Anonymous Statistical Survey of Attributes

1
Anonymous Statistical Surveyof Attributes

• Toru Nakanishi and Yuji Sugiyama
• Okayama Univ., Japan

2
Background
Distributor(Seller)
User(Customer)
3
Background(Contd)
Maybe useful for identifying the user.
4
Background(Contd)
Survey system to generate only the statistical
results is in demand.
5
Requirements in the survey system

• Anonymity of users
• No extra information beyond statistical results
• Correctness of results

6
Related Work

• Sako proposed a protocol to generate statistical
  results of attributes

7
Problem in simple application

• But,

• Sakos protocol may be simply applied to
  anonymous statistical survey.

Is single TTP really trusted ?
Are users honest ?
8
Our anonymous statistical survey system of
attributes

• Users cannot cheat.
• Each TTP doesnt have extra information.

trusted
Quorum is trusted
9
Tool 1 Camenisch-Stadlers group signature

• Whats a group signature ?

10
Tool 1 Camenisch-Stadlers group signature
(Contd)

• Registration

• Signing

11
Tool 2 Threshold Cryptosystem

• Only quorum of a group can decrypt a ciphertext.

???
12
Tool 3 Shuffle
13
Model
14
Our survey system - Registration

• Registration in group signature is executed.

• Registration in group signature is executed.
• zs are published in lists of respective
  attributes.

Female
15
Our survey system - Offering

• The group signature is offered.

16
Our survey system - Generating

• Sent ciphertexts are shuffled.

17
Our survey system Generating (Contd)

• For each shuffled ciphertext, its linked to
  attribute, with no extra information of z.
• Public zs are shuffled by the same random r,

while the ciphertext is randomized by r.
18
Our survey system Generating (Contd)

• Count revealed attributes, and calculate
  statistic.

• Decrypt the ciphertext,

and search in lists of zs.
19
Security

• Correctness

• Anonymity

Anonymity in offering
No extra information in generating
Correctness of offering
Correctness of generating
20
Conclusion

• An anonymous statistical survey system of
  attributes is proposed.
• No extra information for each trustee
• No cheating