4. The cloud computing model allows access to information
and computer resources from anywhere that a network
connection is available. Cloud computing provides a shared
pool of resources, including data storage space, networks,
computer processing power, and specialized corporate and
user applications.
Cloud computing is a practical approach to experience
direct cost benefits and it has the potential to transform a
data center from a capital-intensive set up to a variable
priced environment. The idea of cloud computing is based on
a very fundamental principal of „reusability of IT capabilities'.
4
5. INTRODUCTION(CNTD)
5
Data security
Can’t implement Traditional cryptographic technology .
Cloud - not a third-party warehouse.
Data stored in multiple physical locations in random
manner
Steganographic Approach Using Huffman Coding
ensures explicit dynamic data support
security of data when these data are in the cloud storage.
6. INTRODUCTION(CONT)
The Huffman Tree constructs an optimal prefix code called
a Huffman code.
Let’s say, there are six characters A,B,C,D,E and F as shown
in Fig a .
6
7. INTRODUCTION(CONT) 7
Now for a given code 0 100 100 1101 we can decode them
to get back the original code by traversing the Huffman tree.
8. CLOUD COMPUTING ARCHITECTURE AND
SECURITY ISSUES
DEPLOYMENT MODELS
Private cloud
Community cloud
Public cloud
Hybrid cloud
SERVICE DELIVERY MODELS
Software as a Service(SaaS)
Platform as a Service(PaaS)
Infrastructure as a Srevice(IaaS)
8
INTRODUCTION(CONT)
9. SECURITY ISSUES
9
Phishing
data loss
botnet (Collection of machines are running remotely).
botnet - offers more reliable infrastructure at a relatively low price
for attack.
INTRODUCTION(CONT)
10. Problem statement :
Main problem - loss of control of data stored in the
cloud.
10
Schematic System Architecture for Cloud
INTRODUCTION(CONT)
11. RELATED WORK
cong wang et al.use homomorphic token with distributed
verification of erasure-coded data.
but it is failed to achieve public verifiability and storage correctness.
shantanu pal et al. ensures to find location of adversary or the
attacking party from its target.
it may try to attack them, if adversary knows the location of the other vms.
this may harm the other vms in between.
ateniese et al.proposed the “provable data possession” (pdp)
model to ensure possession of file in untrusted storages.
This scheme used public key based homomorphic tags to audit the data file
and it is providing public verifiability.
11
12. DESIGN OF THE SYSTEM
Storing data into some images. - steganography .
12
Processes to store or retrieve their data :
14. Human Visual System(HVS) has very low sensitivity.
Variable length encoding doesn’t help attacker to recognize
characters.
He/she has no idea about frequency of characters.
Can’t generate Huffman code.
Ultimately we are having a secured system
14
15. Image database
•Images stored in CSP-1
•Set of images sent to CSP-3 - user wants to store data in cloud
File database
•File holds the address of images
Embedded data into Images
•Counts total no. of characters
•Finds frequency of each characters by Huffman code
•Applies Steganography to both frequency of characters & codified
data.
15
16. ALGORITHMS USED IN THE SYSTEM
ALGORITHM 1 : HF-codification()
1. procedure
2. Read file FText which is to be saved in Cloud
3. Compute CN from FText
4. Find the frequency of occurrences of each characters in
Ftext and store them in some chronological order
5. Store frequency in a new file FFreq .
6. FN = Freq-Codification( )
16
17. 17
7. Call Huffman-Tree()
8. Create a file FCode
9. Open FN. Reach EOF of FN where the original
characters of FText will be replaced by the Huffman
codes present in FCode .
10. Calculate the total Bit BCount in FN.
11. Delete FText, F Freq and F Code.
12. Call Steganography() to perform steganography on FN
13. end procedure
18. FILE CODIFICATION
frequency file is read digit by digit & each digit is codified into 4-
bit binary pattern
Algorithm 2: Freq-Codification ()
1. procedure
2. Open FFreq and a new File FN.
3. while ( Read characters from F Freq until EOF )
4. do if (character is a new line character)
5. Append 1111 at the end of FN.
6. else
18
19. 19
7. Convert the digit to its 4-bit binary form.
8. Append those 4-bits at the end of FN.
9. end if
10. end while
11. Append 11111111 at the end of FN.
12. Return FN
13. end procedure
Algorithm 2: Freq-Codification () (CNTD…)
20. Hiding Data within Images Steganography
Deals the pre-requisite requirements like :
load image, store file name, image index
finally call the MdfImg operation which will map data from file to images.
ALGORITHM 3: STEGANOGRAPHY()
1. procedure
2. Load Image_Index = ImageSearch (Image_Database)
3. Store (FName, BCount, Image_Index)
4. MdfImg (Image_Database [Image_Index]);
5. end procedure.
20
21. SEARCHING OF VALID IMAGE
The algorithm searches an image which we can be used to store the
data. It returns the address of a valid image if it is available in image
database.
ALGORITHM 4: IMAGESEARCH(IMAGE_DATABASE)
1. procedure
2. Open Image_Database;
3. for Image_Database(i), i<-1, n do
4. if (Image_Database(i).valid==1)
5. return i
6. end if
7. end for
8. end procedure
21
22. 22
MAPPING DATA FROM A FILE TO IMAGE
It does the actual steganographic operation by storing data into
images.
Algorithm 5: MdfImg (Image_Database [Image_Index])
1. procedure
2. Read Image_Database [Image_Index];
3. Compute Pixel Count for Image_Database[Image_Index];
4. Open FN
5. while (Read Characters until EOF)
6. do if (Pixel Count < B Count) …
23. 7. Last bit of each consecutive pixels of the
Image_Database[Image_Index] is replaced by Store each character.
8. else
9. Load Image_Index1=ImageSearch(Image_Database)
10. Image_Database[Image_Index1].valid=Image_Index1
11. Image_Database [Image_Index1].valid=0
12. end if
13. end while
14. end procedure
23Algorithm 5: MdfImg (Image_Database [Image_Index]) (CNTD)
24. RETRIEVING DATA FROM IMAGE
The following algorithm retrieves the data from the images
which is kept in cloud storage.
Algorithm 6: RetrieveData ()
1. procedure
2. Read File_Database;
3. for F_Database (i), i=1 to m
4. do if (F exits)
5. I=Holds the address of image.
6. end if
7. end for
8. Open Image_Database;
9. Read Image_Database [I];
10. Open a F Temp and a F Freq
24
25. 11. while (Until we get 11111111 in Image_Database [I])
12. do Read 4 bits at a time from 4 consecutive pixels
13. Convert them into decimal form.
14. sum =sum + 4
15. if ( decimal number is within 0 to 9 )
16. Write that digit in F Freq
17. else
18. Write new line character in F Freq
19. end if
20. end while
21. sum= sum - BCount
22. Call Huffman-Tree( ) based on the frequency counts
present in FFreq and create the HuffmanTree.
25
26. 23. while ( sum <= BCount ) do
24. read bits from Image_Database [I]
25. Start traversing the Huffman-Tree from root.
26. When we reach leaf node, we will get character.
27. Append that character in FTemp .
28. Increment sum number of times we collect bits
from Image_Database [I]
29. end while
30. Show F Temp to the user, after user closes the file
FTemp, delete the file FTemp from system.
31. end procedure
26
27. CONSTRUCTION OF HUFFMAN TREE
A priority queue, Q, is used to generate Huffman Tree with levels
(frequency) as key.
Algorithm 7a: Huffman-Tree (X)
1. procedure
2. FN=|X|
3. Q=X
4. for i=1 to N-1
5. do
6. Z=Allocate_node( )
7. Z.left=Extract_min(Q)
8. Z.right=Extract_min(Q)
9. Frequency(Z)=Frequency(Z.left)+Frequency(Z.right)
10. Insert(Q,Z)
11. end for
12. end procedure
27
28. Algorithm 7b: Allocate_node()
1. Procedure
2. Create a node for storing characters and their
frequency from available free memory space.
3. Return the allocated node.
4. End procedure
Algorithm 7c: Extract_min(Q)
1. procedure
2. Remove and return the character with minimum
frequency from the priority queue Q.
3. end procedure
28
30. Security Analysis and Performance
Evaluation
Huffman coding is a variable length coding scheme.
The frequency of each character is stored in some chronological order.
Variable length encoding does not help the attacker to recognize the
characters.
Decoding of bits can only be done by the Huffman tree only.
The frequency file contains only frequency.
Change in chronological order results difficulty of tracking characters.
Ultimately we are having a secured system.
30
31. •SECURITY STRENGTH AGAINST CSP-1
CSP-1 only stores some files.
CSP-1 does not contain the retrieving algorithm, thus
the images containing data are purely safe.
31
•SECURITY STRENGTH AGAINST CSP-2
Retrieving and hiding mechanism are stored in CSP-2.
Knowing only the algorithm will not help the attacker.
•SECURITY STRENGTH AGAINST CSP-3
CSP-3 is responsible for computation.
All files will be deleted after the above operations.
32. CONCLUSION
We applied steganographic approach to ensure data storage
security in cloud computing using Huffman Coding (SAHC).
Through detailed security and performance analysis this
approach gives high security of data when it is on rest in the data
center of any Cloud Service Provider (CSP).
This proposed architecture will be able to provide customer
satisfaction to a great level and it will attract more clients in the
field of cloud computing for industrial as well as future research
firms.
32
33. REFERENCES
[1] Peter Mell, Timothy Grance, “The NIST Definatin of Cloud Computing”, Jan,
2011.http://docs. ismgcorp.com/files/external/Draft-SP-800-145_clouddefinition.pdf.
[2] Amazon.com, “Amazon Web Services (AWS)”, Online at hppt://aws.amazon.com,
2008.
[3] Con Wang, Qian Wang, Kui Ren, and Wenjng Lou,“Ensuring Data Storage Security
in CloudComputing”,17th International workshop on Quality of service, USA, pp1-9,
2009, IBSN:978-42443875-4.
[4] Thomas H. Cormen, Charles E. Leiserson, Ronald L.Rivest, and Clifford Stein.
Introduction to Algorithms,Third Edition, Prentice Hall of India, 2010.
[5] B.P Rimal, Choi Eunmi,I.Lumb, “A Taxonomy and Survey of Cloud Computing
Sytem”, Intl. Joint Conference on INC, IMS and IDC, 2009,pp.44-51, Seoul,Aug, 2009.
DOI : 10.1109/NCM.2009.218.
33
