- Email: [email protected]
A systematic mapping study on soft computing techniques to cloud environment
Available online at www.sciencedirect.com
ScienceDirect
Available online at www.sciencedirect.com Procedia Computer Science 00 (2018) 000–000
ScienceDirect
www.elsevier.com/locate/procedia
Procedia Computer Science 120 (2017) 31–38
9th 9thInternational InternationalConference Conferenceon onTheory Theoryand andApplication Applicationof ofSoft SoftComputing, Computing,Computing Computingwith with Words and Perception, ICSCCW 2017, 24-25 22-23 August August 2017, 2017, Budapest, Budapest, Hungary Hungary
A systematic mapping study on soft computing techniques to cloud environment Obinna H. Ejimogua, Seren Başarana* a
Department of Computer Information Systems, Near East University,POBOX:99138,Nicosia, North Cyprus, Mersin 10, Turkey.
Abstract Cloud computing plays an essential role in storage and transfer of big capacity data due to a rapid increase in size and the number of organizational activities. There exist numerous studies in which diverse soft computing techniques are applied to the cloud environment. The relevant extant literature that were clustered into five main categories with respect to precedence are; task optimization, power optimization, security, service selection and cost optimization. Yet, it was discovered that there is a dearth of systematic review/mapping studies particularly on soft computing techniques in cloud environment so as to obtain exclusive insight, to identify existing gaps and future research directions. Therefore the aim of this paper is to conduct a systematic mapping study of recent literature on soft computing techniques in cloud environment. For this purpose, 163 articles were chosen as primary sources that were published within the last decade, which were classified based on study focus area, type of research, contribution facet and particularly the type of soft computing technique used. Findings revealed that task optimization takes part as the highly preferred research focus area. Secondly, most of the articles found are of validation studies. The contributions of most of the studies are concerned about methods and finally the top three soft computing techniques were detected as particle swarm optimization (PSO), genetic algorithm (GA) and hybrid systems. The results of this study confirm that applying soft computing techniques in cloud computing has gained more and more significant attention recently but there still remain challenges and gaps which calls for further investigation especially in the area of cost optimization and also artificial bee colony. © 2018 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. Keywords:Cloud computing; power optimiztion; soft computing; systematic mapping; task optimization.
* Seren Başaran. Tel.: +90- 392-675-1000(3121); fax: +90-392-6751051. E-mail address: [email protected] 1877-0509© 2018 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. 1877-0509 © 2018 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. 10.1016/j.procs.2017.11.207
32 2
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
1. Introduction Recently cloud computing environment has drawn remarkable attention from organizations because of its capabilities to reduce cost, maximize productivity with high degree of flexibility in handling big data (Pueschel et al., 2009). Considering the burden of manual management in ever growing capacity of cloud data, there has always been a need to seek for more efficient autonomous and computational intelligent solutions to handle the dynamic nature of the cloud. Over the last decades, numerous soft computing techniques have been proposed and utilized for automated optimization of the cloud environment. The prevalent types of soft computing techniques which addressed mainly within the cloud computing related studies about security challenges, task optimization, service selection cost optimization and power optimization include; fuzzy logic, neural networks, genetic algorithm, swarm intelligence and hybrid systems (Esposito et al., 2015; Feller et al., 2011; Ramezani et al., 2014). Although soft computing techniques have caught the interests of many researchers over the years, yet to author’s knowledge, there is currently no systematic mapping study conducted particularly to soft computing in cloud environment. A systematic mapping study not only helps to identify the extent of study in a research area, but also provides a view to pinpoint existing gaps in a particular subject area and summarizes the entire corpus of study in a subject area to encourage further development. In the light of above, a systematic mapping study was conducted to gather, analyze and interpret existing studies that were carried out in the area of soft computing in cloud environment so as to provide an overview and highlight research gaps in the subject area. 2. Motivation and related surveys Although soft computing techniques are popular and have been widely applied to cloud environments, initial attempts to locate relevant systematic mapping studies in the area of soft computing techniques in cloud environment proved to be abortive, which became the fundamental motivation to carry out this systematic study. The only located literature survey studies were briefly summarized as follows: A study conducted by Hormozi et al.(2012) was a survey to investigate the use of machine learning in cloud environment. Later in 2013, Tantar et al.(2013) conducted a small review in the area of computational intelligence for cloud management. Another study by Guzek et al.(2015) was a survey study with the aim of giving computational intelligence researchers an understanding of novelties in the cloud. In addition, Demirci (2015) reported on recent works focused on the use of machine learning for energy optimization in cloud computing environment. Similarly, Zhan et al.(2015) presented a report examining and comparing various evolutionary computation approaches to some areas of the cloud. A recent study by Shishira et al.(2016) was also a survey reporting on optimization algorithm for cloud environment based on three well known meta-heuristic techniques which are ant colony optimization(ACO), particle swarm optimization (PSO) and genetic algorithm (GA).In another survey study conducted by Pooman et al.(2016), some workflow scheduling metaheuristic algorithm in cloud and grids were discussed. A recent study by Masdari et al.(2016) conducted a survey to present a comparative analysis of literatures on particle swarm optimization schemes that has been proposed to the cloud environment and then provides a classification of the scheme based on the type of particle swarm optimization algorithm that was applied to the scheme. 3. Methodology To ensure that reliable information is provided on the topic of soft computing techniques in cloud environments, the systematic mapping was conducted following through frameworks proposed by Petersen et al.(2008) and Kitchenham and Charters(2007).
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
33 3
3.1. Research questions The main objective of this study is to pinpoint an inclusive systematic mapping outline of recent research and highlight research gaps in the topic of applied soft computing in cloud environment. The general goal of this paper is defined by the research question below: RQ1 Which topics with their focus area were investigated in the soft computing applications in cloud environment? RQ2 Which soft computing techniques are explored at the intersection of the focus areas? RQ3 Which types of research and contribution facets are used? RQ4 What were the publication years and medium/venue of publication? 3.2. Source data and search strategy The database sources as recommended by Petersen et al.(2008) were resourceful for indexing ICT and computer science related literature. These main databases are Science Direct, IEEExplore, ACM digital library and Springerlink. Search keywords used were defined as the initial step in constructing the search string based on the recommendations by Kitchenham and Charters(2007). This was driven mainly with the topic and the research questions that were proposed, a total of 6362 were returned as results after applying the following search string: ((“soft computing” OR “fuzzy logic” OR “machine learning” OR “neural network” OR “support vector machines” OR “evolutionary computation” OR “genetic algorithm” OR “differential evolution” OR “ant colony optimization” OR “particle swarm optimization”) AND (“cloud computing” OR “cloud services” OR “cloud”)). The second step involved filtering the 6362 studies using the inclusions and exclusion criteria to examine the title, a total of 412 papers were selected. The third phase was to filter the studies by reading the abstract and keyword, doing this resulted in selecting 283 studies and finally the last step of selection was the full text review selection, the introduction, methodology and conclusion were examined for relevancy, research questions, inclusion and exclusion criteria were utilized for selecting the primary studies. A total of 163 studies were selected as primary studies. Figure 1 below shows the selection process for the selection of the studies. Phase 1: Application of Search string
3.3. Study selection
Phase 2: Title based selection
Phase 3: Abstract based selection
Phase 4: Full text based selection
Fig. 1. Selection process
In order to identify empirical studies in the area of soft computing techniques in the cloud environment, subsequent guidelines were followed. The main criteria is taking into account the studies only published within the last decade. The following are the inclusion and exclusion criteria that were used to filter the studies in the paper: Inclusion criteria – The study must report on soft computing techniques in cloud environment from computer science or information system perspectives The study must provide empirical data or supporting evidence
Exclusion criteria – Studies related to cloud but no application of soft computing Studies related to soft computing but no relation to cloud system Articles that just mentions our focus Articles that are from other fields of study Articles not in English language, articles in form of books or abstracts
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
34 4
3.4. Search results As stated earlier the initial search results were total of 6362 articles for all sources but after going through the three additional filtering phases, as Table 1 below demonstrates the filtered results yielded to a total of 163 articles in the form of journals, conferences and workshop proceedings. Table 1. Studies after filtration. Source
Phase 1
Phase 2
Phase 3
Phase 4
ScienceDirect
611
36
19
10
IEEE
2367
221
159
96
ACM
456
46
24
10
Springer
1100
109
81
47
Total
6362
412
283
163
4. Classification scheme In other to classify the primary studies of this article, the classification scheme recommended by Petersen et al.(2008) was adopted, the classification scheme was carried out from four different viewpoints: focus area, contribution facet, research facet and additionally included for the purpose of this study type of soft computing techniques. In the focus area classification, the studies were classified into five main categories which are; task optimization, service selection, security, power optimization and cost optimization. The studies were further classified by their contribution facet which was spliced into five (5) clusters; method, model, process, tools and metrics. The third classification scheme is the research type describing the kind of research that was carried out in a study, the research type classification used in this study are; validation research, evaluation research, solution proposal, opinion paper, philosophical papers and experience papers. The final classification scheme is the soft computing technique, which helps to classify the study based on the soft computing technique applied to cloud computing. According to Chaturvedi (2008) there exist four main branches to soft computing which are artificial neural network (ANN), fuzzy logic (FL), evolutionary computation (EC) and hybrid systems. Due to the broadness of EC, it was further classified into particle swarm optimization (PSO), ant colony optimization (ACO), artificial bee colony (ABC), genetic algorithm (GA), differential evolution (DE) and EC approaches that didn’t fall in any known category was classified as other evolutionary computation. 5. Results The answer to the research questions were presented as follows: 5.1. RQ1 What are the studies that investigate soft computing techniques in the cloud and their focus area? There were 163 studies found that investigate soft computing techniques applied to cloud environment as shown in Table 2. These studies were divided into five (5) major research focus areas as; task optimization, power optimization, cost optimization, security and service selection. Most of the studies focus on task scheduling optimization (75%), with a few studies focusing on power optimization (10%), security (7%). Service selection (5%) and cost (3%). According to Table 3, after the year 2011 there is a considerable amount of increase in the number of articles regarding task optimization of the cloud followed by security.
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
35 5
Table 2. Distribution of studies by focus area. Category
Studies(S)
Total
%
Task optimization
S1,S2,S3,S4,S5,S7,S9,S10,S12,S13,S14, S15,S16,S17,S19,S20,S21,S22,S23,S24,S25,
122
75
S26,S31,S32,S33,S35,S36,S38,S39,S40,S42,S44,S45,S46,S47,S48,S49,S50,S52,S54,S56,S57, S59,S60,S62,S63,S65,S66,S67,S68,S69,S70,S71,S72,S73,S75,S76,S79,S82,S85,S88,S89,S91, S92,S93,S95,S96,S97,S98,S100,S103,S104,S105,S106,S107,S108,S109,S110,S111,S112, S113,S114,S115,S117,S118,S119,S121,S122,S123,S124,S125,S126,S127,S129,S130,S131, S132,S133,S134,S135,S136,S137,S138,S139,S140,S141,S144,S145,S147,S149,S150,S151, S152,S153,S154,S157,S158,S159,S160,161,,S163 Power optimization
S27,S28,S29,S30,S51,S64,S74,S77,S78,S83,S101,S116,S120,S128,S143,S148,162
17
10
Security
S6,S18,S37,S43,S55,S81,S87,S90,S94,S142,S146,S156
12
7
Service selection
S11,S41,S53,S58,S80,S84,S86,S99,S102
9
5
Cost optimization
S34,S61,S155
3
3
Table 3. Further distribution of research focus area by year. Focus area
2009
2010
2011
2012
2013
2014
2015
2016
Total
%
Task optimization
1
2
5
14
15
25
28
32
122
75
Cost optimization
-
-
-
1
-
1
1
-
3
3
Power optimization
-
-
5
0
1
3
5
3
17
10
Security
-
-
-
2
-
1
2
7
12
7
Service selection
-
-
-
0
1
2
1
5
9
5
5.2. RQ2 What are thesoft computing techniques explored at the intersection of the focus areas? Five main soft computing techniques were found in the studies. The majority were focused on evolutionary computation and its branches (71%), hybrid systems (14%), fuzzy logic (10%) and artificial neural network (5%). According to findings shown in Table 4, PSO has been the soft computing technique with the highest attention over the years and has had the highest application in task optimization followed by GA, see Fig. 2. Table 4. Distribution of soft computing technique by publication year. Soft computing focus
2009
2010
2011
2012
2013
2014
2015
2016
Total
%
ANN
-
-
2
-
1
1
3
2
9
5
FL
-
-
-
-
2
3
4
7
16
10
GA
1
1
5
7
4
5
6
10
39
23
PSO
-
1
-
7
7
9
12
11
47
31
ACO
-
-
3
-
2
7
3
3
18
11
ABC
-
-
-
-
-
1
2
3
6
4
DE
-
-
-
-
-
1
-
2
3
2
Other EC
-
-
-
-
1
-
-
1
2
1
Hybrid
-
-
-
3
-
5
7
8
23
14
Total
1
2
10
17
17
32
37
47
163
100
ObinnaObinna H. Ejimogu et al./ Procedia Computer Science 00 (2018)120 000–000 H. Ejimogu et al. / Procedia Computer Science (2017) 31–38
636
Fig. 2. Distribution of studies by soft computing techniques to cloud environment and research focus
5.3. RQ3 What were the research types and contribution facets of these studies? As Table 5 and Table 6 show with regards to research type, the study found more frequency on publications of validation research and solution proposals with little or none at all in the other area, while in contribution facets majority of publications were of methods contribution type. Table 5. Distribution of study by contribution. Contribution
Total
Table 6. Distribution of study by research type. %
Research type
Total
%
Method
140
86
Validation
144
88
Process
22
13
Evaluation
1
1
Model
1
1
Philosophical
1
1
Metrics
0
0
Solution proposal
17
10
Tools
0
0
Opinion papers
0
0
Experience papers
0
0
5.4. RQ4What were the publication year and publication medium of the studies? The study investigates soft computing techniques to cloud environment from 2006-2016 however no studies were found from 2006 to 2008, with a first appearance in 2009 and a more prominent growth after 2011, also the studies found were made up of 129 conference proceedings, 30 journal articles and 4 workshop proceedings, Table 7 shows the top 9 publications forums with the highest paper frequency representing only 22% of the total studies while Fig.3 shows the exponential growth of studies focused on soft computing techniques in cloud environment.
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
37 7
Table 7. Top 9 publication medium. Publication venues
No. of papers
International conference on fuzzy systems
4
Congress on evolutionary computation
4
International conference in swarm intelligence
4
International conference on advances in computing, communications and informatics
3
International conference on cloud computing and intelligence systems
3
International conference on cloud networking
3
International conference on computational intelligence and communication networks
3
Procedia technology journal
2
Future generation and computer systems journal
2
50 45 40 35 30 25
47
20 32
15 10 5 0
0
0
0
1
2
2006
2007
2008
2009
2010
17
17
2012
2013
37
10 2011
2014
2015
2016
Fig. 3. Research studies per year
6. Conclusion and future research The core aim of these study was to provide an overview of various soft computing techniques that have been applied to cloud computing over the last ten years from 2006 to 2016 and 163 studies were identified as primary studies. The study shows a recent rise in focus on soft computing techniques in cloud environment from 2012. Majority of the studies focuses on task optimization and also PSO as the most popular technique proposed for cloud optimization followed by GA. This study finds a research gap in the area of ABC and also cost optimization, only three (3) studies were found in relation with soft computing to cost optimization, further work is recommended in this area because in a survey by Rightscale (2017), cost is stated as the fastest growing concern in the cloud and investigating soft computing techniques to optimize cost could help tackle this challenge, also only 1% of the study base their study on real world application while the vast number of the studies were validation, no studies were found in regards to tools and metrics contribution studies of soft computing techniques to cloud environment. In summary, many studies have been conducted in the area of applying soft computing techniques to the cloud so as to make the cloud more autonomous and intelligent. This mapping studies shows a growing increase in the
38 8
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
interest of soft computing techniques to cloud. The academic community should be encourage to conduct more practical studies and the industry to work to see that these proposed techniques are implemented to develop a solid foundation for improving soft computing techniques in the cloud environment. Appendix Due to the nature of this study, a large number of primary studies were included hence with limited space, a complete list of the primary studies could be visited through this link https://goo.gl/jwnXHI References Chaturvedi, D., 2008. Soft Computing Techniques and its Applications in Electrical Engineering.In: Studies in Computational Intelligence. Springer, Berlin, 1-10. Demirci, M., 2015. A Survey of Machine Learning Applications for Energy-Efficient Resource Management in Cloud Computing Environments, 14th International Conference on Machine Learning and Applications. Florida, USA, 1185-1190. Esposito, C., Ficco, M., Palmieri, F., Castiglione, A.,2015. Smart Cloud Storage Service Selection Based on Fuzzy Logic, Theory of Evidence and Game Theory. IEEE Transactions on Computers 65, 8, 2348-2362. Feller, E., Rilling, L., Morin, C., 2011. Energy-Aware Ant Colony Based Workload Placement in Clouds, 12th International Conference on Grid Computing. Lyon, France,23-26. Guzek, M., Bouvry, P., Talbi, G.,2015. A Survey of Evolutionary Computation for Resource Management of Processing in Cloud Computing. IEEE Computational Intelligence Magazine 10, 2, 53-67. Hormozi, E., Hormozi, H., Akbari, K., Javan, S.,2012. A Using of Machine Learning into cloud environment (a survey): Managing and scheduling of resources in cloud systems, 7th International Conference on P2P, Parallel, Grid, Cloud and Internet Computing. BC, Canada,363-368. Kitchenham, B., Charters, S., 2007. Guidelines for performing Systematic Literature Reviews in Software Engineering. Keele University and Durham University Joint Report. Masdari, M., Salehi, F., Jalali, M., Bidaki, M., 2016. A Survey of PSO-Based Scheduling Algorithms in Cloud Computing. Journal of Network and Systems Management 25, 1, 1-37. Petersen, J., Feldt, R., Mujtaba, S., Mattson, M., 2008. Systematic Mapping Studies in Software Engineering, 12th International Conference on Evaluation and Assessment in Software Engineering. Bari, Italy, 68-77. Pooman, C., Maitreyee, D., Naveen, A., 2016. Meta-Heuristic Based Approach for Workflow Scheduling in Cloud Computing: A Survey, International Conference on Artificial Intelligence and Evolutionary Computations in Engineering Systems. Chennai, India, 1331-1345. Pueschel, T., Ludwigs, A., Freiburg, U., Neumann, D., 2009. Management of Cloud Infrastructures: Policy- Based Revenue Optimization, International Conference on Information System.Arizona, USA, 178. Ramezani, F., Lu, J., Hussain, K., 2014. Task-Based System Load Balancing in Cloud Computing Using Particle Swarm Optimization. Journal of Parallel Programming 42, 5, 739-754. Rightscale, 2017. Retrieved from http://assets.rightscale.co m/uploads/pdfs/RightScale-2017-State-of-the-Cloud-Report.pdf. Shishira, S., Kandasamy, A., Chandrasekaran, K., 2016. Survey on Meta Heuristic Optimization Techniques in Cloud Computing, 5th International Conference on Advances in Computing, Communications and Informatics. Jaipur, India, 1434-1440. Tantar, A., Bouvry, P., Dorronsoro, B., Talbi, G., 2013. Computational Intelligence for Cloud Managementcurrent trends and opportunities, IEEE Congress on Evolutionary Computations. Cancun, Mexico, 1286-1293. Zhan, H., Liu, F., Gong, J., Zhang, J., Chung, S., Li, Y., 2015. Cloud Computing Resource Scheduling and a Survey of its Evolutionary Approaches. ACM Computing Surveys 47, 4, 1-33.
ScienceDirect
Available online at www.sciencedirect.com Procedia Computer Science 00 (2018) 000–000
ScienceDirect
www.elsevier.com/locate/procedia
Procedia Computer Science 120 (2017) 31–38
9th 9thInternational InternationalConference Conferenceon onTheory Theoryand andApplication Applicationof ofSoft SoftComputing, Computing,Computing Computingwith with Words and Perception, ICSCCW 2017, 24-25 22-23 August August 2017, 2017, Budapest, Budapest, Hungary Hungary
A systematic mapping study on soft computing techniques to cloud environment Obinna H. Ejimogua, Seren Başarana* a
Department of Computer Information Systems, Near East University,POBOX:99138,Nicosia, North Cyprus, Mersin 10, Turkey.
Abstract Cloud computing plays an essential role in storage and transfer of big capacity data due to a rapid increase in size and the number of organizational activities. There exist numerous studies in which diverse soft computing techniques are applied to the cloud environment. The relevant extant literature that were clustered into five main categories with respect to precedence are; task optimization, power optimization, security, service selection and cost optimization. Yet, it was discovered that there is a dearth of systematic review/mapping studies particularly on soft computing techniques in cloud environment so as to obtain exclusive insight, to identify existing gaps and future research directions. Therefore the aim of this paper is to conduct a systematic mapping study of recent literature on soft computing techniques in cloud environment. For this purpose, 163 articles were chosen as primary sources that were published within the last decade, which were classified based on study focus area, type of research, contribution facet and particularly the type of soft computing technique used. Findings revealed that task optimization takes part as the highly preferred research focus area. Secondly, most of the articles found are of validation studies. The contributions of most of the studies are concerned about methods and finally the top three soft computing techniques were detected as particle swarm optimization (PSO), genetic algorithm (GA) and hybrid systems. The results of this study confirm that applying soft computing techniques in cloud computing has gained more and more significant attention recently but there still remain challenges and gaps which calls for further investigation especially in the area of cost optimization and also artificial bee colony. © 2018 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. Keywords:Cloud computing; power optimiztion; soft computing; systematic mapping; task optimization.
* Seren Başaran. Tel.: +90- 392-675-1000(3121); fax: +90-392-6751051. E-mail address: [email protected] 1877-0509© 2018 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. 1877-0509 © 2018 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 9th International Conference on Theory and application of Soft Computing, Computing with Words and Perception. 10.1016/j.procs.2017.11.207
32 2
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
1. Introduction Recently cloud computing environment has drawn remarkable attention from organizations because of its capabilities to reduce cost, maximize productivity with high degree of flexibility in handling big data (Pueschel et al., 2009). Considering the burden of manual management in ever growing capacity of cloud data, there has always been a need to seek for more efficient autonomous and computational intelligent solutions to handle the dynamic nature of the cloud. Over the last decades, numerous soft computing techniques have been proposed and utilized for automated optimization of the cloud environment. The prevalent types of soft computing techniques which addressed mainly within the cloud computing related studies about security challenges, task optimization, service selection cost optimization and power optimization include; fuzzy logic, neural networks, genetic algorithm, swarm intelligence and hybrid systems (Esposito et al., 2015; Feller et al., 2011; Ramezani et al., 2014). Although soft computing techniques have caught the interests of many researchers over the years, yet to author’s knowledge, there is currently no systematic mapping study conducted particularly to soft computing in cloud environment. A systematic mapping study not only helps to identify the extent of study in a research area, but also provides a view to pinpoint existing gaps in a particular subject area and summarizes the entire corpus of study in a subject area to encourage further development. In the light of above, a systematic mapping study was conducted to gather, analyze and interpret existing studies that were carried out in the area of soft computing in cloud environment so as to provide an overview and highlight research gaps in the subject area. 2. Motivation and related surveys Although soft computing techniques are popular and have been widely applied to cloud environments, initial attempts to locate relevant systematic mapping studies in the area of soft computing techniques in cloud environment proved to be abortive, which became the fundamental motivation to carry out this systematic study. The only located literature survey studies were briefly summarized as follows: A study conducted by Hormozi et al.(2012) was a survey to investigate the use of machine learning in cloud environment. Later in 2013, Tantar et al.(2013) conducted a small review in the area of computational intelligence for cloud management. Another study by Guzek et al.(2015) was a survey study with the aim of giving computational intelligence researchers an understanding of novelties in the cloud. In addition, Demirci (2015) reported on recent works focused on the use of machine learning for energy optimization in cloud computing environment. Similarly, Zhan et al.(2015) presented a report examining and comparing various evolutionary computation approaches to some areas of the cloud. A recent study by Shishira et al.(2016) was also a survey reporting on optimization algorithm for cloud environment based on three well known meta-heuristic techniques which are ant colony optimization(ACO), particle swarm optimization (PSO) and genetic algorithm (GA).In another survey study conducted by Pooman et al.(2016), some workflow scheduling metaheuristic algorithm in cloud and grids were discussed. A recent study by Masdari et al.(2016) conducted a survey to present a comparative analysis of literatures on particle swarm optimization schemes that has been proposed to the cloud environment and then provides a classification of the scheme based on the type of particle swarm optimization algorithm that was applied to the scheme. 3. Methodology To ensure that reliable information is provided on the topic of soft computing techniques in cloud environments, the systematic mapping was conducted following through frameworks proposed by Petersen et al.(2008) and Kitchenham and Charters(2007).
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
33 3
3.1. Research questions The main objective of this study is to pinpoint an inclusive systematic mapping outline of recent research and highlight research gaps in the topic of applied soft computing in cloud environment. The general goal of this paper is defined by the research question below: RQ1 Which topics with their focus area were investigated in the soft computing applications in cloud environment? RQ2 Which soft computing techniques are explored at the intersection of the focus areas? RQ3 Which types of research and contribution facets are used? RQ4 What were the publication years and medium/venue of publication? 3.2. Source data and search strategy The database sources as recommended by Petersen et al.(2008) were resourceful for indexing ICT and computer science related literature. These main databases are Science Direct, IEEExplore, ACM digital library and Springerlink. Search keywords used were defined as the initial step in constructing the search string based on the recommendations by Kitchenham and Charters(2007). This was driven mainly with the topic and the research questions that were proposed, a total of 6362 were returned as results after applying the following search string: ((“soft computing” OR “fuzzy logic” OR “machine learning” OR “neural network” OR “support vector machines” OR “evolutionary computation” OR “genetic algorithm” OR “differential evolution” OR “ant colony optimization” OR “particle swarm optimization”) AND (“cloud computing” OR “cloud services” OR “cloud”)). The second step involved filtering the 6362 studies using the inclusions and exclusion criteria to examine the title, a total of 412 papers were selected. The third phase was to filter the studies by reading the abstract and keyword, doing this resulted in selecting 283 studies and finally the last step of selection was the full text review selection, the introduction, methodology and conclusion were examined for relevancy, research questions, inclusion and exclusion criteria were utilized for selecting the primary studies. A total of 163 studies were selected as primary studies. Figure 1 below shows the selection process for the selection of the studies. Phase 1: Application of Search string
3.3. Study selection
Phase 2: Title based selection
Phase 3: Abstract based selection
Phase 4: Full text based selection
Fig. 1. Selection process
In order to identify empirical studies in the area of soft computing techniques in the cloud environment, subsequent guidelines were followed. The main criteria is taking into account the studies only published within the last decade. The following are the inclusion and exclusion criteria that were used to filter the studies in the paper: Inclusion criteria – The study must report on soft computing techniques in cloud environment from computer science or information system perspectives The study must provide empirical data or supporting evidence
Exclusion criteria – Studies related to cloud but no application of soft computing Studies related to soft computing but no relation to cloud system Articles that just mentions our focus Articles that are from other fields of study Articles not in English language, articles in form of books or abstracts
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
34 4
3.4. Search results As stated earlier the initial search results were total of 6362 articles for all sources but after going through the three additional filtering phases, as Table 1 below demonstrates the filtered results yielded to a total of 163 articles in the form of journals, conferences and workshop proceedings. Table 1. Studies after filtration. Source
Phase 1
Phase 2
Phase 3
Phase 4
ScienceDirect
611
36
19
10
IEEE
2367
221
159
96
ACM
456
46
24
10
Springer
1100
109
81
47
Total
6362
412
283
163
4. Classification scheme In other to classify the primary studies of this article, the classification scheme recommended by Petersen et al.(2008) was adopted, the classification scheme was carried out from four different viewpoints: focus area, contribution facet, research facet and additionally included for the purpose of this study type of soft computing techniques. In the focus area classification, the studies were classified into five main categories which are; task optimization, service selection, security, power optimization and cost optimization. The studies were further classified by their contribution facet which was spliced into five (5) clusters; method, model, process, tools and metrics. The third classification scheme is the research type describing the kind of research that was carried out in a study, the research type classification used in this study are; validation research, evaluation research, solution proposal, opinion paper, philosophical papers and experience papers. The final classification scheme is the soft computing technique, which helps to classify the study based on the soft computing technique applied to cloud computing. According to Chaturvedi (2008) there exist four main branches to soft computing which are artificial neural network (ANN), fuzzy logic (FL), evolutionary computation (EC) and hybrid systems. Due to the broadness of EC, it was further classified into particle swarm optimization (PSO), ant colony optimization (ACO), artificial bee colony (ABC), genetic algorithm (GA), differential evolution (DE) and EC approaches that didn’t fall in any known category was classified as other evolutionary computation. 5. Results The answer to the research questions were presented as follows: 5.1. RQ1 What are the studies that investigate soft computing techniques in the cloud and their focus area? There were 163 studies found that investigate soft computing techniques applied to cloud environment as shown in Table 2. These studies were divided into five (5) major research focus areas as; task optimization, power optimization, cost optimization, security and service selection. Most of the studies focus on task scheduling optimization (75%), with a few studies focusing on power optimization (10%), security (7%). Service selection (5%) and cost (3%). According to Table 3, after the year 2011 there is a considerable amount of increase in the number of articles regarding task optimization of the cloud followed by security.
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
35 5
Table 2. Distribution of studies by focus area. Category
Studies(S)
Total
%
Task optimization
S1,S2,S3,S4,S5,S7,S9,S10,S12,S13,S14, S15,S16,S17,S19,S20,S21,S22,S23,S24,S25,
122
75
S26,S31,S32,S33,S35,S36,S38,S39,S40,S42,S44,S45,S46,S47,S48,S49,S50,S52,S54,S56,S57, S59,S60,S62,S63,S65,S66,S67,S68,S69,S70,S71,S72,S73,S75,S76,S79,S82,S85,S88,S89,S91, S92,S93,S95,S96,S97,S98,S100,S103,S104,S105,S106,S107,S108,S109,S110,S111,S112, S113,S114,S115,S117,S118,S119,S121,S122,S123,S124,S125,S126,S127,S129,S130,S131, S132,S133,S134,S135,S136,S137,S138,S139,S140,S141,S144,S145,S147,S149,S150,S151, S152,S153,S154,S157,S158,S159,S160,161,,S163 Power optimization
S27,S28,S29,S30,S51,S64,S74,S77,S78,S83,S101,S116,S120,S128,S143,S148,162
17
10
Security
S6,S18,S37,S43,S55,S81,S87,S90,S94,S142,S146,S156
12
7
Service selection
S11,S41,S53,S58,S80,S84,S86,S99,S102
9
5
Cost optimization
S34,S61,S155
3
3
Table 3. Further distribution of research focus area by year. Focus area
2009
2010
2011
2012
2013
2014
2015
2016
Total
%
Task optimization
1
2
5
14
15
25
28
32
122
75
Cost optimization
-
-
-
1
-
1
1
-
3
3
Power optimization
-
-
5
0
1
3
5
3
17
10
Security
-
-
-
2
-
1
2
7
12
7
Service selection
-
-
-
0
1
2
1
5
9
5
5.2. RQ2 What are thesoft computing techniques explored at the intersection of the focus areas? Five main soft computing techniques were found in the studies. The majority were focused on evolutionary computation and its branches (71%), hybrid systems (14%), fuzzy logic (10%) and artificial neural network (5%). According to findings shown in Table 4, PSO has been the soft computing technique with the highest attention over the years and has had the highest application in task optimization followed by GA, see Fig. 2. Table 4. Distribution of soft computing technique by publication year. Soft computing focus
2009
2010
2011
2012
2013
2014
2015
2016
Total
%
ANN
-
-
2
-
1
1
3
2
9
5
FL
-
-
-
-
2
3
4
7
16
10
GA
1
1
5
7
4
5
6
10
39
23
PSO
-
1
-
7
7
9
12
11
47
31
ACO
-
-
3
-
2
7
3
3
18
11
ABC
-
-
-
-
-
1
2
3
6
4
DE
-
-
-
-
-
1
-
2
3
2
Other EC
-
-
-
-
1
-
-
1
2
1
Hybrid
-
-
-
3
-
5
7
8
23
14
Total
1
2
10
17
17
32
37
47
163
100
ObinnaObinna H. Ejimogu et al./ Procedia Computer Science 00 (2018)120 000–000 H. Ejimogu et al. / Procedia Computer Science (2017) 31–38
636
Fig. 2. Distribution of studies by soft computing techniques to cloud environment and research focus
5.3. RQ3 What were the research types and contribution facets of these studies? As Table 5 and Table 6 show with regards to research type, the study found more frequency on publications of validation research and solution proposals with little or none at all in the other area, while in contribution facets majority of publications were of methods contribution type. Table 5. Distribution of study by contribution. Contribution
Total
Table 6. Distribution of study by research type. %
Research type
Total
%
Method
140
86
Validation
144
88
Process
22
13
Evaluation
1
1
Model
1
1
Philosophical
1
1
Metrics
0
0
Solution proposal
17
10
Tools
0
0
Opinion papers
0
0
Experience papers
0
0
5.4. RQ4What were the publication year and publication medium of the studies? The study investigates soft computing techniques to cloud environment from 2006-2016 however no studies were found from 2006 to 2008, with a first appearance in 2009 and a more prominent growth after 2011, also the studies found were made up of 129 conference proceedings, 30 journal articles and 4 workshop proceedings, Table 7 shows the top 9 publications forums with the highest paper frequency representing only 22% of the total studies while Fig.3 shows the exponential growth of studies focused on soft computing techniques in cloud environment.
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al. / Procedia Computer Science 00 (2018) 000–000
37 7
Table 7. Top 9 publication medium. Publication venues
No. of papers
International conference on fuzzy systems
4
Congress on evolutionary computation
4
International conference in swarm intelligence
4
International conference on advances in computing, communications and informatics
3
International conference on cloud computing and intelligence systems
3
International conference on cloud networking
3
International conference on computational intelligence and communication networks
3
Procedia technology journal
2
Future generation and computer systems journal
2
50 45 40 35 30 25
47
20 32
15 10 5 0
0
0
0
1
2
2006
2007
2008
2009
2010
17
17
2012
2013
37
10 2011
2014
2015
2016
Fig. 3. Research studies per year
6. Conclusion and future research The core aim of these study was to provide an overview of various soft computing techniques that have been applied to cloud computing over the last ten years from 2006 to 2016 and 163 studies were identified as primary studies. The study shows a recent rise in focus on soft computing techniques in cloud environment from 2012. Majority of the studies focuses on task optimization and also PSO as the most popular technique proposed for cloud optimization followed by GA. This study finds a research gap in the area of ABC and also cost optimization, only three (3) studies were found in relation with soft computing to cost optimization, further work is recommended in this area because in a survey by Rightscale (2017), cost is stated as the fastest growing concern in the cloud and investigating soft computing techniques to optimize cost could help tackle this challenge, also only 1% of the study base their study on real world application while the vast number of the studies were validation, no studies were found in regards to tools and metrics contribution studies of soft computing techniques to cloud environment. In summary, many studies have been conducted in the area of applying soft computing techniques to the cloud so as to make the cloud more autonomous and intelligent. This mapping studies shows a growing increase in the
38 8
Obinna H. Ejimogu et al. / Procedia Computer Science 120 (2017) 31–38 Obinna H. Ejimogu et al./ Procedia Computer Science 00 (2018) 000–000
interest of soft computing techniques to cloud. The academic community should be encourage to conduct more practical studies and the industry to work to see that these proposed techniques are implemented to develop a solid foundation for improving soft computing techniques in the cloud environment. Appendix Due to the nature of this study, a large number of primary studies were included hence with limited space, a complete list of the primary studies could be visited through this link https://goo.gl/jwnXHI References Chaturvedi, D., 2008. Soft Computing Techniques and its Applications in Electrical Engineering.In: Studies in Computational Intelligence. Springer, Berlin, 1-10. Demirci, M., 2015. A Survey of Machine Learning Applications for Energy-Efficient Resource Management in Cloud Computing Environments, 14th International Conference on Machine Learning and Applications. Florida, USA, 1185-1190. Esposito, C., Ficco, M., Palmieri, F., Castiglione, A.,2015. Smart Cloud Storage Service Selection Based on Fuzzy Logic, Theory of Evidence and Game Theory. IEEE Transactions on Computers 65, 8, 2348-2362. Feller, E., Rilling, L., Morin, C., 2011. Energy-Aware Ant Colony Based Workload Placement in Clouds, 12th International Conference on Grid Computing. Lyon, France,23-26. Guzek, M., Bouvry, P., Talbi, G.,2015. A Survey of Evolutionary Computation for Resource Management of Processing in Cloud Computing. IEEE Computational Intelligence Magazine 10, 2, 53-67. Hormozi, E., Hormozi, H., Akbari, K., Javan, S.,2012. A Using of Machine Learning into cloud environment (a survey): Managing and scheduling of resources in cloud systems, 7th International Conference on P2P, Parallel, Grid, Cloud and Internet Computing. BC, Canada,363-368. Kitchenham, B., Charters, S., 2007. Guidelines for performing Systematic Literature Reviews in Software Engineering. Keele University and Durham University Joint Report. Masdari, M., Salehi, F., Jalali, M., Bidaki, M., 2016. A Survey of PSO-Based Scheduling Algorithms in Cloud Computing. Journal of Network and Systems Management 25, 1, 1-37. Petersen, J., Feldt, R., Mujtaba, S., Mattson, M., 2008. Systematic Mapping Studies in Software Engineering, 12th International Conference on Evaluation and Assessment in Software Engineering. Bari, Italy, 68-77. Pooman, C., Maitreyee, D., Naveen, A., 2016. Meta-Heuristic Based Approach for Workflow Scheduling in Cloud Computing: A Survey, International Conference on Artificial Intelligence and Evolutionary Computations in Engineering Systems. Chennai, India, 1331-1345. Pueschel, T., Ludwigs, A., Freiburg, U., Neumann, D., 2009. Management of Cloud Infrastructures: Policy- Based Revenue Optimization, International Conference on Information System.Arizona, USA, 178. Ramezani, F., Lu, J., Hussain, K., 2014. Task-Based System Load Balancing in Cloud Computing Using Particle Swarm Optimization. Journal of Parallel Programming 42, 5, 739-754. Rightscale, 2017. Retrieved from http://assets.rightscale.co m/uploads/pdfs/RightScale-2017-State-of-the-Cloud-Report.pdf. Shishira, S., Kandasamy, A., Chandrasekaran, K., 2016. Survey on Meta Heuristic Optimization Techniques in Cloud Computing, 5th International Conference on Advances in Computing, Communications and Informatics. Jaipur, India, 1434-1440. Tantar, A., Bouvry, P., Dorronsoro, B., Talbi, G., 2013. Computational Intelligence for Cloud Managementcurrent trends and opportunities, IEEE Congress on Evolutionary Computations. Cancun, Mexico, 1286-1293. Zhan, H., Liu, F., Gong, J., Zhang, J., Chung, S., Li, Y., 2015. Cloud Computing Resource Scheduling and a Survey of its Evolutionary Approaches. ACM Computing Surveys 47, 4, 1-33.