1. How can we prove that cloud files are encrypted.?
Hourglass Schemes
Marten Van Dijk ,Ari Juels , Alina Oprea
Ronald , Emil Stefanov, Nikos Triandopoulos
Presented by Siva Prasad Reddy Nooli

2. Overview Introduction General Framework Hourglass Scheme
Alternative approaches
Framework Description (Encoding and Operation)
Hourglass Protocol Phases
Hourglass Functions (Butterfly, Permutation and RSA)
Hourglass Performance
Conclusion

3. Introduction Uncontrolled data leakage is an enormous problem today
Cloud providers offer the service with simple API
Data encryption is most required to secure client data
Very difficult to provide transparency to tenants about handling the sensitive data
The General Framework uses watermark so that source of data leakage is identified

4. General Framework
General framework is constructed that will design the protocols for ensuring that data of client is stored “at the rest” in an encoding of client’s choice
General Framework will help cloud service provider to understand the source of data leakage with embedded watermark
The framework introduces the concept of encoding cloud data in clients choice and remotely verifying the data
A misbehaved cloud provider would have to double their storage in order to reply correctly to client challenges

5. Hourglass Scheme Named after the ancient time telling device Hourglass
Hourglass Scheme translates cipher text file into encapsulation file
Encapsulation file is denoted by H
Cipher text file is denoted by G
Time period for encapsulation is denoted by T
When challenged by client the cloud provider should able to retrieve the file in < T

6. Hourglass Schemes Solution
encryption
hourglass
Original File
Encrypted File
Encapsulated File
client verifies encryption
client assists
client verifies by periodically challenging random file blocks
client uploads file
(Source: arijuels.com)

7. Example of Hourglass Function
Inversion of small image hash function will help in gaining intuition that assumes computational source
F is consisted of “n” blocks each of l bits, that helps the client to verify format preserving encryption to F
Recovery of G from H is easily possible by computing hash values of Gi
For small values of l, transformation of Gi to Hi is computationally costly on average
Now let the case be such that the client wishes to verify format preserving encryption to F for obtaining cipher text G

8. Example of Hourglass Function… (Cont.)
Let h: {0, 1}* {0, 1}l, denotes fixed-length hash function.
Hi = h-1 (Gi) for every i that belongs to {1, … n}
Recovery of G from H is very practical from this function Gi = h (Hi) for i ϵ [1; n]
Even for fairly small values of l (e.g 35-bit or 40-bit blocks)The transformation from Gi to Hi is computationally cost on average
A honest server that has stored H just retrieves and serves Hi and is a quick operation
A cheating server that has stored only plain text F will take more to generate the file Hi on the fly

9. Alternative approaches
Trusted computing offers authentic information
Ensure that all the data of the client is stored on the enabled hardware
Data can be handled by client from different location .
(Source: Diva-portal.org)

10. Framework Description
Framework Encoding: Specifies an encoding algorithm under specified input tenant
Framework Operation: Hourglass Scheme HG is specified
(Source: Fu and Kone,)

11. Framework Encoding
Components of Framework encoding

12. Framework Operation
Framework Operation Components

13. Generic Hourglass Protocol
Assurance to client that data is outsourced in encoded format G
Fast output encode by watermarking and encryption
Help server to store file F on the air
Randomly selected blocks of H is chosen by client to ensure that F is stored in H format
Hourglass Protocol has three phases- file encoding, hourglass encapsulation and format checking

14. General Hourglass Protocol

15. Hourglass Protocol Phases
1. File Encoding 2. Hourglass Encapsulation 3. Format Checking

16. Hourglass Function (Butterfly)
Example of butterfly network with case n = 8
Each set of values Gj [1],…Gj [n] is row of nodes
Two edges connect to each other at “w”
Hourglass involves n log2n
Total computational cost O(n log n)
Structure of overlapping binary series
Function is full-file PRP d = log2n
Cryptographic operation w to pair of blocks in sequence

17. Hourglass Function (Permutation based)
Permutation based hourglass function
Hourglass Function compute H to be the permutation of data element in G permutation that excessively G symbols across H, achieve highest security
Hourglass Function (Permutation based)
No cryptographic operation
Security depends on performance characteristics of rotational drives
Block size “l” of encryption algorithm must be larger than symbol size z
Security Analysis is consisted of Drive Model, Time Bound and Parallelism

18. Hourglass Function (RSA Based)
Apply RSA signing to translate encode format G to H for file F
Each block of encoded file is mapped with hourglass block
Resource is computation based and no assumption for storage usage

19. Typical RSA Hourglass function (Source:arijuels.com)… 𝑭:
𝑭 𝟏
𝑭 𝟐
𝑭 𝟑
𝑭 𝟒
𝑭 𝒏 … 𝑮:
𝑮 𝟏
𝑮 𝟐
𝑮 𝟑
𝑮 𝟒
𝑮 𝒏 … 𝑯:
𝑯 𝟏
𝑯 𝟐
𝑯 𝟑
𝑯 𝟒
𝑯 𝒏

20. Comparison of Hourglass Functions
𝑶 𝒏 RSA exponentiations
𝑶 𝒏 𝐥𝐨𝐠 𝒏 AES operations
𝑶 𝒏 random memory accesses
less practical
more practical
RSA
Butterfly
Permutation
less assumptions
more assumptions
RSA assumptions
storage speed
seek inefficiency
in rotational drives
(Source: arijuels.com)

21. Hourglass Performance
Butterfly Function is minimum of 4 times faster than local machine due to AES support
Permutation hourglass function is 8 times faster on Amazon due to cryptographic operation
RSA is less efficient than butterfly and permutation

22. In-memory performance of Hourglass Function (Butterfly-Permutation Scheme) local machine Amazon EC2

23. Conclusion
Hourglass scheme with cryptographic construction is introduced
Hourglass schemes leverage server resource bounds to achieve their security
Clients will get remote access to verify their files in terms of security
Hourglass Scheme will prove to be most valuable way of penetrating abstraction layer of cloud
Hourglass Scheme will restore security assurances at times of cloud- based outsourcing

24. Reference List
Hourglass Schemes: How to prove that cloud files are Encrypted :- Marten van Dijk, Ari Juels, Alina Oprea, Ronald L. Rivest , Emil Stefanov, Nikos Triandopoulos
M. Abadi, M. Burrows, M. Manasse, and T. Wobber. Moderately hard, memory-bound functions. ACM Trans. Internet Technol., 5:299– 327, May 2005
E. Giberti. Honesty box: EBS performance revisited. Blog posting, available at
M. Jakobsson and A. Juels. Proofs of work and bread pudding protocols. In Communications and Multimedia Security, pages 258– 272, 1999
M. H. Manshaei, Q. Zhu, T. Alpcan, and J.-p. Hubaux. Game theory meets network security and privacy. Main, V(April):1–44, 2010

25. Thank you.!