An Energy & Cost Efficient Task Consolidation Algorithm for Cloud Computing Systems

Abstract

The power consumption of untapped resources, especially during a cloud background, represents a significant sum of the specific power use. By its nature, a resource allotment approach that takes into account the use of resources would direct to better power efficiency; this, in clouds, expands even additional, and with virtualization techniques often jobs are easily combined. Job consolidation is an effective way to expand the use of resources and sequentially reduce power consumption. Current studies have determined that server power utilization extends linearly with processor resources. This hopeful fact highlights the importance of the involvement of standardization to reduce energy utilization. However, merging tasks can also cause freedom from resources that will remain idle as the attraction continues. There are some remarkable efforts to decrease idle energy draw, usually by putting computer resources into some kind of power-saving/sleep mode. Throughout this article, we represent 2 power-conscious task reinforcement approaches to maximize resource use and explicitly consider both passive and active power consumption. Our inferences map each job to the resource at which the power consumption to perform the job is implicitly or explicitly reduced without degrading the performance of that task. Supporting our investigational outcome, our inference methods reveal the most promising power-saving potential.KeywordsLoad balancingCloud computingPower-aware computing

Full text

An Energy & Cost Efficient Task Consolidation
Algorithm for Cloud Computing Systems
Sachin Kumar1(B), Saurabh Pal1, Satya Singh2, Raghvendra Pratap Singh3,
Sanjay Kumar Singh3, and Priya Jaiswal1
1Department of Computer Applications, V.B.S.P.U, Jaunpur, U.P., India
jaiswalsachin009@gmail.com
2Department of Computer Science and Applications, M.G.K.V.P, Varanasi, U.P., India
3Department of Computer Science and Engineering, Kashi Institute of Technology-Varanasi,
Varanasi, U.P., India
Abstract. The power consumption of untapped resources, especially during a
cloud background, represents a significant sum of the specific power use. By its
nature, a resource allotment approach that takes into account the use of resources
would direct to better power efficiency; this, in clouds, expands even additional,
and with virtualization techniques often jobs are easily combined. Job consolida-
tion is an effective way to expand the use of resources and sequentially reduce
power consumption. Current studies have determined that server power utiliza-
tion extends linearly with processor resources. This hopeful fact highlights the
importance of the involvement of standardization to reduce energy utilization.
However, merging tasks can also cause freedom from resources that will remain
idle as the attraction continues. There are some remarkable efforts to decrease
idle energy draw, usually by putting computer resources into some kind of power-
saving/sleep mode. Throughout this article, we represent 2 power-conscious task
reinforcement approaches to maximize resource use and explicitly consider both
passive and activepower consumption. Our inferences map each job to the resource
at which the power consumption to perform the job is implicitly or explicitly
reduced without degrading the performance of that task. Supporting our investi-
gational outcome, our inference methods reveal the most promising power-saving
potential.
Keywords: Load balancing ·Cloud computing ·Power-aware computing
1 Introduction
Cloud computing is a beneficial model for both providers and consumers. Cloud com-
puting usually contains several applications that may be heterogeneous and distributed.
Virtualization technology has made cloud computing more fruitful. The deployments of
cloud applications have several advantages like reliability and scalability; on the other
hand at its core, the cloud aims to provide more cost-effective solutions for providers as
well as consumers. In economic terms, buyers should buy resources as per their require-
ments while service providers can make good use of underutilized cloud resources. As
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022
S. Rajagopal et al. (Eds.): ASCIS 2022, CCIS 1759, pp. 446–454, 2022.
https://doi.org/10.1007/978-3-031-23092-9_35

An Energy & Cost Efficient Task Consolidation Algorithm 447
per the service provider’s point of view, maximization of profits may be a higher pri-
ority. For this purpose, reducing the consumption of energy plays an important role.
Alternatively, by increasing the use of resources, service providers can reduce power
consumption.
Resource use and energy consumption in cloud computing are extremely correlated.
Especially, computing resources use a large amount of power for their low utilization
in comparison to the sufficiently used loader. As per the latest review [1–4], task con-
solidation is an efficient method for increasing the use of resources and reducing power
consumption. Task consolidation technology is largely used by cloud technologies that
make it easier to run multiple applications on a single cloud platform at the same time.
Recent studies have determined that server power consumption is measured sequen-
tially with resource usage [5–7]. This information also calls for a significant contribution
to the standardization of ta sks in reducing energy consumption. However, merging tasks
can also free up resources that could exist idling however still pulling force. Our infer-
ence sets every task for the cloud resource in which power consumption is reduced to
perform the job without any deterioration in performance.
We have calculated results based on the objective function for energy consumption.
This means when more than one task was combined for a single resource only then
the consumption of energy will be drastically decreased. Our inductive methods show
promise in the ability to save energy.
The remainder of the research document is prepared in the following manner.
Section 2explains the cloud applications, energy, and task integration model used in
this research article. Section 3explains relevant work. Inference in MaxUtil and ECTC
is described in Sect. 4followed by outcomes and wrapping up of performance evaluation
in Sects. 5& 6 respectively.
2 Related Works
The paradigms of green and cloud computing are interrelated and growing. Cloud energy
efficiency has become one of the major research challenges. Advances in computer
hardware equipment [8], like solid-state drives, low power energy-efficient CPUs, and
monitors of computers have assisted a lot to alleviate this energy problem to a certain
extent. Meanwhile, energy issues are also handled by using several software approaches
like allocation of resources [9–15] and standardization of tasks [16–19].
The allocation of resource and scheduling policy is mainly facilitated using the
grace period retrieval with the support of a dynamic potentiometer [20] is integrated into
many processors. This technology provisionally reduces the supply voltage to re duce
the computation speed.
The task consolidation in [16] is handled using a traditional container filling problem
with 2 main properties, for example, disk usage and CPU. The algorithm proposed in
[16] attempts to standardize performance and power consumption.
In [17], an analytical model has been presented to standardize internet-oriented tasks.
The model takes into account the functions required for services such as e-commerce
network services or an e-book database. The main objective is to maximize the use of
resources to reduce power consumption.

448 S. Kumar et al.
The mechanism for consolidating tasks was developed in [18,19] to manage power
declination through various techniques, such as [18]. Unlike computing job consolidation
strategies, the approach used in [18] adopts 2 techniques, memory compression and asked
for discrimination.
A virtual power approach is suggested in [19] the integration of tasks in the integration
of energy management “Hard” and “Soft” scaling techniques. These 2 techniques belong
to energy management utilities equipped with physical processors and virtual machines.
In [21], a supportive Nash bargaining and game model is introduced to deal with
the network load balancing issue. The key goal is to reduce power utilization while
preserving the limited quality of services, such as time.
In [22], a similar task has been performed as in [21] in that they deal with fixed
scheduling situations with independent functions. Additionally, both take advantage of
dynamic voltage frequency scaling energy reduction technology.
3 Task Consolidation Algorithm
Job consolidation is an efficient way to manage computing resources, mainly in the
long and short term cloud. As for the short-term is concerned, volume flows on arriving
jobs can be treated as “power-saving” by decreasing the number of running computing
resources, and planting excess computing resources into the energy-saving mode, or
by systematically switching off some non working computing resources. For the long
period of time, cloud service providers should follow energy saving models; it relieves
the excessive load of computing operational rate due to increased provisions. The main
orientation of this research article; despite of the result of merging tasks; our method
may be used for file estimation.
In this part, we have presented 2 power-conscious job fusing methods, MaxUtil and
ECTC. In this maxUtil unifies resourceful decision-making tasks; it is one of the key
indicators of power efficiency under our settings.
Description of the Algorithm
MaxUtil and ECTC trail analogous steps with the main difference in their cost functions
(Fig. 2). In short for a given job, 2 inferences are validated for each computing resources
and select the most power-efficient computing resource for the job. Assessment of the
most power efficient computing resource depends on the heuristic used. The real power
utilization of the existing job is calculated by cost function of ECTC by subtracting the
minimum power (pmin) utilization. No power consumption in the overlapping period
of time between those jobs and the present job is taken into account explicitly. The job
function tends to distinguish a job that is performed alone. The Fi,j value of a job tj is
defined on the computing resource ri attained through the cost function of the ECTC as
follows:
Fi,j={(P×Vj +Pmin)×τ0)−((P×Vj +Pmin)+τ1+P×Vj ×τ2)}
(1)
The relation for this cost function is that the power consumption at its lowest use is much
larger than that at its idle state.

An Energy & Cost Efficient Task Consolidation Algorithm 449
Input: A set R of r cloud resources and a task tj and
Output: A task-resource match
Fig. 1. Description of Algorithm
The cost function of MaxUtil is considered with average use during computation
time for the existing job-as an essential component of it. This cost function intends to
amplify the intensity of uniformity. The 1st benefit is to reduce power utilization. The 2nd
benefit is that MaxUtil’s cost function implicitly reduces the number of active computing
resources because it tends to intensify the use of few computing resources compared to
the cost of the ECTC function. The value Fi,j of the task tj on the computing resource ri
using cost function of MaxUtil’s is given below:
Fi,j = τ0
τ=1Ui
τ0(2)
3.1 Discussion and Analysis of performance
As integrated into our power model, the consumption of power is directly proportional
to the use of resources. At a glance, for any 2 task computation resource matches, 1 with
higher use can be chosen. On the other hand, determining the correct matching is not
completely reliant on the existing job. The decision made by ECTC is based that instead
of consuming (the only) energy for that job. In Fig. 3a, job 3 (t3) arrives at 14 s after
job zero, job one, job 2 and is assigned to computing resource one (r1) based on power
utilization, still the utilization of computing resources uses 0(r0) is more.
4 Experimental Evaluations
In the experimental evaluation section, we have explained the settings and methods
including the characteristics of the job and their creation. Then the experimental obser-
vation is presented based on the consumption of energy. While computing resource usage
may be a better measurement of performance, Average usage rates are not presented on
all resources because they are already presented by consumption of energy.

450 S. Kumar et al.
Fig. 2. Consolidation examples for tasks in Table 1using ECTC
Fig. 3. Consolidation examples for tasks in Table 1using MaxUtil
Table 1. Task properties
Tas k Processing time Arrival time Utilization %
Zero Twe n t y Zero Forty
One Eight Three Fifty
Two Twenty Three Seven Twe nty
Three Ten Fourteen Forty
Four Fifteen Twe n t y Seventy
4.1 Experiments
The performance of MaxUtil and ECTC is calculated thoroughly using a large number
of experiments using a variety of tasks. Along with the characteristics of the task, we
have used three algorithms i.e. ECTC, MaxUtil, and random. These three algorithms
have been with the integration of job migration.
Since (as far as we know) the current task merging algorithms are cannot be directly
compared with our inference, comparisons were made between Randomization, MaxU-
til, and ECTC. Especially most power -saving technologies are closely related to specific
deadlines and/or interrelated tasks; In addition, it does express the relationship between
energy consumption and resource utilization i.e. standardization of tasks is not taken
into account. This is the current task of introducing unification techniques in the Sect. 3
fundamental differences appear from our inference in scheduling and power models.

An Energy & Cost Efficient Task Consolidation Algorithm 451
As per the early experiments with those 3 experimental methods (MaxUtil, ECTC,
and randomized), we noticed that in some conditions the transfer of a few jobs can
decrease power consumption. This result encouraged us to implement randomization,
ECTC, MaxUtil for task migrations. This transfer is taking into account for every running
job at what time the use of resources varies, i.e. get the job done or start the job.
4.2 Results
The outcomes achieved from widespread simulations are explained in Table 2.The
outcome of different cloud resources is shown in Fig. 4. Nevertheless, Simulation on
cloud platform performed with fifty dissimilar jobs as mentioned in Sect. 5.1.
The savings of power in Table 2are relative rates of the outcomes attained from the
experimentation carried out by means of random algorithms. These results are shown in
Table 2. Figure 4shows the capability of energy saving of Max and ECTC in general.
MaxUtil and ECTC Outperform stochastic algorithms – whatever their dependence on
immigration – by 18% and 13%, correspondingly.
Saving energy with uncertainty and high resource usage is still in demand. The use
of cloud resources is most appropriate for the consolidation of tasks as presented in
Fig. 4a. It is basically because the jobs that have been carried over tend to have a short
duration remaining to process & these jobs are creating a hindrance for upcoming new
jobs. As a result, more power depreciation is there in comparison when immigration is
not considered.
Table 2. Relative energy saving
Usage Pattern MaxUtil Algorithm
Low High Random
Energy Saving % Twenty Five Twenty Three Four Five Twelve Eleven
Migration No Yes No Yes No Ye s
Average % Thirteen
Usage Pattern ECTC Algorithm
Low High Random
Energy Saving % Twirty Three Thirty Two Nine Nine Seventeen Sixteen
Migration No Yes No Yes No Ye s
Average % Eighteen

452 S. Kumar et al.
Fig. 4. Consumption of energy using task consolidation approaches: (a) Low Resource Utilization
(b) High Resource Utilization (c) Random Resource Utilization
5 Conclusion
TASK consolidation especially in cloud computing is a significant method to develop
energy efficiency. As per the fact that e nergy consumption is directly related to resource
utilization, we succeeded in modeling their association and developing two energy-
aware task inferences. Cost functions are effectively integrated into these inferences
Energy-saving capabilities and capacity demonstrated by our experimental evaluation.
The outcome of this article will not only reduce the electricity invoices for cloud service
providers, but it also involves potential earnings by saving operating charges. It also
plays an important role in th e reduction of carbon footprint in cloud computing.

An Energy & Cost Efficient Task Consolidation Algorithm 453
References
1. Singh, P., Prakash, V., Bathla, G., Singh, R.K.: QoS aware task consolidation approach for
maintaining SLA violations in cloud computing. Comput. Electr. Eng. 99, 107789 (2022)
2. Nayak, S.K., Panda, S.K., Das, S., Pande, S.K.: A renewable energy-based task consolidation
algorithm for cloud computing. In Control Applications in Modern Power System, pp. 453–
463. Springer, Singapore (2021)
3. Pattnayak, P.: Optimizing power saving in cloud computing environments through server
consolidation. In: Advances in Micro-Electronics, Embedded Systems and IoT, pp. 325–336.
Springer, Singapore (2022)
4. Arshad, U., Aleem, M., Srivastava, G., Lin, J.C.W.: Utilizing power consumption and SLA
violations using dynamic VM consolidation in cloud data centers. Renew. Sustain. Energy
Rev. 167, 112782 (2022)
5. Varvello, M., Katevas, K., Plesa, M., Haddadi, H., Bustamante, F., Livshits, B.: BatteryLab:
A collaborative platform for power monitoring. In: International Conference on Passive and
Active Network Measurement (pp. 97–121). Springer, Cham (2022, March)
6. Bustamante, F., Livshits, B.: BatteryLab: a collaborative platform for power monitoring. In:
Passive and Active Measurement: 23rd International Conference, PAM 2022, Virtual Event,
March 28–30, 2022: Proceedings (Vol. 13210, p. 97). Springer Nature (2022)
7. Song, M., Lee, Y., Kim, K.: Reward-oriented task offloading under limited edge server power
for multiaccess edge computing. IEEE Internet Things J. 8(17) 13425–13438 (2021)
8. Venkatachalam, V., Franz, M.: Power reduction techniques for microprocessor systems. ACM
Computing Surveys (CSUR) 37(3), 195–237 (2005)
9. Bal, P.K., Mohapatra, S.K., Das, T.K., Srinivasan, K., Hu, Y.C.: A Joint Resource allocation,
security with efficient task scheduling in cloud computing using hybrid machine learning
techniques. Sensors 22(3), 1242 (2022)
10. Al-Wesabi, F.N., Obayya, M., Hamza, M.A., Alzahrani, J.S., Gupta, D., Kumar, S.: Energy
aware resource optimization using unified metaheuristic optimization algorithm allocation
for cloud computing environment. Sustain. Comput.: Inform. Syst. 35, 100686 (2022)
11. Nanjappan, M., Albert, P.: Hybrid-based novel approach for resource scheduling using
MCFCM and PSO in cloud computing environment. Concurr. Comput.: Pract. Exp. 34(7),
e5517 (2022)
12. Kumar, C., Marston, S., Sen, R., Narisetty, A.: Greening the cloud: a load balancing
mechanism to optimize cloud computing networks. J. Manag. Inf. Syst. 39(2), 513–541 (2022)
13. Belgacem, A.: Dynamic resource allocation in cloud computing: analysis and taxonomies.
Computing 104(3), 681–710 (2021). https://doi.org/10.1007/s00607-021-01045-2
14. Peng, K., Huang, H., Zhao, B., Jolfaei, A., Xu, X., Bilal, M.: Intelligent computation offloading
and resource allocation in IIoT with end-edge-cloud computing Using NSGA-III. IEEE Trans.
Netw. Sci. Eng. (2022)
15. Wadhwa, H., Aron, R.: TRAM: Technique for resource allocation and management in fog
computing environment. J. Supercomput. 78(1), 667–690 (2021). https://doi.org/10.1007/s11
227-021-03885-3
16. Srikantaiah, S., Kansal, A., Zhao, F.: Energy aware consolidation for cloud computing (2008)
17. Song, Y., Zhang, Y., Sun, Y., Shi, W.: Utility analysis for internet-oriented server consolidation
in VM-based data centers. In: 2009 IEEE International Conference on Cluster Computing
and Workshops, pp. 1–10. IEEE (2009, August)
18. Torres, J., Carrera, D., Hogan, K., Gavaldà, R., Beltran, V., Poggi, N.: Reducing wasted
resources to help achieve green data centers. In: 2008 IEEE International Symposium on
Parallel and Distributed Processing, pp. 1–8. IEEE (2008, April)

454 S. Kumar et al.
19. Nathuji, R., Schwan, K.: Virtualpower: coordinated power management in virtualized
enterprise systems. ACM SIGOPS Oper. Syst. Rev. 41 6 265 278 (2007)
20. Kuroda, T., et al.: Variable supply-voltage scheme for low-power high-speed CMOS digital
design. IEEE J. Solid-State Circuits 33(3), 454–462 (1998)
21. Subrata, R., Zomaya, A.Y., Landfeldt, B.: Cooperative power-aware scheduling in grid
computing environments. J. Parallel Distrib. Comput. 70(2), 84–91 (2010)
22. Khan, S.U., Ahmad, I.: A cooperative game theoretical technique for joint optimization of
energy consumption and response time in computational grids. IEEE Trans. Parallel Distrib.
Syst. 20 3 346 360 (2008)