Perspectives in Science (2016) 8, 689—691 CORE Metadata, citation and similar papers at core.ac.uk Available online at www.sciencedirect.com Provided by Elsevier - Publisher Connector ScienceDirect jo urnal homepage: www.elsevier.com/pisc A review on software testing approaches ଝ for cloud applications ∗ Tamanna Siddiqui, Riaz Ahmad Department of Computer Science, Aligarh Muslim University, India Received 20 February 2016; received in revised form 9 June 2016; accepted 9 June 2016 Available online 4 July 2016 KEYWORDS Summary Cloud computing has actually been invented to be the latest computing standard that will work several distinctive research areas, such as software testing. Testing cloud applica- Cloud testing; Clustering; tions will keep its unique characteristics that involve more recent testing techniques. Software Prioritization; testing helps to reduce the need for hardware and software services and also provide adaptable Privacy; and valuable cloud platform. Testing within the cloud platform is easily manageable based on Security; new test models and criteria. Prioritization approach is made responsive to build much better relationship between test cases. These test cases are clustered dependent on priority level. Test cases The resources can be used properly by applying load balancing algorithm. Cloud guarantees maximum usage of existing resources. But, security defined as a primary problem in cloud. At the present time, organizations are progressively moving excited about deploying and making use of ready-prepared business applications, with particular short-term to the marketplace. The possible lack of capital budgets for software planning and on principle deployments, along with the swift progression of cloud these are the reasons why one should make the interest on business application. However, these are the interests that help make the SaaS based busi- ness application on-demand. In this paper different approaches has been discussed that will help to extend the cloud environment. Also, the study of several well-known software testing approaches. © 2016 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction making use of virtual machines (VM) to do testing is not a unique idea. The issue starts when those VM are unavailable as needed. Some issues that faced like servers, data source Software testing is used to find out the functionality and as well as applications not representing virtualized environ- validity of a program over performance and analysis. By ments. To address this challenge migrate the tests way to the cloud. Cloud computing provides you with other norms ଝ of testing agility which gives a paradigm shift that a pay- This article belongs to the special issue on Engineering and Mate- per-use technique but not creating in-house data canters. rial Sciences. ∗ Corresponding author. Tel.: +91 9412273378. The users have the ability improve their computing choices E-mail address: [email protected] (R. Ahmad). to fulfill their requirements on-demand (Prathibhan et al., http://dx.doi.org/10.1016/j.pisc.2016.06.060 2213-0209/© 2016 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/). 690 T. Siddiqui, R. Ahmad Figure 1 Cloud computing testing techniques. 2014; Doss et al., 2012). It seen to apply virtualized hard- Clustering is a bottom-up clustering method each observa- ware appropriately unrestricted storage also software vital tion starts in its own cluster (Liu et al., 2013). It utilized for thin-client network right to use the infrastructure. By in several field because of it offers ability to use arbitrary putting on prioritization strategies simply to make the most clustering dissimilarity or distance function. Clusters are of the scope and plan the test cases in a sequential order to design by a couple of objects determined by the squared- get the code coverage at the quick rate. Clustering models error objective functions because highly responsive to initial are useful to software providers to find his services to the conditions when various size and non-globular state. Algo- cloud users and helps the service expert to extend use of rithms design by using the part of the data in Partition-based software service toward the cloud nature. Clustering. Once every cluster expands clustering elements and minimizes the sum of squared distance from the mean within every cluster. Moreover it used to minimize the clus- Testing in cloud computing tering criteria by iteratively moving the data between the clusters till an ideal partition is seen. The testing framework in cloud computing consists of four most important techniques (Fig. 1). Load distribution scenarios Prioritization techniques Load balancing aims to improve resource use, extend throughput, eliminate response time and avoid overload Prioritization techniques are used to improve the function of any single resource. Load balancing technique will and give a direction to develop and run test cases. Test involve different levels of the cloud in load balancing deci- cases priority assign by using graph coverage value. It is usu- sion. It mainly found in master slave form (Penmatsa and ally defined as the amount of the complexities that involved Chronopoulos, 2011). The static load balancing algorithms dependent test cases which is determined by total number helpful to defines the job to the node according to the of dependents and greatest path of overall dependents of potential of the node to process new queries. The process is the test case. Weight factors to be used for prioritization calculated on previously ability to use the nodes properties, strategies that are customer priority, requirement complex- working power, space for storing capacity and functions. In ness and volatility (Anitha and Srinath, 2014). High factor dynamic includes several features of nodes functions and values show need for prioritization of test case relating to network capacity. It is also combos of prior gathered details the requirement. Test coverage analysis is simply an amount about the nodes in the cloud. According to the attribute applied in software testing named code coverage analysis gathered, Algorithm assigns and reassigns the task to the (Indumathi and Selvamani, 2015). It deals only source codes nodes dynamically depending on the prior gathered details of program that used within the testing process. The cov- about the nodes in cloud area. In Centralized Load Balanc- erage prioritization practice prioritizes the test cases in ing (CLB) technique, the assigning and arranging actions are decreasing order and determined by the sum of ReqSlice produced by one node that chargeable for keeping the infor- (number of requirements declared by test case) and ReqEx- mation based upon entire cloud set-up. It is usually highly ercise (number of requirements utilized by test case). The overloaded and low fault tolerant (Berander and Andrews, exact time is checked for measuring the expense for every 2005). test case because test case prioritization is performed based on cost examination in cost based prioritization (Mohanty Security mechanism et al., 2011). Diffie—Hellman algorithm offered a method to sharing keys Clustering techniques between sender and receiver for encryption and decryption of information and data. It arranges most group members in Clustering can be considered the key unsupervised learn- a logic ring to sharing DH public keys Salzbrunn et al. (2008); ing situation. So, the other issues it focuses on finding a Liao and Chao (2008). The arbitrary parameters produce new structure in an amount of unlabelled data. Agglomerative shared keys for every message that is exchanged between Software testing approaches for cloud applications 691 Table 1 Software testing techniques in cloud platform. Techniques Sub techniques Quality measurement Test case prioritization techniques DSPT IDT, MFD, CS, RT RBPT RC, CC, TCL CBPT Weight in each row using UWD & WD CEBPT IFT, TST, Whole test case cost Clustering algorithm AC CT, half distance matrix computation K-MC Dataset sizes, overhead at the server side PBC Hedgehog representation of the flow Load balancing scenarios HLB Execution time and total performance SLB Expected response time, fairness index, system utilization DLB Power utilization expense, average response time, Usage rates CLB Capability matrix, load matrix Security mechanism DHA I.F., decryption, computation of point on the curve DESA Profitable rate under toleration distance AES Timing simulation each other. Wang’s approach relates to the security and References safety problems in cloud. It offers data access process that is depending on owner—create user read case. DES algorithm Anitha, M.V., Srinath, 2014. A review on software testing framework needs a 56-bit key in execution and one bit used for odd in cloud computing. Int. J. Comput. Sci. Inf. Technol. 5. parity on every octet. AES plays the role of a substitution- Berander, P. , Andrews, A., 2005. Requirements prioritization. In: permutation network according to a design principle. The Engineering and Managing Software Requirements. Springer, pp. finalized result of cipher text is received by several repeti- 69—94. Doss, R., Kumar, R., Prakash, V. , Krishnan, S.G., 2012. Software as tive transformations known as AES cipher. But it takes extra a service (SaaS) testing challenges — an in-depth analysis. Int. round of communication