This section highlights the contribution made by the author’s in this paper. The research mainly aims to improve the task allocation system of VM’s in cloud data center by enhancing load balancing approach. Due to the modification in tasks allocation system different parameter of time consumptions improved in cloud datacenter. Figure 2 presents the frame work of the proposed approach. Our contribution of the paper can be summarized as below.
[1] A survey of existing improvement in different section of ABC algorithm used for load balancing purpose.
[2] The proposed algorithm modification sections are discussed with explanation.
[3] The proposed algorithm addresses the VM violation issue like under/overloaded task distribution in cloud datacenter.
[4] The proposed algorithm improved different time consumptions parameters in cloud data center.
[5] The proposed algorithm results are comparing with well-known algorithm.
[6] Discuss the result along with future implementation of the proposed algorithm.
3.1 Improved Employed bee in the proposed HMBC Algorithm
The proposed HMBC Algorithm inherits the basic idea from ABC algorithm and tries to implements load balancing technique for improvements in task allocation system of VM in the cloud computing. Different parameters like makespan, total processing time, response time of algorithms, response time of datacenter and degree of imbalance are used to check the performance of the proposed algorithm in the cloud computing. For that reason, three main sections of ABC algorithm are modified in the proposed HMBC i.e. onlooker bees, employed bees and load balancer. Normally, an ABC algorithm consists of different parameters which are employed bees, onlooker bees, and scout bees. During the execution process, they change into sub-groups. Initialization of the population consist of two groups of bees which are known as employed and onlooker both containing the same number of bees. When we analyzed the algorithm, employed bee and onlooker bee used same equation for obtaining the candidate solution which makes an imbalanced choice or worst case can occur due to the previous update rule [22]. Therefore, we used new update rules in this section which we derived from Bat algorithm update rule. Bat algorithm use virtual Bat for the update rules for their position or food source as given in the Eqs. 1 and 2.
$${ \text{V}}_{\text{i}}^{\text{t}}={\text{V}}_{\text{i}}^{\text{t}-1}+({\text{X}}_{\text{i}}^{\text{t}-1}-{\text{X}}_{\text{*}}){\text{f}}_{\text{i}}$$
1
$${ \text{X}}_{\text{i}}^{\text{t}}={\text{X}}_{\text{i}}^{\text{t}-1}-{\text{V}}_{\text{i}}^{\text{t}}$$
2
Where \({\text{V}}_{\text{i}}^{\text{t}}\) present the velocity and \({\text{X}}_{\text{i}}^{\text{t}}\) present the location at t (cycle) iteration in space or possible dimension solution [23, 24]. Table 4 shows the Pseudocode of improved employed bee in the proposed algorithm.
Table .4.Improved Employed bee in the proposed Algorithm
Algorithm 2: Employed bee section
|
1
2
|
For i = 1 to n
Sen the population of bees in VM at random search method.
|
3
|
\({ V}_{i}^{t}={V}_{i}^{t-1}+({X}_{i}^{t-1}-{X}_{\text{*}}){f}_{i}\)
\({ X}_{i}^{t}={X}_{i}^{t-1}-{V}_{i}^{t}\)
|
4
|
Calculate the fitness value of each VM\({f}_{j }=({p}_{j}\times {m}_{j}{bw}_{j}\)
Where j = index of VM that ith bees found.
|
5
|
End for.
|
3.2 Proposed Algorithm selection approach for Onlooker bee
When the employed bee of ABC algorithm exploits the nectar source and shares all the information upon returning to beehive about the food source by making wangle dance. As mentioned in the Eq. (3) with the roulette wheel selection the onlooker bee gets the information with probability and the fitness value of the solution (quality of food source). Onlooker bee try better food sources as per rules according to the current cycle. After the first cycle, the successive cycles start with employed bee moving to the food source position update by onlooker bee. During this process, some of the food source remains unchanged for predetermined number of trails which causes time consumption in DC’s. Therefore, we improve the search selection approach for onlooker bee by taking range sector along with the fitness value [25]. So, the average fitness value obtained from the employed bee is calculated as.
$${ Avg}_{t}^{pit}\frac{I}{n}{\sum }_{i=0}^{n}{fit}_{i}$$
3
Table .5. Selection approach for onlooker bee in the proposed Algorithm
Algorithm 2:Onlooker bee phase
|
1
|
The first (m) onlooker bees as food source (VM) are choice with the highest fitness value and performed neighborhood search.
|
2
|
Numbers of employed bee are sent in to new position for food source and when they return the share the information.
|
3
|
\({ Avg}_{t}^{pit}\frac{I}{n}{\sum }_{i=0}^{n}{fit}_{i}\)
|
4
|
The onlooker bee calculate the fitness value of food source according eq in line 5.
|
5
|
\({\text{f}\text{i}\text{t}}_{\text{i}\text{j}}=\frac{\sum _{\text{i}=1}^{\text{n}}\text{d}\text{a}\text{t}\text{a} \text{l}\text{e}\text{n}\text{g}\text{t}\text{h} \text{i},\text{j}}{ \text{d}\text{a}\text{t}\text{a} \text{s}\text{i}\text{z}\text{e} }\)
|
6
|
The onlooker bees choose the best food source (VM) and assign task to VM.
|
3.3 Load balancer section in the proposed HMBC Algorithm
After the completion of all iterations, task is assigned to different VM’s using load balancer section. Based on the standard deviation, it checks the processing time of a single VM or a group of VM’s. Here, a VM is also checked for overloaded/under loaded. It is calculated by using the standard deviation which is given below.
$$\text{S}\text{D}=\sqrt{\frac{1}{\text{n}}\sum _{\text{j}=0}^{\text{n}}\left(\text{X}\text{j}-\text{X}\right)2}$$
4
Processing time of data by a single VM can be measured by;
$${ \text{X}}_{\text{j}}={\sum }_{1}^{\text{d}}\frac{\text{d}\text{a}\text{t}\text{a}-\text{l}\text{e}\text{n}\text{g}\text{t}\text{h} \left(\text{i}\right)}{\text{C}\text{a}\text{p}\text{a}\text{c}\text{i}\text{t}\text{y} }$$
5
Means of processing time of all VMs can be calculated as;
$${ X}_{j}={\sum }_{1}^{\text{d}}\frac{\text{d}\text{a}\text{t}\text{a}-\text{l}\text{e}\text{n}\text{g}\text{t}\text{h} \left(\text{i}\right)}{\text{n}-1}$$
6
In the given section load balancer section, standard deviation is used for the distribution of data to a single or a group of VM’s in a cloud data center [26]. Figure 2 shows the working of proposed algorithm.