173 Chapter 8 Simulation of Protocols 8.1. Introduction In this chapter, we analyze the security and performance of proposed protocols: HDVP, RSA-DPAP, ECC-DPAP, PVDSSP and EDVP by using simulation results. The simulation results were realized by using Network Simulator (NS-2), MATLAB 9.0, Statistical Tool box and proposed verification protocols applied to outsourced data storage applications in cloud to show the security and performance of these verification protocols. For the sake of completeness, we implemented proposed protocols in windows. Our experiments are conducted on a system with an Intel Core 2 processor running at 2.4 GHz, 4GB RAM, and a 7200 RPM Western Digital 320 GB Serial ATA drive with an 8 MB buffer with. All programs are written with help of Pairing-Based Cryptography (PBC) library version 0.4.18, the crypto library of OpenSSL version 0.9.8h and Sobol_Data Set library. Our implementation utilizes storage services/application: Amazon Simple Storage Service (S3). Storage service: Amazon Simple Storage Service (S3) is a scalable, pay-per use online storage service. The Clients can store an unlimited amount of data, paying for only the storage space and bandwidth that they are using, without initial startup fee. The basic data unit in S3 is an object, and the basic container for objects in S3 is called a bucket. For example, objects contain both data and metadata. A single object has a size limit of 5GB, but there is no limit on the number of objects per bucket. Moreover, a small script on Amazon Elastic Compute Cloud (EC2) is used to provide the support for verification protocol and dynamic data operations. 8.2. Experimental Results In this section, we present and discuss the experimental results for security and performance of all our proposed protocols and compare the results. 8.2.1. Security Here, we conduct experimental results for testing the Integrity, Confidentiality and Availability of the data for data storage applications.
30
Embed
Chapter 8 Simulation of Protocols - INFLIBNETshodhganga.inflibnet.ac.in/bitstream/10603/20850/17/17_chapter 8.pdfChapter 8 Simulation of Protocols 8.1. ... RSA-DPAP, ECC-DPAP, PVDSSP
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
173
Chapter 8 Simulation of Protocols
8.1. Introduction
In this chapter, we analyze the security and performance of proposed protocols: HDVP,
RSA-DPAP, ECC-DPAP, PVDSSP and EDVP by using simulation results.
The simulation results were realized by using Network Simulator (NS-2), MATLAB 9.0,
Statistical Tool box and proposed verification protocols applied to outsourced data storage
applications in cloud to show the security and performance of these verification protocols.
For the sake of completeness, we implemented proposed protocols in windows. Our
experiments are conducted on a system with an Intel Core 2 processor running at 2.4 GHz, 4GB
RAM, and a 7200 RPM Western Digital 320 GB Serial ATA drive with an 8 MB buffer with. All
programs are written with help of Pairing-Based Cryptography (PBC) library version 0.4.18, the
crypto library of OpenSSL version 0.9.8h and Sobol_Data Set library. Our implementation
utilizes storage services/application: Amazon Simple Storage Service (S3).
Storage service: Amazon Simple Storage Service (S3) is a scalable, pay-per use online
storage service. The Clients can store an unlimited amount of data, paying for only the storage
space and bandwidth that they are using, without initial startup fee. The basic data unit in S3 is
an object, and the basic container for objects in S3 is called a bucket. For example, objects
contain both data and metadata. A single object has a size limit of 5GB, but there is no limit on
the number of objects per bucket. Moreover, a small script on Amazon Elastic Compute Cloud
(EC2) is used to provide the support for verification protocol and dynamic data operations.
8.2. Experimental Results
In this section, we present and discuss the experimental results for security and performance
of all our proposed protocols and compare the results.
8.2.1. Security
Here, we conduct experimental results for testing the Integrity, Confidentiality and
Availability of the data for data storage applications.
174
a)Integrity
To test the Integrity of data, we consider two parameters: Probability Detection and
Verification Time.
1) Probability Detection is the corruption of data should be detected with high probability as
soon as possible. The Probabilistic detection of data corruption assurance on data Integrity
increased with the iteration of the verification protocols. The main problem is the detection of
such corruption in less time.
2) Verification time is the time taken for probability detection of it being corrupted blocks,
less time is always preferable.
We simulated the proposed verification protocols and existing protocol with 100, 500 and
1000 node cloud network using NS2 to test the Integrity in terms of verification time for the
detecting data corruptions with high probability (99%) and compare the verification time of all
Integrity verification protocols. We assume that nodes that data is deleted or modified by the
store. Each simulation step is identifying the corrupted data on the verification time. We consider
the different randomly corrupt a percentage ranging from 1% to 10% of the data with 1GB of the
data file.
Fig. 8.1-8.3 presents the verification time (in seconds) for the detection of different data
corruptions range from 1% to 20% of 1GB file with 99% probability using proposed protocols
and Wang et al.[165] protocol in 100, 500 and 1000 node cloud network.
175
1% 5% 10% 15% 20%0
50
100
150
200
Data Corruption
Tim
e(S
)
Wang et al.[165](99%)
HDVP(99%)
RSA-DPAP(99%)
ECC-DPAP(99%)
PVDSSP(99%)
EDVP(99%)
Fig. 8.1 Comparisons of verification time between proposed protocols and Wang’s protocol
for the detection of different data corruptions with detection probability maintained at 99
percent in 100 nodes.
176
1 5 10 15 200
50
100
150
200
250
300
350
Data Corruption(%)
Tim
e(S
)Wang et al.[165](99%)
HDVP(99%)
RSA-DPAP(99%)
ECC-DPAP(99%)
PVDSSP(99%)
EDVP(99%)
Fig. 8.2 Comparisons of verification time between proposed protocols and Wang’s protocol
for the detection of different data corruptions with detection probability maintained at 99
percent in 500 nodes.
177
1 5 10 15 200
50
100
150
200
250
300
350
400
Data Corruption(%)
Tim
e(S
)
Wang et al.[165](99%)
HDVP(99%)
RSA-DPAP(99%)
ECC-DPAP(99%)
PVDSSP(99%)
EDVP(99%)
Fig. 8.3 Comparisons of verification time between proposed protocols and Wang’s protocol
for the detection of different data corruptions with detection probability maintained at 99
percent in 1000 nodes.
As observed from Fig. 8.1-8.3, the proposed protocols are very fast at detecting data
corruptions in cloud than Wang et al. [165]. In proposed protocols, the HDVP is useful for small
size applications and it is not suitable for large size data storage application when Clients having
less constrained resources. The RSA-DPAP suitable for large size data storage application but it
creates heavy overhead on processer due to large key size. Similarly, ECC-DPAP is more
suitable for small, medium and large size applications even when Clients having less constrained
resources(PDA, smart phones) and detects the corruptions faster than RSA-DPAP due to the less
key size. The PVDSSP is also useful for all types of applications and it takes very less
verification time to detect data corruptions. Finaly, the EDVP protocol detects the corruptions
more efficiently when compare to all above protocols.
178
Statistical Inference on Integrity of proposed protocols using one-way ANOVA
Consider a one-way ANOVA having experimental results of proposed verification protocols
for verification time to detect the data corruptions.
The hypothesis is assumed as follows:
Null hypothesis H0: There is no significant difference in the verification time of the
proposed algorithms tested.
Alternate hypothesis H1: there is significant difference between the verification times of
the proposed algorithms tested.
Table 8.1: ANOVA Table for Comparison of the Verification Time of Proposed Protocols No. of Nodes source SS df MS F Prob>F
100
Columns 12948.4 4 3237.1 11.59 0.0004932
Error 5585 24 279.25
Total 18533.4 28
500
Columns 35713.6 4 8928.39 9.86 0.0001
Error 18118.7 24 905.93
Total 53832.3 28
1000
Columns 47000 4 11750.01 7.26 0.0009
Error 32355.9 28 1671.79
Total 79355.9 28
SS: Sum of Squares, df: degrees of freedom, MS: mean square F:F-distribution, Prob: Probabability
The test statistic is the F value of 11.59, 9.86, and 7.26 from Table 8.1 for 100, 500 and
1000 nodes respectively. Using an α of .05, we have that F.05; 4, 24 = 2.87 from the F distribution
table. Since the test statistic is much larger than the critical value, we reject the null hypothesis of
equal verification time means and conclude that there is a (statistically) significant difference
among the verification times of proposed protocols. The p-value for 11.59, 9.86, and 7.26 are
0.000493, 0.0001, and 0.0009 respectively from Table 8.1, so the test statistic is significant at
that level.
The p-value returned by anova1 depends on assumptions about the random disturbances εij
in the model equation. For the p-value to be correct, these disturbances need to be independent,