Broadcasting a Means to Disseminate Public Data in a Wireless Environment—Issues and Solutions - Advances in Computers - page 19

case for tune-in time and response time. It was concluded that, in general, the ( 1 ,m)
indexing scheme reduces power consumption at the expense of an increase in the
response time; and the distributed indexing scheme, relative to the ( 1 ,m) indexing
scheme, increases the response time at a much lower rate at the expense of higher
power consumption.
Access frequencies of the data items have been used to generate an index tree.
Shivakumar and Venkatasubramanian [59] proposed application of tree-based index-
ing using Huffman-encoding scheme that is sensitive to data access patterns. The
work proposed in [18] constructs an index tree based on the data access skew. To
reduce the number of index probes, this work considered two cases: fixed index fan-
out and variant index fan-outs. It was shown that the cost of index probes can be
minimized if imbalanced index tree based on skewed data access is employed. This
reduces the number of index probes for hot data items at the expense of more index
probe for less frequently accessed data items (cold data items).
To conclude, the tree-based indexing schemes are more suitable for application
domains where information is accessed from the broadcast channel randomly, and
the signature-based indexing schemes are more suitable in retrieving sequentially
structured data elements [29,30] . In addition, tree-based indexing schemes have
shown superiority over the signature-based indexing schemes when the user request
is directed towards interrelated objects clustered on the broadcast channel(s). Fur-
thermore, tree-based indexing schemes relative to signature-based indexing schemes
are more suitable in reducing the overall power consumption. This is due to the fact
that a tree-based indexing provides global information regarding the physical loca-
tion of the data frames on the broadcast channel. On the other hand, signature-based
indexing schemes are more effective in retrieving data frames based on multiple at-
tributes.
3 . 3 S i n g l e B r o a d c a s t C h a n n e l O r g a n i z a t i o n
Organization of data objects as a means of reducing access latency has been the
subject of intensive research in the past. Whether the physical storage medium is
a flat memory or a disk rack structure, an appropriate data placement algorithm
should attempt to detect data locality and cluster related data close to one another.
The objects in an object-oriented paradigm are normally associated with one an-
other through semantic links—inheritance, aggregation, or association. An object-
clustering algorithm maps a complex object into a linear sequence of objects along
these semantic links. A complex object can be expressed as a hierarchy or a directed
acyclic graph (DAG) in which objects are represented as the vertexes (or nodes) and
edges (or links) are the relationships among these objects. Such clustering can im-
prove the response time by an order of magnitude [7,13,19] .