Database Reference
In-Depth Information
4 Algorithms for Nash Equilibrium Semantic Cache
Schemes
This section discusses the concrete implementation of the algorithms. Our algo-
rithm is consisted of three sub-algorithms:
1) Network Manager Algorithm:
The Network Manager Algorithm manages
the network among the clients and the servers, which includes three functions:
Node Move Function, Building Routing Table Function, and Searching Routing
Table Function.
2) Query Processor Algorithm:
Query Processor Algorithm generates queries
and processes queries via the semantic caches and servers. It randomly generates
a query that is a data structure. Afterward, firstly it queries to locate the node,
then it queries in the neighboring node and finally it queries to the server.
3) Semantic Cache Manager Algorithm:
Thecorepartofourworkfocus
on the Semantic Cache Manager Algorithm, which calculates semantic cache
scalar coecients, manages the store and replaces of the semantic cache. First,
it calculates the Semantic Weight, Time Weight, and Location Weight of every
semantic-cache block using the basic information about frequency, size, cost,
time, and location. Second, it finds the inserting place and calculates the store
space. Finally if the space is large enough, then store, otherwise replace.
5 Performance Evaluation
We examine the performance of the proposed scheme in a mobile computing en-
vironment through a simulation study. Our simulator is developed by ourselves
and implemented in C++, which simulates a simple but typical mobile comput-
ing model. Then, we describe the conducted experiments, and further analyze
the results. The simulation experiments are executed on an IBMx255 server that
runs Linux with four Intel Xeon MP 3.0GHz/400MHz processors and 18G DDR
main memory.
5.1 Simulation Environment
In the simulation environment, we set two servers and multiple mobile clients,
which use a wireless link among them. They move in the range (
X
1
,X
2
,Y
1
,Y
2).
We assume that at least one or two servers maintain a complete copy of the
database and act as a sink and a database server. The queries are submitted
by the different mobile nodes in different situations, including the location, time
and requirement. The semantic cache is located on the mobile nodes. First a
query is processed locally; second the query is sent to the neighboring nodes for
processing, when it cannot be answered via the cache; finally, it is sent to the
server nodes, when it cannot be answered via the neighboring nodes.
Mobile Clients Model:
The mobile clients are simulated to be composed of
the following modules: network manager, which manages the network among
the mobile nodes and the central servers; query processor, which generates and
Search WWH ::
Custom Search