1. New Bounds for PMAC, TMAC, and XCBC Kazuhiko Minematsu and Toshiyasu Matsushima, NEC Corp. and Waseda University Fast Software Encryption 2007, March 26-28, Luxembourg City, Luxembourg

2. 1 Introduction  Message authentication code (MAC) from block ciphers (BCs)  “ BC-only ” modes: no special function other than a block cipher Ex. Encrypted CBC-MAC (EMAC)

3. 2 Security notion of MACs  Advantage in distinguishing MAC from the (keyed) random oracle (RO),, using CPA  Small advantage implies small MAC forgery prob. Note: We only consider the info-theoretic security, but our results have simple computational counterparts : number of queries : max. message length (in n-bit) : total number of queried blocks can contain (but not vice versa)

4. 3 Related works on EMAC  Previous EMAC security bound is:  when it is implemented w/ two n-bit uniform random permutations (URPs), and EMAC w/ two URPs [BR00] room for improvement?

5. 4 Related works on EMAC (contd.)  Bellare, Pietrzak, and Rogaway [BPR05] is a function that grows very slowly with Note: Pietrzak [P06] obtained a tighter bound for a range of parameters (much smaller than )  If, the bound is roughly

6. 5 Our contribution  New security bounds for  PMAC (a parallelizable MAC)  TMAC and XCBC (successors of EMAC)  Old: or  New: for PMAC, and for TMAC & XCBC  compared w/, from quadratic to (almost) linear degradation wrt  compared w/, better in most (but not all) cases

7. 6 Analysis of PMAC

8. 7 PMAC (Black-Rogaway[BR02], Rogaway[R04])  Hashing with mask-encrypt-sum (PHASH)  still BC-only: masks are generated w/ few bitshifts and XORs PMAC ([R04] version w/ 128 bit block size) PHASH input

9. 8 Overview of old proof [R04]  “ Perfect ” PMAC using independent URPs as an intermediate function  Use triangle inequality Perfect PMAC PMAC RO  Old bound: (also, as )

10. 9 Overview of new proof  A different intermediate function, the modified PMAC (MPMAC)  PHASH + independent finalization MPMACPMAC RO

11. 10 MPMAC vs. Random Oracle  What we need is: (a stronger form of ) differential probability of PHASH... used for MPMAC vs. RO used for PMAC vs. MPMAC...

12. 11 Diff. probability of PHASH  A subset of input blocks may generate the same URP input  Odd (Even) collision involves odd (even) number of input blocks  Let denote odd collisions with non- zero URP inputs  Then, c ritical event is, as it implies the sum = 0 or w/ prob. 1 (as )... even collisionodd collision...

13. 12 Diff. probability of PHASH (contd.)  is at most  Given, PHASH sum is almost uniform (point probability is at most ) for any Lemma 2  From Lemma 2, the advantage between MPMAC and RO is:

14. 13 PMAC vs. MPMAC  Four “ good ” events defined as: the sets of URP inputs in PHASH and in the finalization (+ dummy mask for MPMAC) have no intersection  Using Maurer ’ s method [M02], the advantage is at most the max. prob. of “ bad ” events in MPMAC, denoted by

15. 14 New bound for PMAC  A careful analysis using Lemma 2 provides if MPMACPMACRO Theorem 2

16. 15  As long as there is a small (but not too small) fraction of long messages, the new bound is better  Much better under some practical cases (e.g., all messages have similar lengths) Comparison of new and old bounds  New ( ) < old ( ) iff  Ex: New bound is 2 -32, old bound is 2 -48 ~2 -16 If 99.9% messages are one-block, old bound is better If at least 1% messages are -block, new bound is better (if we ignore constants)

17. 16 Analysis of TMAC and XCBC

18. 17 TMAC [KI03] and XCBC [BR00]  Successors of EMAC  fewer BC calls (no double encryption)  one BC key + one or two n-bit keys is independent of TMAC

19. 18 Proof sketch for TMAC (XCBC is the same)  Modified TMAC (MTMAC) and bad events similar to those for PMAC  Adv. between TMAC and MTMAC is  much simpler analysis due to the independence of  Adv. between MTMAC and RO is EMAC bound of [BPR05], i.e.,

20. 19 New bounds for TMAC and XCBC  Old bounds are or for  TMAC ’ s new bound is: Theorem 3 (XCBC ’ s bound is the same) [BR00][KI03][IK03s]  Bound comparison is almost the same as PMAC ’ s case, in case the second term is negligible

21. 20 Short comments on OMAC [IK03o]  OMAC (aka CMAC) is one-key CBC-MAC  improvement to TMAC and XCBC mask is or, where  MOMAC and bad events are similarly defined  however, the probabilities of some new bad events have to be evaluated such as an extension of CBC collision analysis [BPR05] is needed (open problem)

22. 21 Conclusion  New bounds for PMAC, TMAC, and XCBC  from quadratic to (almost) linear degradation wrt the max. message length  Future directions  OMAC  further improvement (still far from the lower bound )

23. 22 Thank you!