Broadcasting a Means to Disseminate Public Data in a Wireless Environment—Issues and Solutions - Advances in Computers - page 15

F IG . 4. A broadcast sequence.
up to the present time, i.e., 100. Applying the aforementioned equation will result
in G 1 = 12 . 5, G 2 = 9 . 18, and G 3 = . 5. As a result, item 1 is the next item to be
broadcast.
Finally, the scope of this allocation algorithm has also been extended to the domain
of the multiple broadcast channels.
3.1.2 Hybrid-Based Schemes
Studies reported in [8,52] have shown that push based model and pull-based model
have complementary properties. In general, two conditions support a push-based
model: (i) strict coherency requirements and/or (ii) the communication bottleneck.
On the other hand, a pull-based model is suited in an environment where: (i) client
has relaxed coherency requirements, (ii) the servers are heavily loaded, or (iii) fault
tolerance is important. This study motivates the hybrid data dissemination, where the
push and pull techniques for data dissemination are mixed. The users in the network
play a more active role in deciding which data to be broadcast. The pushed data items
are still the main means to disseminate public data. However, with the integration of
an independent back channel, clients can directly query the server database and pull
the requested data from the air channel. It should be noted that, in this class of data
dissemination methods, unlike flat broadcasting and broadcast disks, all of the data
base contents are not broadcast during a broadcast cycle.
The hybrid broadcasting of data items was originally proposed and employed in
the Boston Community Information System [24] —broadcast and interactive commu-
nication to provide up-to-the-minute information to an entire metropolitan area. The
experiment showed that users valued both components of the hybrid organization.
Wong and Dykeman [63] also proposed hybrid teletext-videotex platform for both
periodically pushed and upon-request pulled data items with some ad hoc partition
of the data into two groups. Acharya et al. [3] proposed a back channel to allow the
clients to explicitly request data items on the broadcast channel.
3.1.2.1 Adaptive Data Broadcast. This work was motivated by two is-
sues; the dynamic nature of the clients' needs and requests and the passive nature of
the clients receiving the broadcast. Similar to the concept of cache memories that
is adaptive to the changing workload, the push-pull based broadcast model [60]