Maintaining the Data Integrity and Data Replication in Cloud using Modified Genetic Algorithm (Mga) and Greedy Search Algorithm (Gsa)
M.Ramanan* and J.Arockia Stephen Raj
Department of Physical Sciences and IT,Agricultural Engineering College and Research Institute, Tamil Nadu Agricultural University, Coimbatore-641003, Tamil Nadu, India
Article Publishing History Article Received on : 28 Oct 2020 Article Accepted on : 24 Dec 2020 Article Published : 30 Dec 2020 Plagiarism Check: Yes Reviewed by: Dr. Tanveer Ahmad Tarray
Cloud computing is an emerging paradigm thatprovides computing, communication and storage resourcesas a service over a network. Data Integrity is an important aspect which ensures the quality of the data stored in the cloud. Data Replication is the key factor which restores the corrupted data stor
KEYWORDS:
Cloud Computing; Data Integrity; Data Replication; Greedy Search Algorithm (Gsa); Modified Genetic Algorithm (Mga)
Copy the following to cite this article:
Ramanan M, Raj J. A. S .Maintaining the Data Integrity and Data Replication in Cloud using Modified Genetic Algorithm (Mga) and Greedy Search Algorithm (Gsa). Orient.J. Comp. Sci. and Technol; 13(2,3).
Copy the following to cite this URL:
Ramanan M, Raj J. A. S .Maintaining the Data Integrity and Data Replication in Cloud using Modified Genetic Algorithm (Mga) and Greedy Search Algorithm (Gsa). Orient.J. Comp. Sci. and Technol; 13(2,3). Available from: https://bit.ly/39DueD1
Introduction
Cloud Computing provides a platform
for delivery of computing serviceslike storage, networking, analytics etc.,
over the internet [1]. Such advancement incomputing technologies has made major
IT organizations start to move theirdata towards cloud environment. The Cloud
platform offers a lightweightflexible way of storing and accessing data with
innovative newer technologiesand resources. This will facilitate to reduce the
operational cost and to executethe computational needs in a high end
infrastructure effectively.The major challenges encountered by the cloud
storage environment arethe problem of maintaining the integrity of data and
effectively creating andmanaging the replicas of data.
The integrity of data is considered to be themaintenance of the intactness of data [2]. During various operations like accessingthe data, data retrieval, data replication, data storage etc., the integrity of datashould be maintained. The data can be accessed, modified or updated only if asuitable operation is authorized. Data integrity may be hampered at the storagelevel due to various factors. The media types that can result in a data corruptionwill be the bit rot. Bit rot results in data duplications, metadata corruption andcontroller failures. This is a very critical factor which ruins the integrity byaltering the bits of data for any reason. For example, the text file integrity mayget affected by including a single space character within the file. In these cases,making an alteration to some words can render the situation risky.
Cloud providers offersa reliable
environment to the users for storing and accessing the data in the cloud, but
integrity of the data is hard to maintain due to human errors and software
failures. Many research works are being carried out to maintain data integrity.
These mechanisms uses signature based authentication mechanisms using a Third
Party Actuary (TPA). This TPA will be able to check the data integrity by
accessing only a part of the data [5].
The study of data storage and
replication in cloud environmentprovides an unrivalled opportunity to know
about various storage methodsinvolved in cloud. The main objective of a good
cloud storage is to identify anyabnormal activity and resolve the problems
accordingly. There are severalexisting solutions for cloud security framework
and many approach forvulnerabilities coverage and cost optimization, which
mitigates identified set ofvulnerabilities using selected set of techniques by
minimizing cost andmaximizing coverage[11]. However, there are still number of
key challenges yet tobe solved in selecting the appropriate methodologies for
effective cloud storageand replication
In this paper, the MGA and GSA
approaches are based on deterministic approach forfinding near optimal solution
is discussed. Both MGA and GSA methods reduce the lagby storing the data nearer
to the services.The MGA method chooses a set of evenly distributed data blocks,
applies a fitnessfunction and selects the best nodes for replication. This
process will continue untila best solution is obtained. The GSA approach starts
with an inceptive solution for a problem and fixes the problem in a step by
step manner so as to obtain optimal solution in a least time [12].
The cloud environment provides high
computation power to its users by abstracting the computations involved in
background. The abstraction can be implemented as private cloud or public cloud.
There are several factorsthat require processing and management of large data
with state of art computational standards. Lots of research work is being
carried out to improve the computational capabilities (Rashmi Ranjana et al.,
2015).
The approachesdiscussed in this paper can be used to reduce the lag in accessing the remote data. By replicating data in various data center, the performance can be increased. The accesses through local databases pose challenges with respect to synchronization of data and many replicas storage. In this work, an optimized data replication algorithm is proposed.
Related Works
A formal evaluation of the probable types of fine grained updates of data has been provided by Liu et al ., (2014). A scheme has been suggested that can completely take care of the verified auditing as well as intricate update requests. An enhancement has also been suggested on the basis of the approach; this scheme is known to spontaneously decrease the burdens of communication whenever trivial updates are being verified. It has been proven via theoretical and practical verifications that the scheme can increase both the security as well as the flexibility. It can also sufficiently decrease the overheads for big data schemes involving several small updates like the social media applications as well as business transactions.
For removing the complexity of the certificate management in conventional heuristics for cloud data integrity verification, an Identity-based Cloud Data Integrity Checking (ID-CDIC) protocol has been suggested by Yu et al., (2016). Different sized file blocks as well as public auditing can be supported by the proposed concrete construction from Ron Rivest, the Adi Shamir and the Leonard Adleman (RSA) signature. Additionally, a formal security model for ID CDIC can be provided. This can verify the formulation’s security. This is under RSA assumption with large public exponents in the random oracle model. The model of the heuristic has been developed for showing the performance of this scheme. It has been shown via implementation outcomes that in real life situations, the suggested ID-CDIC has been extremely practical as well as adaptable.
A case has been taken up by Gaetani
et al., (2017). This is from the European SUNFISH project. It involves the
formulation of cloud federation platform for public sector which is a secure
design. This comprises the data integrity requirements of the cloud computing
set ups. For adopting the databases based on block chain, there is a need to
address the research questions. Firstly, the persisting research questions have
been detailed. Then, for addressing these questions, the intrinsic challenges
have been addressed. This is followed by outlining the basic formulation of the
database based on block chain in cloud computing paradigm.
The replication of data in cloud
computing data centers has been explored by Boru et al., (2015). This approach
is both effective in terms of energy as well as the system bandwidth consumed,
unlike the other schemes. Because of decreased communication delays, there is
an improvisation in the Quality of Service that has been obtained. The outcomes
of analysis from the pervasive simulations as well as the mathematical
prototype have proven that there is a tradeoff between the performance as well
as energy efficiency. It also guides the formulation of future solutions for
replication of data.
A suggestion has been made for Optimal Performance and Security (DROPS) by Ali et al., (2015). This is for splitting and replicating data. This has a collective approach to tackle both the security and the performance issues. The DROPS scheme splits the file into various parts and this data which is split is replicated over the cloud nodes. Since only a tiny part of the data file is contained in every node, it ensures that even when a successful attack takes place, the data exposed to the attacker will have no critical information. Additionally, the nodes that have the fragmented data are all at a distance of each other using a graph T-coloring so that their placement remains elusive to the attacker. The systems are also relieved from computationally expensive schemes as the DROP technique does not depend on the conventional basis of data as well as security.
Zhang et al., (2016) made a
proposal of Provable Multiple Replication Data Possession protocol having full
dynamics called the MR-DPDP. A genuine structure of the data referred to as the
Merkle hash tree is used in MR-DPDP. This comprises grading which can support
the spontaneous updates on data, along with the data verification as well as
protection. The file blocks of different sizes are supported by the formulation
with the RSA signature. This uses the proof of security and also evaluates the
performance for showing the MR-DPDP that lesser communication overheads are
encountered when the data blocks are being updated and the proof of integrity
for several replicas is being verified.
There are severalexisting solutions for cloud security framework and many approach forvulnerabilities coverage and cost optimization, which mitigates identified set ofvulnerabilities using selected set of techniques by minimizing cost andmaximizing coverage. However, there are still number of key challenges yet tobe solved in selecting the appropriate methodologies for effective cloud storageand replication.
Methodology
In this section, an improved methodology of maintaining data integrity and data replication using the Modified Genetic Algorithm (MGA) and Greedy Search algorithm are discussed.
Modified Genetic Algorithm (MGA)
The problem of data replication is: let F={ f1 —-f 2…...f i} is a file group belonging to a data hub.
is the file blocks of data hub, and
sub blocks belonging to ith data file fi ,which is distributed evenly. A five row relation which is defined as ,
where
are the sub section identification, request count, size of the section, replica numbers and the last access time respectively (Hussein & Mousa 2012).
When a block bk is requested by user uj for a node dndi with a bandwidth
The total band width used required to support n users should be less than dbwi, as shown below
Where si is the concurrent network sessions peak value for data node dndi , bsk is the section size of section bk, dsti is the average service time of data node dndi , dbwi is the network bandwidth of data node . The section availability of a section bk is denoted as BAk. P (BAk) is the probability of section bk in an available state. So the section vailability can be calculated as
If the data file fi is distributed into ni fixed sections denoted by
which are distributed on different data nodes.
is the set of the numbers of replicas of the blocks of bi . The availability and unavailability of data file fi is given as:
The various stages that a data is processed in MGA algorithm is shown in Figure 1.
A set of randomly selected chromosomes are considered as initial population set in genetic algorithm. Every answer in the group of nodes is referred to each single node. Every node is represented as a chromosome for generating different operations. From adistinct group, a node will be selected, and operations are applied on them to form the next generation. A specific criterion is used for mating the chromosomes.
Fitness Function
The efficiency of a node depends on the fitness value. It is the measure of the supremacy of node in the group. Fitness (F) is calculated as the shortest distance between two nodes as shown in
The Probability of selecting best nodes using the fitness function is shown in
Here Pj is the probability of choosing jth node, N is number of nodes, Fj is the probability of jth node to be fit. The fitness value shows the efficiency of anode in the group. Therefore, the individual node will survive according to the fitness value. Hence, this function is the key factor in this algorithm.
Selection
The mechanism selects a solution for moving ahead in creating next generation nodes. This is a key operation to increase the productivity of the nodes. Several methods are used to select accurate nodesseveral selection strategies (Goyal & Agrawal 2013).
Crossover
Crossover operation creates individual nodes by reforming their parent nodes using hybridization method.
Mutation
Mutation is the next step that introduces diversity in the groups. This is applied to homogeneous population. This process alters the gene values of the nodes from its initial state. This results in a new set of values being added to the resource pool. With this the MGA may be able to get a optimal result.
The steps involved in Modified Genetic Algorithm for data replication is shown below,
begin
generate initial population with the selected nodes
while (node_count > 0)
calculate fitness F by calculating distance between two nodes
compute F= inverse of min(Ed(m,n))
evaluate each of the node for fitness
do
select parent node
crossover of parent node
mutation of descendant nodes
inspection of new nodes
select new nodes
Goto to step 2 until best solution obtained
end
The problem of data replication without considering their size and capacity may result in poor solution. The MGA algorithm is able to produce an improved replication cost than earlier genetic algorithm by improving the fitness function which is able to handle dynamic data and frequent data updates.
Greedy Search Algorithm (GSA)
Greedy algorithm selects the best option in each
iteration so as to proceed with best results which improves the performance. This
GSA method divides a large problem into smaller and smaller problems solution
will get a global solution (Pan et al., 2016).Greedy algorithm uses the minimum
spanning tree in the chart and Huffman encoding. It can be used as alternative
where precise result is not required.
The methodology of greedy algorithm is: Begins with
a primary solution of a problem to approach final goal in a step by step manner.
It will stop at a point called saturation level.
A simple method of data replication using greedy search technique is created. For each number of site Zi and item Uk, the repetition function value Ak is defined as
The above repetition function Ak represents the replication benefit in terms of network cost, if the item Uk replicated at Zi. This benefit is that the value Ak is computed by using the difference between the communication cost occurred from the current read requests. So it reduces the communication cost by half than the previous methods. However, for the adverse value of replication function Ak replicating the ith site in its local view of kth site will not produce an accurate result. This does not mean that network cost will be be always at its peak. But improving the site parameters the network cost can be managed.
A list Li is created for Ui containing all the nodes that are replicated. A node Nk can be replicated only when the storage capacity bi of the site is greater than the replication space requested by the node. A positive value should be generated while finding the difference betwwen the storage space and requsted replication space. Additionally, a list AL is maintained to include any additional resources provisioned in cloud and thus the replication capacity is improved.
The steps involved in Greedy Search algorithm is shown below,
begin
create list Li and additional list L1
while (Li and L1 ≠ NULL)
Initialize Um=0, Nm={}
/*Um contains present max value of Ak, Nm preserves node identity */
Select a node Uk belongs to list L1 in circular order
for each Nk in Li
calculate replication function Ak
if (Um<Aki) then Um= Aki and Nm=k
else
Li = Li – {Nk}
/* performes node update with neighbours */
for nodes in L1 change SNmi
New nodes included */
Ai= Ai – Nk
/* Remove resource on replication completion */
If Li = NULL then L1=L1- {Ui}
end while
This Greedy search replication algorithm generates improved results but it cannot guarantee that the final solution is the best. The MGA algorithm outperforms this algorithm in terms of capability of handling large data volume. The greedy algorithm can be applied extensively but cannot guarantee a optimal solution unlike genetic algorithm.
Results and Discussion
In this section, Modified Genetic Algorithm (MGA) and Greedy Search Algorithm (GSA) is compared with the random method for the data replication performance. A cloud environment with a range of minimum of 250 nodes to a maximum of 1000 nodes is taken for the experimentation. Evaluation results have been computed in terms of time taken for worst case recovery, average case recovery, recovery of 20% corrupt data and recovery of 40% corrupt data as shown in tables 1 to 4 and figures 2 to 5.
Table 1: Time taken for recovering worst case data corruption
Number of nodes
in Cloud
Random Method
MGA
Greedy Search
250
2542
2021
2502
500
3668
3301
3522
750
5209
4674
5108
1000
8142
6822
8022
Figure 2: Time taken for recovering worst case data corruption
From the figure 2, it can be observed that the Modified GA has lower worst case recover time of 22.83% than the random method and 21.68% than GSA for 250 nodes, 10.53% lesser than the random method and 8.87% than GSA for 500 nodes, 10.82% lesser than the random method and 8.87% than GSA for 750 nodes and 17.64% lesser than the random method and 16.16% than GSA for 1000 nodes.
Table 2: Time taken for recovering average case data corruption
Number of nodes
in Cloud
Random Method
MGA
Greedy Search
250
14
11
13
500
17
14
16
750
21
19
20
1000
27
24
25
Figure 3: Time taken for recovering average case data corruption
From the figure 3, it can be observed that the MGA has lower average case recover time of 24% than the random method and 16.66% than GSA for 250 nodes, 19.35% lesser than the random method and 13.33% than GSA for 500 nodes, 10% lesser than the random method and 5.12% than GSA for 750 nodes and 11.76% lesser than the random method and 4.08% than GSA for 1000 nodes.
Table 3 Time taken for recovering after 20% of data corruption
Number of nodes
in Cloud
Random Method
MGA
Greedy Search
250
6
5
5
500
8
7
8
750
11
8
10
1000
12
9
11
Figure 4: Time taken for recovering after 20% of data corruption
From the figure 4, it can be observed that after 20% of data corruption MGA has lower recovery time by 18.18% than the random method and GSA for 250 nodes, 14% lesser than the random method and 13.33% than GSA for 500 nodes, 31.57% lesser than the random method and 22.22% than GSA for 750 nodes, by 20% lesser recovery time than random method and GSA for 1000 nodes.
Table 4: Time taken for recovering after 40% of data corruption
Number of nodes
in Cloud
Random Method
MGA
Greedy Search
250
7
7
6
500
9
7
8
750
13
9
11
1000
13
12
11
Figure 5: Time taken for recovering after 40% of data corruption
From the figure 5, it can be observed that after 40% of data corruption MGA has same recovery time as that of random method and 15.38% higher than the GSA for 250 nodes, 25% lesser than the random method and 13.33% than GSA for 500 nodes, 36.36% lesser than the random method and 20% than GSA for 750 nodes and 8% lesser than the random method and 8.69% than GSA for 1000 nodes.
Conclusion
In this work, we propose the Modified Genetic Algorithm (MGA) and Greedy Search Method (GSM) to handle the drawbacks existing in the previous schemes. Both of these algorithms are based on deterministic approach in which lag in accessing remote data is reduced by storing the data closer to the services. The GSA approach considers an initial solution for a problem and fixes a set of goal in a step by step manner so as to obtain better solutions in a least time.The MGA method chooses a particular set of data blocks and applies the fitness function which selects best nodes for replication. Both of these methods find out the best nodes for replication, but they select a particular set of data blocks. If the best nodes are not available in those blocks, they will choose next set of blocks which results in increase of replication cost. But dynamic data are effectively handled by these methods by using genetic approach for population selection.
Acknowledgement
We acknowledge the support provided by The Professor and Head, Dept. of Physical Sciences and IT, TNAU, Coimbatore.
Conflict of
Interest
The authors do not have any conflict of interest.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Zunnurhain, K., & Vrbsky, S. V. (2011). Security in cloud computing. In Proceedings of the 2011 International Conference on Security & Management.
Hussein, M. K., & Mousa, M. H. (2012). A light-weight data replication for cloud data centers environment. International Journal of Engineering and Innovative Technology, 1(6), 169-175.
Boru, D., Kliazovich, D., Granelli, F., Bouvry, P., & Zomaya, A. Y. (2015). Energy-efficient data replication in cloud computing datacenters. Cluster computing, 18(1), 385-402. CrossRef
Liu, C. W., Hsien, W. F., Yang, C. C., & Hwang, M. S. (2016). A Survey of Public Auditing for Shared Data Storage with User Revocation in Cloud Computing. IJ Network Security, 18(4), 650-666. CrossRef
Zhang, J., & Dong, Q. (2016). Efficient ID-based public auditing for the outsourced data in cloud storage. Information Sciences, 343, 1-14. CrossRef
Wang, B., Li, B., Li, H., & Li, F. (2013, October). Certificateless public auditing for data integrity in the cloud. In Communications and Network Security (CNS), 2013 IEEE Conference on (pp. 136-144). IEEE. CrossRef
Lin, J. W., Chen, C. H., & Chang, J. M. (2013). QoS-aware data replication for data-intensive applications in cloud computing systems. IEEE Transactions on Cloud Computing, 1(1), 101-115. CrossRef
Boru, D., Kliazovich, D., Granelli, F., Bouvry, P., & Zomaya, A. Y. (2015). Energy-efficient data replication in cloud computing datacenters. Cluster computing, 18(1), 385-402. CrossRef
Zhang, Y., Xu, C., Li, H., & Liang, X. (2016). Cryptographic public verification of data integrity for cloud storage systems. IEEE Cloud Computing, 3(5), 44-52. CrossRef
Williams, H., & Bishop, M. (2014). Stochastic diffusion search: a comparison of swarm intelligence parameter estimation algorithms with ransac. Algorithms, 7(2), 206-228. CrossRef
El-henawy, I. M., & Ismail, M. M. (2014). A Hybrid Swarm Intelligence Technique for Solving Integer Multi-objective Problems. International Journal of Computer Applications, 87(3). CrossRef
al-Rifaie, M. M., Bishop, M. J., & Blackwell, T. (2011, July). An investigation into the merger of stochastic diffusion search and particle swarm optimisation. In Proceedings of the 13th annual conference on Genetic and evolutionary computation (pp. 37-44). ACM. CrossRef
Lin, J. W., Chen, C. H., & Chang, J. M. (2013). QoS-aware data replication for data-intensive applications in cloud computing systems. IEEE Transactions on Cloud Computing, 1(1), 101-115. CrossRef
Ye, Z., Zhou, X., &Bouguettaya, A. (2011, April). Genetic algorithm based QoS-aware service compositions in cloud computing. In International Conference on Database Systems for Advanced Applications (pp. 321-334). Springer, Berlin, Heidelberg. CrossRef
Sriprasadh, K., & Prakash Kumar, M. (2014). Ant colony optimization technique for secure various data retrieval in cloud computing. Int. J. Comput. Sci. Inf. Technol, 5(6), 7528-7531.
Huo, Y., Zhuang, Y., Gu, J., Ni, S., & Xue, Y. (2015). Discrete gbest-guided artificial bee colony algorithm for cloud service composition. Applied Intelligence, 42(4), 661-678. CrossRef
Połap, D., Woźniak, M., Napoli, C., & Tramontana, E. (2015). Real-time cloud-based game management system via cuckoo search algorithm. International Journal of Electronics and Telecommunications, 61(4), 333-338. CrossRef
Florence, A. P., & Shanthi, V. (2014). A load balancing model using firefly algorithm in cloud computing. Journal of Computer Science, 10(7), 1156. CrossRef
Jacob, L., Jeyakrishanan, V., & Sengottuvelan, P. (2014). Resource scheduling in cloud using bacterial foraging optimization algorithm. International Journal of Computer Applications, 92(1). CrossRef
Hu, X. (2015). Adaptive optimization of cloud security resource dispatching SFLA algorithm. Int. J. Eng. Sci.(IJES), 4(3), 39-43. CrossRef
Dashti, S. E., & Rahmani, A. M. (2016). Dynamic VMs placement for energy efficiency by PSO in cloud computing. Journal of Experimental & Theoretical Artificial Intelligence, 28(1-2), 97-112. CrossRef
Juels, A., & Wattenberg, M. (1996). Stochastic hillclimbing as a baseline method for evaluating genetic algorithms. In Advances in Neural Information Processing Systems (pp. 430-436). CrossRef
De, M. K., Slawomir, N. J., & Mark, B. (2006). Stochastic diffusion search: Partial function evaluation in swarm intelligence dynamic optimisation. In Stigmergic optimization (pp. 185-207). Springer, Berlin, Heidelberg. CrossRef
Zhan, S., & Huo, H. (2012). Improved PSO-based task scheduling algorithm in cloud computing. Journal of Information & Computational Science, 9(13), 3821-3829.
Wen, X., Huang, M., & Shi, J. (2012, October). Study on resources scheduling based on ACO allgorithm and PSO algorithm in cloud computing. In Distributed Computing and Applications to Business, Engineering & Science (DCABES), 2012 11th International Symposium on (pp. 219-222). IEEE. CrossRef
Zhou, J., & Yao, X. (2017). A hybrid approach combining modified artificial bee colony and cuckoo search algorithms for multi-objective cloud manufacturing service composition. International Journal of Production Research, 55(16), 4765-4784. CrossRef
Zhou, J., & Yao, X. (2017). A hybrid artificial bee colony algorithm for optimal selection of QoS-based cloud manufacturing service composition. The International Journal of Advanced Manufacturing Technology, 88(9-12), 3371-3387. CrossRef
Cavazos, J., Moss, J. E. B., & O’Boyle, M. F. (2006, January). Hybrid optimizations: Which optimization algorithm to use?. In Compiler Construction (pp. 124-138). Springer Berlin Heidelberg. CrossRef
Sedano, A., Sancibrian, R., de Juan, A., Viadero, F., & Egana, F. (2012). Hybrid optimization approach for the design of mechanisms using a new error estimator. Mathematical Problems in Engineering, 2012. CrossRef
Wang, G., & Guo, L. (2013). A novel hybrid bat algorithm with harmony search for global numerical optimization. Journal of Applied Mathematics, 2013. CrossRef
Milan, T. U. B. A. (2013). Artificial Bee Colony (ABC) algorithm with crossover and mutation. Appl. Soft Comput, 687-697.
Khanam, R., Kumar, R. R., & Kumar, C. (2018, March). QoS based cloud service composition with optimal set of services using PSO. In 2018 4th International Conference on Recent Advances in Information Technology (RAIT) (pp. 1-6). IEEE. CrossRef
Dai, Y., Lou, Y., & Lu, X. (2015, August). A task scheduling algorithm based on genetic algorithm and ant colony optimization algorithm with multi-QoS constraints in cloud computing. In Intelligent Human-Machine Systems and Cybernetics (IHMSC), 2015 7th International Conference on (Vol. 2, pp. 428-431). IEEE. CrossRef
Arindam Das and Ajanta De Sarkar, “On Fault Tolerance of Resources in Computational Grids”, International Journal of Grid Computing and Applications, Vol.3, No.3, Sept. 2012, DOI: 10.5121/ijgca.2012.3301. CrossRef
Armbrust M et al (2009) Above the clouds: a Berkeley view of cloud computing. UC Berkeley Technical Report
Armbrust, M.; Fox, A.; Griffith, R.; Joseph, A. D.; Katz, R. H.; Konwinski, A.; Lee, G.; Patterson, D. A.; Rabkin, A.; Stoica, I. &Zaharia, M. (2009), ‘Above the Clouds: A Berkeley View of Cloud Computing'(UCB/EECS-2009-28), Technical report, EECS Department, University of California, Berkeley.
Birk, Dominik. “Technical challenges of forensic investigations in cloud computing environments.” workshop on cryptography and security in clouds. 2011. CrossRef
Buyya, R., Sulistio, A.: Service and utility oriented distributed computing systems: challenges and opportunities for modeling and simulation communities utility-oriented computing systems. Symp. A.Q. J. Mod. For. Lit. 68–81 (2008) CrossRef
Foster, I., Zhao, Y., Raicu, I., Lu, S.: Cloud computing and grid computing 360-degree compared. Grid Comput. Environ. Work 1–10 (2008) CrossRef
Grover J., Sharma, M.: Cloud computing and its security issues—a review (2014). CrossRef
Kim, W.: Cloud computing: today and tomorrow. J. Object Technol. 8(1) (2009) CrossRef
Kulkarni, G., Waghmare, R., Palwe, R., Waykule, V., Bankar, H., &Koli, K. (2012, October). Cloud storage architecture. In Telecommunication Systems, Services, and Applications (TSSA), 2012 7th International Conference on (pp. 76-81). IEEE. CrossRef
Modi, C., Patel, D., Borisaniya, B., Patel, A., Rajarajan, M.: A survey on security issues and solutions at different layers of Cloud computing, pp. 1–32 (2012). CrossRef
Mukundha, C., &Vidyamadhuri, K. (2017). Cloud Computing Models: A Survey. Advances in Computational Sciences and Technology, 10(5), 747-761.
Nepal, Surya, et al. “DIaaS: Data integrity as a service in the cloud.” Cloud Computing (CLOUD), 2011 IEEE International Conference on. IEEE, 2011. CrossRef
NIST Definition of Cloud Computing- https://www.nist.gov/programsprojects/cloud-computing
Popović, K., &Hocenski, Ž. (2010, May). Cloud computing security issues and challenges. In MIPRO, 2010 proceedings of the 33rd international convention (pp. 344-349). IEEE.
PDF Downloads:
404
Open
Access -
