Performance Evaluation
In this section, we evaluate the performance of the proposed game-based cloud resource allocation.
Simulation Setup
In the simulation, there is a media cloud to lease cloud resource to brokers. The total size of cloud resource is B = 100 k which denotes that the cloud can process 100 x 103 tasks per unit time. In addition, the MSN has two brokers and two communities. In each community, there are 20 users. The media cloud waits Tw,mc = 500 for the next strategy update and each broker waits Tw,b = 100 for the next generation strategy update. The speed of adjustment for each broker on the bought resource size and price are p = 0.1 and witE = 1, respectively. The speed of adjustment for media cloud on price of resource is wr = 0.01. The detailed values of parameters in this simulation are listed in Table 5.1.
Table 5.1 Simulation parameters
|
Parameter |
Value |
|
J: the number of communities |
2 |
|
N1, N2: the number of users in community 1 and community 2 |
20,20 |
|
I: the number of cloud brokers in the network |
2 |
|
B: the maximum of resource can be allocated by media cloud |
100 k (tasks/s) |
|
[at, j, [h, j}, j 6 {1, 2}: the parameters on QoE of users in community 1 |
[2, 10} |
|
{a2, j, $2, j}, j 6 [1, 2}: the parameters on QoE of users in community 2 |
[1.5, 10} |
|
[s, j, S2, j}: the payoff parameters of mobile users in community 1 and community 2 |
[2, 2} |
|
Di: the reserved resource provided to brokers |
0 |
|
g: the discount parameter |
1 |
|
Tw,mc: the waiting time for strategy adjustment of media cloud |
500 |
|
Tw,b: the waiting time for strategy adjustment of a broker |
100 |
