithms generates an access
forest and uses a set of heuristics to choose the “most suitable access pattern” with
the goals to reduce the number of passes. It should be noted that the heuristic so-
lutions, in general, offer a “good” solution, not always. For example the protocols
discussed in Section 7 , in some instances could require unnecessary number of passes
and/or channel switches. The following two examples are intended to motivate this
issue.
Assume retrieval of eight objects from a three-channel configuration ( Fig. 36 ).
The application of the bottom up retrieval approach requires four passes (objects 3,

F IG . 36. Generation of unnecessary passes by the Bottom Up Retrieval Approach .