2. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Summary
2
What is Tokenization in two words
Definition & Functionalities
Use cases
How does it work ?
Random Token System
Format Preserving Encryption
Conclusion
3. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [DEFINITION]
▶ Definition
– Tokenization is a process of replacing sensitive data by non sensitive ones
(tokens) with respect of the following properties:
• Tokens bears enough information to be useful (e.g. The entity manipuling
token can accomplish transaction as it was the sensitive data).
• Tokens does not compromise security
– Tokenization system tries to minimize the integration impact on existing
infrastructure
▶ Who offer such service
3
4. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [FUNCTIONALITIES]
▶ Function description of a Tokenization system
– Conversion (Convert sensitive data into a token and vice versa)
– Conversion policy (Format definition, Mode of operation)
– Communication Canal : Authentication, Integrity, Confidentiality
4
Tokenization
System
External
System
5. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [USE-CASES]
▶ Focusing on payment (but not limited to)
– Context :
• Sensitive data : PAN,…
• PCI compliancy
– Use cases
• MOTO
• Face2Face
5
CardHolder
AcquirerIssuer
Acceptor
6. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [USE-CASES]
▶ Focusing on payment (but not limited to)
– Context :
• Sensitive data : PAN,…
• PCI compliancy
– Use cases
• MOTO
• Proximity payment
6
CardHolder
AcquirerIssuer
Acceptor
E2E-Encryption
7. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [USE-CASES]
▶ Focusing on payment (but not limited to)
– Context :
• Sensitive data : PAN,…
• PCI compliancy
– Use cases
• MOTO
• Proximity payment
7
CardHolder
AcquirerIssuer
Acceptor
Secure MOTO
8. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
What is tokenization in two words [USE-CASES]
▶ Focusing on payment (but not limited to)
– Context :
• Sensitive data : PAN,…
• PCI compliancy
– Use cases
• MOTO
• Proximity payment
8
CardHolder
AcquirerIssuer
Acceptor
Process
transaction
9. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Summary
9
What is Tokenization in two words
Definition & Functionalities
Use cases
How does it work ?
Random Token System
Format Preserving Encryption
Conclusion
10. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Tokenization and Format Preserving Encryption: A Case Study
▶PCI-DSS(Payment Card Industry Data Security Standard) :
• Security requirements for entities processing cards data
(processing, transmission and storage)
▶Objectives:
• Reduce PCI evaluation perimeter
• Choose a suitable algorithm that tokenize a PAN
▶Constraints:
• The algorithm must be collision free
• In a certain mode the algorithm must be “not reversible”
• In certain mode the algorithm must not takes secret parameters
10
How does it works ? [Objectives-Constraints]
11. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
How does it works [RandomToken]
▶ Random Token
– Card data are
• ciphered (classic algorithms)
• stored in a database
– System generate an associated token
• Format respect
• Checks for no Collision
11
Tokenization and Format Preserving Encryption: A Case Study
12. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Summary
12
What is Tokenization in two words
Definition & Functionalities
Use cases
How does it work ?
Random Token System
Format Preserving Encryption
Conclusion
13. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶ FPE : Format Preserving Encryption.
▶ Introduced by Brightwell [BS97]
o Encryption scheme with
o format preserving property
▶ Format definition is a key point
– Follow PCI guidelines :
• you have to differentiate a Token from a PAN
▶ NIST is considering 3 FPE algorithms
▶ Applications :
• Security Social Number
• Credit Card Number
13
How does it works [FPE based tokenization]
Tokenization and Format Preserving Encryption: A Case Study
First introduction of Format Preserving
Encryption [BS97]
14. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶ NIST is considering 3 FPE algorithms
• FFX [FFX10]
• BPS [BPS10]
• FCEM [FCEM10]
14
How does it works [FPE based tokenization]
Tokenization and Format Preserving Encryption: A Case Study
FFX BPS
FCEM
15. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
15
▶ Feistel
o Inventé par Horst Feistel .
o Round notion
o Input are split in 2
o F : cipher function
o Secret key K
o Key Derivation algorithm
o During a round
Ai+1 = Bi
Bi+1 = Ai Fki(Bi)
o Example
DES : 16 tours.
+
How does it works [Cryptographic-Approach]
Tokenization and Format Preserving Encryption: A Case Study
16. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Feature FFX BPS FCEM
Feistel based Yes Yes No
#Rounds 12 8 2
Cipher function AES AES/TDES/SHA AES
#Function is used 12 8 8
Reversibility Yes Yes Yes
Tweak Yes Yes No
16
How does it works [Cryptographic-Approach]
Tokenization and Format Preserving Encryption: A Case Study
▶ Cryptographic notions
– Tweak Notion : Add variability in cryptographic schemes
– Patarin attack : Differentiate ciphertext from random string
17. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Feature Random Token FPE
Multi Site Difficult Medium
Key deployment Medium Hard
Format preserving Easy Easy
Performance Low Fast
Token/Data link No (except in DB) Algorithm
17
Tokenization and Format Preserving Encryption: A Case Study
How does it works [Analysis]
18. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Summary
18
What is Tokenization in two words
Definition & Functionalities
Use cases
How does it work ?
Random Token System
Format Preserving Encryption
Conclusion
19. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
Conclusion [VISION]
▶ Which choice ?
19
CardHolder
AcquirerIssuer
Acceptor
Secure MOTO
Process
transaction
FPE
RTS
E2E-Encryption
FPE FPE
Tokenization and Format Preserving Encryption: A Case Study
20. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶ Tokenization in payment context
It allows the reduction of PCI audit perimeter in a payment application
Waiting for NIST approval.
▶ Depending on use case:
Random Tokenization:
In case of internal processing
FPE based Tokenzaton
In case of multi site,
In case of multi-party protocols
20
Tokenization and Format Preserving Encryption: A Case Study
Conclusion []
22. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
22
BPS
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
23. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
23
▶ BPS :
▶ Autor: Brier E, Peyrin T & Stern J.
▶ Published in 2010.
▶ BPS : "a Format Preserving Encryption
Proposal ".
▶ Features:
• 8 round.
• Tweak of 64 bits split in 2 sub tweak
o TL et TR
• F : AES or one way function.
• K : secret key
• reversible.
• Patarin resistant.
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
24. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶
24
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
25. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
25
FFX
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
26. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶ FFX :
▶ Autors : Bellare M, Rogaway P & Spies T.
▶ Published in 2009 and 2010.
▶ FFX : "Format Preserving Feistel-based Encryption"
▶ Features:
• 12 round,
• 64 bits tweak,
• FK : AES-128 or one-way function
• K : secret key
• reversible
26
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
27. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
27
27
FCEM
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
28. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
28
▶ Autor :Ulf T Matsson.
▶ Published in 2009.
▶ FCEM : "Format Controlling Encryption Mode".
▶ Features:
• 8 steps
o Index Value Data
o Encryption of Left
o Encryption of Right
o Scrambled
o Rippled Left to Right
o Rippled Right to Left
o Encryption and Update
o The last transformation
• F : AES-128
• K : secret key
• reversible
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
29. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
▶ Index Value data :
• Rewriting input as hexa values.
• Example:
o X : 1122334455667788
o Index Value data : 01010202030304040505060607070808
▶ Encryption of Left :
• left part encryption
• Example :
o Index Value data : 01010202030304040505060607070808
o Sortie de FK: 00C01F49D0C2C050188D8FDFADCDF846
o RightUpdate : 0507070905010008
▶ Encryption of Right :
• Same idea
• We get LeftUpdate : 0101080503060303
29
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
30. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
30
▶ Scrambled :
• Concat LeftUpdate and RightUpdate .
• Example:
o CipherScrambled : 01010805030603030507070905010008
▶ RippledLeftToRight :
• Scrambled modifying by :
o CipherScrambled : 01010805030603030507070905010008
o 01 ⊕ 01 = (0 × 16) + 1 + (0 × 16) + 1 = 02 ≡ 02 (mod10).
o RippledLeftToRight = 0102
o RippledLeftToRight = 01020005080407000503090803040402
▶ RippledRightToLeft :
• Same idea
• RippledLeftToRight = 04030101060804070702000103000602
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study
31. | 08-09-2011 | Cauchie stéphane
Carte & Identification 2011
31
31
▶ Encryption and Modular Sum :
• RippledLeftToRight : 04030101060804070702000103000602
Survey on FPE
Tokenization and Format Preserving Encryption: A Case Study