Effective Cost Mechanism for Cloudlet Retransmission and Prioritized VM Scheduling Mechanism over Broker Virtual Machine Communication Framework

International Journal on Cloud Computing: Services and Architecture(IJCCSA),Vol.2, No.3, June 2012
DOI : 10.5121/ijccsa.2012.2305 41
Effective Cost Mechanism for Cloudlet
Retransmission and Prioritized VM Scheduling
Mechanism over Broker Virtual Machine
Communication Framework
Gaurav Raj1
and Sonika Setia2
1
Asst. Prof. , Lovely Professional University, Phagwara
er.gaurav.raj@gmail.com
2
M. Tech. Student, Lovely Professional University, Phagwara
sonika.setia@hotmail.com
Abstract
In current scenario cloud computing is most widely increasing platform for task execution. Lot of research
is going on to cut down the cost and execution time. In this paper, we propose an efficient algorithm to
have an effective and fast execution of task assigned by the user. We proposed an effective communication
framework between broker and virtual machine for assigning the task and fetching the results in optimum
time and cost using Broker Virtual Machine Communication Framework (BVCF). We implement it over
cloudsim under VM scheduling policies by modification based on Virtual Machine Cost. Scheduling over
Virtual Machine as well as over Cloudlets and Retransmission of Cloudlets are the basic building blocks of
the proposed work on which the whole architecture is dependent. Execution of cloudlets is being analyzed
over Round Robin and FCFS scheduling policy.
Keywords
Broker Cloud Machine Communication Framework, Cloud Computing, Cost, Scheduling, Virtual Machine.
1. Introduction
In Current Scenario, with a prior disquisition of the subject, the task is divided and disseminated
into same size cloudlets. These Cloudlets as well as Virtual Machines are scheduled according to
the First Come First Serve Scheduling Policy i.e. FCFS.

42
Fig. 1 General Cloud Computing Scenario
In general Cloud Computing scenario user submits the task to be performed / executed. Cloud
Coordinator (CC) divides the task into equal sized cloudlets and passes it to DataCenter (DC).
Due to default policy, it takes a lot of time because the cloudlets are processed and emanates one
at a time in FCFS manner as and when they reach to VM. In addition to the time consuming
factor, the cost factor also acts negatively for this scheduling policy. VM executes the cloudlets
present in the queue as they reach the VM’s. In a gist, this default policy is extremely Time-
Consuming, Cost insensitive and inefficient.
2. Related Work
Gaurav Raj [1], Author proposed Cost effective Broker Cloud Management where the lowest
cost has been calculated in accordance to the communication link. Total cost is calculated on the
basis of Hop Count, Bandwidth and Network Delay. An Optimum Route Cost Finder (ORCF)
algorithm has been proposed for finding the optimum path on the basis of cost.
Rodrigo N. Calheiros et al. [2], Authors gives a generalized framework for modeling and
experimentation of the cloud and its working and they called this framework as CloudSim. It
gives the benefit to everyone to work and analyse different scenarios over the framework. This
reduces the cost as actual implementation requires lot of cost and is almost impossible to
implement the desired thing initially at cloud. So, the framework gives the right to simulate
various cases before actual implementation.
Rajkumar Buyya et al. [3], Authors proposed the market-based resource management that
combines the customer service management and the risk management. There is also algorithm
and mechanisms for the VM allocation. It also includes market maker, market registry and their
services.
Suraj Pandey et al. [4], Authors formulate a model to reduce the execution cost of data. It shows
that the amount of data being transferred is inversely proportional to the communication cost.
Comparison has been made between costs of multiple executions of data with Amazon
CloudFront’s ‘nearest’ single data selection. Dividing the data to distributed DC’s in proportion
to their access cost results into cost savings.
User DC
CC VM-1 ………..VM-N
VMM
M

43
3. Proposed Cost Effective Framework with Scheduling Algorithm
Our main aim is to schedule the cloudlets as well as VM so that it provides more efficient and
effective results.
In BVCF, User assigns the task to CC through Broker which acts as intermediate. CC divides the
task into same sized cloudlets and sends it further to DC. Every DC has its own VM Manager
(VMM) which manages the Virtual Machine Pool (VMP). VMM requests for the resources to the
Resource Provisioner (RsP) and it then further asks for access permission from Resource
Provider (RP) and responds back to the VMM. After the permissions being granted by RP, RsP
provide resources, based on minimum cost of resources, to VMM for creating VM. The cost is
calculated using the below formula:
TC = α HC + β ND + γ B + δ SC
Here, HC is Hop Count, ND is Network Delay, B is Bandwidth and SC is Security Cost . And α, β,
γ and δ are variables [1].
Analysing the total cost, VMM prioritizes the different VM’s and chooses the VM with the
highest priority or lowest cost. The VMM sends the cloudlet ID list consists of all the ID’s of the
cloudlets received from the CC. Then, VMM sends actual cloudlets whose ID is being checked
by the VM and matched with the ID in cloudlet ID list. After receiving the accurate cloudlet VM
sends the Acknowledgement otherwise the retransmission message is sent back. After this, the
cloudlet is executed by two type of scheduling policies i.e.
Fig. 2 Broker Virtual Machine Communication Framework (BVCF)
user
Broker
CC
RsP
DC VM
M
VMP
VMP
VMP
VMP
RP
RC
N
RC
3
RC
2
RC
1

44
FCFS over cloudlets – The cloudlets will execute in the sequence as they are entering the VM.
Round-Robin over cloudlets – Initially the Time Quantum is assigned by the User and each and
every cloudlet executes for that time quantum and left cloudlets will be executed afterwards for
that time quantum only.
3.1 Proposed Scheduling and Retransmission Algorithm:
The main algorithm for the proposed architecture is divided into further three sub algorithms
which are:
i. Cost Computation Algorithm – Computation of Cost
ii. Cloudlet Retransmission and Prioritized VM Algorithm – Prioritizing the VM’s and
Sending and Retransmitting the cloudlets.
iii. Cloudlet Scheduling Algorithm – Scheduling the Cloudlets.
3.1.1 Algorithm-1:
It is named as Cost Computation Algorithm. Its basic purpose is to compute the cost at the time
of Virtual Machine creation.
3.1.2 Algorithm-2:
It is named as Cloudlet Retransmission and Prioritized VM Algorithm. Its purpose is to
acknowledge or retransmit the cloudlets which are received by the VM.
Algorithm-1 Cost Computation Algorithm
Step-1: User assigns task to Cloud Co-ordinator.
Step-2: Cloud Co-ordinator divides the task into cloudlets.
Cloudlet_list  task
Step-3: Virtual Machine Manager creates the Virtual Machine.
Step-4: At the time of creation, Resource Provider allocates the resource to Virtual
Machine.
Virtual Machine  Resource Provider [Resources]
Step-5: Total Cost is calculated keeping in mind four factors i.e. Hop count, Bandwidth,
Network
Delay and Security Cost.
TC = α HC + β ND + γ B + δ SC

45
3.1.3 Algorithm-3
It is named as Cloudlet Scheduling Algorithm. Its purpose is to schedule the cloudlets according
to FCFS and Round Robin Scheduling policy.
Algorithm-2 Cloudlet Retransmission and Prioritized VM Algorithm
Step-1: VMM sets priority of VM’s according to TC
Set Priority of VM_list[i] using VMPriority 1/ TC
Prioritized VM list = VM list
Step-2: VMM arrange / sorts the cloudlet list and store them as cloudlet_ID_list.
Step 3: Sends the cloudlet_ID_list to VM
Step-3: VMM sends cloudlet_ID_list to VM as Predicted Sequence List.
Predicted Sequence List  Cloudlet_ID_List
Step-4: VMM sends actual cloudlets to VM.
While cloudlet_list == Empty
VM_cloudlet_list[i] = VMM_cloudlet_list[i]
Step-5: If actual_cloudlet[ID] = Predicted Sequence List[ID]
Then, continue till (received_cloudlet_list).Length == 10 and
send acknowledgement[receiving_cloudlet_list[ID]].
Else wait[unit_time] and
Retransmit[cloudlet] and counter++.
Here, counter = received_cloudlet_list[ID].
Algorithm-3 Cloudlet Scheduling Algorithm
Step-1: Choose the type of Scheduling
1. Priority Based First Come First Serve Scheduling
2. Priority Based Round Robin Scheduling
User enters the choice.
Choice == 1 || 2
Step-2: If choice is 1.
Then
goto Label 1.
If choice is 2.
Then
goto Label 2.
Step-3: Executed cloudlets are returned to Cloud Co-ordinator.
CC [cloudlet_list]  VMM [executed_cloudlet_list]
Step-4: Cloud Co-ordinator combines all the cloudlets to form task.
Combine [cloudlets]
Step-5: Executed Task returned back to User by Cloud Co-ordinator.
User[executed task]  combine[cloudlets]

46
FCFS and Round Robin Scheduling are shown in Label 1 and Label 2 respectively.
3.2 Sequence Diagram for the Procedure:
1. User assigns the task to CC.
2. CC divides the assigned task into same sized cloudlets.
3. CC sends the cloudlets to VMM.
4. VMM sends the list of the needed resources to the RsP.
5. RsP requests the resources from RP.
6. RP provides access to use the resources.
7. RsP grants the access to VMM.
8. VMM creates the VM using the resources.
9. Total Cost of VM is calculated by using factors like Hop Count, Network Delay,
Bandwidth and Security Cost.
10. VMM prioritizes the VMs according to the cost factor and selects the VM with highest
priority or lowest cost.
11. VMM sends the cloudlet ID list to VM.
12. VMM sends the actual cloudlet to VM.
13. VM matches the ID of actual cloudlet with the sequence of cloudlet ID list.
14. If both ID matches then, VM sends the acknowledgement to VMM. Otherwise, VM sends
Retransmit message.
Label 1: FCFS Scheduling
For every cloudlets received
Executed_list[cloudlets]  Execution [Actual_cloudlet]
VMM[executed_cloudlet_list]  Executed_list[cloudlets]
goto Step3
Label 2: Round Robin Scheduling
[initialize] time quantum as tq = 10
Repeat Steps till (actual_cloudlet_list).Length == NULL
{
Execute[actual_cloudlet] till tq
Executed_list[cloudlets]  Execution[Actual_cloudlet]
actual_cloudlet  cloudlet_list[next_cloudlet]
}
VMM [executed_cloudlet_list]  Executed_list [cloudlets]
goto Step3

47
15. Request for execution of the cloudlet is sent to the VM by VMM.
16. Cloudlet scheduling is done by VM according to FCFS and Round Robin.
17. VM sends the executed cloudlets to the VMM.
18. VMM further passes the executed cloudlets to CC.
19. CC combines all the executed cloudlets together to form the task.
20. CC sends the executed task back to the user.
Fig. 3 Sequence Diagram of work flows for the task execution
In Fig, 3, there are six actors i.e. User, Cloud Co-ordinator, VM Manager, Virtual Machine,
Resource Provisioner and Resource Provider. The whole process takes place in eighteen steps
showing which actor is performing which task and how the assigned task is executed. All the
steps or method shown in the sequence diagram are explained below as:

48
3.3 Use Case Diagram of the Procedure:
There are six actors in the above Use Case Diagram:
a) User
b) Cloud Coordinator (CC)
c) Virtual Machine Manager or VM Manager (VMM)
d) Virtual Machine (VM)
e) Resource Provisioner (RsP)
f) Resource Provider or Resource Owner(RP)
Fig. 4 Use Case Diagram of BVCF, showing all the actors and the work performed by the actors.
Cloud Coordinator itself divides or combines the task to cloudlets so it is not connected to other
actor.
Similarly, VM Scheduling is done by VM Manager and Cloudlet Scheduling is done in Virtual
Machine.

49
Include Relationship is shown if the use case further have some use case. Like in this diagram,
Schedule VM includes the Prioritizing VM’s using Total Cost. This means schedule VM depends
upon the prioritization of VM using Total Cost.
4. Results
After implementing the proposed algorithms over BVCF, results for computation of cost are
obtained which further helps in the study of comparison between the FCFS and Round Robin
Scheduling policy.
The whole procedure is being done with the two different scheduling policies and the cost of both
the procedures is shown in Fig 5.
0
5000
10000
15000
20000
25000
FCFS Round Robin
Total Cost
Fig. 5 Comparison between FCFS and Round Robin
Same numbers of virtual machines are considered while calculating total cost for both cases i.e.
FCFS and Round Robin. The location of user i.e. from where the user is demanding the services
is also considered same for both the cases. The total cost for executing all cloudlets is more in
case of FCFS scheduling policies. And due to the Acknowledgment and Retransmission criteria,
all cloudlets must run. This is why no cloudlet results in paused mode, or stopped mode.
Fig. 5 shows that the total cost of all cloudlets being executed with the help of Round Robin is
comparatively lower than that of being executed by FCFS scheduling policy. So, the proposed
algorithm succeeds in reducing the cost.
5. Conclusion
Cost and time are always the grassroots that come into mind either before buying or selling any
services. In this paper, we have proposed an algorithm for optimized Cost of VM, prioritizing the
VM and cloudlet scheduling through FCFS and Round Robin. BVCF explains the whole
proposed scenario in which the execution of task is performed. The proposed algorithm helps in
the Fast Execution due to Round Robin Scheduling Policy applied on the equally sized cloudlets.
All cloudlets will execute as after each and every successfully received cloudlet VM sends the

50
acknowledgement and for the unsuccessful cloudlets sends the retransmit message. It also results
into Lower Cost as the VM’s are prioritized according to its Cost only.
6. Future Work
Cloud Computing is a perpetual research as all its aspects haven’t unfolded yet. BVCF is just a
framework where most of the mechanisms can be implemented so as to analyze or optimize the
performance of any of the midget area. Security is the major issue in every area and we can also
implement various security mechanisms over BVCF. Resources can also be allocated accordingly
to reduce the cost. VM’s images can be saved in the database so that if the same task is being
assigned by the user then we can allocate the same VM with optimal resources. Cost of VM can
also include certain other factors on which it is dependent.
References
[1] Gaurav Raj, (2011), An Effective Broker Cloud Management System, at ACAI'11 in ACM
Conference.
[2] Rodrigo N. Calheiros, Rajiv Ranjan1, Cesar A. F. De Rose, and Raj Kumar Buyya, (2009),
CloudSim: A Novel Framework for Modeling and Simulation of Cloud Computing Infrastructures
and Services, at IEEE, 2009.
[3] Buyya, R., Chee Shin Yeo, Venugopal S., (2008), Market-Oriented Cloud Computing: Vision, Hype,
and Reality for Delivering IT Services as Computing Utilities, at IEEE.
[4] Suraj Pandey, Kapil Kumar Gupta, Adam Barker, and Rajkumar Buyya, (2009), Minimizing Cost
when using Globally Distributed Cloud Services: A Case Study in Analysis of Intrusion Detection
Workflow Application, Technical Report, CLOUDS-TR-2009-6, Cloud Computing and Distributed
Systems Laboratory, The University of Melbourne, Australia.
[5] Abirami S.P., Shalini Ramanathan, (2012), Linear Scheduling Strategy for Resource Allocation in
Cloud Environment, at IJCCSA.
Authors
1. Sonika Setia
Student of M.Tech in Information Technology from Lovely Professional University,
Phagwara.
2. Gaurav Raj
Asst. Prof. at Lovely Professional University, Phagwara and has done M.Tech from Motilal
Nehru National Institute of Technology, Allahabad.

Effective Cost Mechanism for Cloudlet Retransmission and Prioritized VM Scheduling Mechanism over Broker Virtual Machine Communication Framework

More Related Content

Similar to Effective Cost Mechanism for Cloudlet Retransmission and Prioritized VM Scheduling Mechanism over Broker Virtual Machine Communication Framework (20)

Recently uploaded (20)

Effective Cost Mechanism for Cloudlet Retransmission and Prioritized VM Scheduling Mechanism over Broker Virtual Machine Communication Framework