Exact Security Analysis of Hash-then-Mask Type Probabilistic MAC - - PowerPoint PPT Presentation

exact security analysis of hash then mask type
SMART_READER_LITE
LIVE PREVIEW

Exact Security Analysis of Hash-then-Mask Type Probabilistic MAC - - PowerPoint PPT Presentation

Message Authentication Code HtM Construction Contributions Conclusion Exact Security Analysis of Hash-then-Mask Type Probabilistic MAC Constructions Avijit Dutta and Ashwin Jha and Mridul Nandi Indian Statistical Institute, Kolkata 27th


slide-1
SLIDE 1

Message Authentication Code HtM Construction Contributions Conclusion

Exact Security Analysis of Hash-then-Mask Type Probabilistic MAC Constructions

Avijit Dutta and Ashwin Jha and Mridul Nandi

Indian Statistical Institute, Kolkata

27th September, 2016

A.Dutta Exact Security Analysis of HtM Construction

slide-2
SLIDE 2

Message Authentication Code HtM Construction Contributions Conclusion

Outline of the talk

1 Message Authentication Code. 2 HtM Construction. 3 Contributions. 4 Conclusion A.Dutta Exact Security Analysis of HtM Construction

slide-3
SLIDE 3

Message Authentication Code HtM Construction Contributions Conclusion

MAC (Stateless and Deterministic): The Popular Story

1 Alice and Bob share a secret key K. A.Dutta Exact Security Analysis of HtM Construction

slide-4
SLIDE 4

Message Authentication Code HtM Construction Contributions Conclusion

MAC (Stateless and Deterministic): The Popular Story

1 Alice and Bob share a secret key K. 2 Alice sends a message M with a tag T = MACK(M)

corresponding to the message M to Bob.

A.Dutta Exact Security Analysis of HtM Construction

slide-5
SLIDE 5

Message Authentication Code HtM Construction Contributions Conclusion

MAC (Stateless and Deterministic): The Popular Story

1 Alice and Bob share a secret key K. 2 Alice sends a message M with a tag T = MACK(M)

corresponding to the message M to Bob.

3 Data Integrity: Bob verifies the sender and the message by

computing VERK(M, T) = 1.

A.Dutta Exact Security Analysis of HtM Construction

slide-6
SLIDE 6

Message Authentication Code HtM Construction Contributions Conclusion

MAC (Stateless and Deterministic): The Popular Story

1 Alice and Bob share a secret key K. 2 Alice sends a message M with a tag T = MACK(M)

corresponding to the message M to Bob.

3 Data Integrity: Bob verifies the sender and the message by

computing VERK(M, T) = 1. Unforgeability Adversary asks for tags for queries of his choice. Goal is to generate any fresh, valid (message, tag) pair. Security Requirement: It should be HARD

A.Dutta Exact Security Analysis of HtM Construction

slide-7
SLIDE 7

Message Authentication Code HtM Construction Contributions Conclusion

MAC (Stateful or Probabilistic): The Popular Story

  • Alice sends a message M, an auxiliary variable IV with a tag

T = MACK(M, IV ) corresponding to the message M and IV to Bob.

  • Data Integrity: Bob verifies the sender and the message by

computing VERK(M, IV , T) = 1. Stateful MAC : When IV is a counter / nonce. (e.g XMACC, PCS) Probabilistic MAC : When IV is random. (e.g XMACR, EHtM) Unforgeability Adversary asks for T for queries M (Signing Query). Adversary asks fresh (M, IV , T) triplet and obtains 1 or 0. Succeed if the response is 1 (Verification Query). Security: Should be HARD to obtain response 1

A.Dutta Exact Security Analysis of HtM Construction

slide-8
SLIDE 8

Message Authentication Code HtM Construction Contributions Conclusion

Pseudo Random Function (PRF)

PRF Keyed function which is indistinguishable from a Random Function (RF) Indistinguishability Responses of adversary queries are given either using the function or a RF. Goal is to distinguish the function from a RF. Security Requirement: It should be HARD

A.Dutta Exact Security Analysis of HtM Construction

slide-9
SLIDE 9

Message Authentication Code HtM Construction Contributions Conclusion

Universal and AXU-Hash

Universal Hash H is a n bit Universal Hash, if for all distinct values, the collision probability of H is negligible. Almost-XOR-Universal Hash H is a n bit AXU Hash, if for all distinct values x, x′ and for all y, Pr[H(x) ⊕ H(x′) = y] is negligible.

A.Dutta Exact Security Analysis of HtM Construction

slide-10
SLIDE 10

Message Authentication Code HtM Construction Contributions Conclusion

Existing Result on Probablistic MAC

Candidate Construction Rand Eff. Bound XMACR[BGR’95] (r, H(m) ⊕ f (r)) n 1Hxu, 1F[n, n] O( q2

2n + qvǫ)

MACRX3[BGK’99] (r1, r2, r3, 3n 1Hxu, 3F[n, n] O( q3

23n + qvǫ) 3

  • i=1

f (ri) ⊕ H(m)) RMAC[JJV’02] (r, f r

2 (CBCf1(m))

n (ℓ + 1)P[n] O( ℓ(q+qv )

2n

) FRMAC[JJ’04] (r, πr(H(m))) n 1Hu, 1P[n, n] O(ℓ(q + qv)ǫ) RWMAC[M’10] (r, g(r, H(m))) n 1Hu, 1F[2n, n] O( q2ǫ

2n + qvǫ)

EHtM[M’10] (r, f (r) ⊕ g(r ⊕ H(m)) n 1Hxu, 2F[n, n] O( q3ǫ

2n + qvǫ) A.Dutta Exact Security Analysis of HtM Construction

slide-11
SLIDE 11

Message Authentication Code HtM Construction Contributions Conclusion

HtM: Probabilistic MAC

r m t f – n – n –

n

– n

C1 : t := f(r) ⊕ m

r m t f g – n – n –

n

– n – n

C3 : t := f(r) ⊕ g(m)

r m t f g – n – n – n – n –

n

C5 : t := f(r) ⊕ g(r ⊕ m)

r m t f H – n – n –

n

– ℓ – n

C2 : t := f(r) ⊕ H(m)

r m t f H g – n – n – ℓ – n – n –

n

C4 : t := f(r) ⊕ g(H(m))

r m t f H g – n – ℓ – n – n – n –

n

C6 : t := f(r) ⊕ g(r ⊕ H(m))

A.Dutta Exact Security Analysis of HtM Construction

slide-12
SLIDE 12

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Our Contribution

  • 1. Tight PRF, pPRF and MAC Security Analysis of Different

Types of HtM Constructions.

  • 2. An Impossibility Result on Probabilistic MAC:

Unlike deterministic MAC, in probabilistic MAC, there is no such ideal system, indistinguishable to which, ensures forging advantage.

C1 C2 C3 C4 C5 C6 PRF X X X X X Θ(2n/2) pPRF Θ(2

n 2 )

Θ(2

n 2 )

Θ(2

n 2 )

Θ(2

n 2 )

Θ(2

3n 4 )

Θ(2

3n 4 )

MAC X Θ(2

n 2 )

Θ(2

n 2 )

Θ(2

n 2 )

Θ(2

2n 3 )

Θ(2

3n 4 )

A.Dutta Exact Security Analysis of HtM Construction

slide-13
SLIDE 13

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

PRF Attack Idea of C1,C2,C3,C4

(r1, y1) (r1, y2) (r2, y1) (r2, y2)

SUMf,g(r, y) = f (r) ⊕ g(y)

A.Dutta Exact Security Analysis of HtM Construction

slide-14
SLIDE 14

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

PRF Attack Idea of C1,C2,C3,C4

(r1, y1) (r1, y2) (r2, y1) (r2, y2)

SUMf,g(r, y) = f (r) ⊕ g(y) Alternating Cycle (Alt-Cycle)

  • For an Alt-Cycle C,

4

  • i=1

SUMC

f,g(ri, yi) = 0 (distinguishing event)

  • For C1, C2 : g is identity function.
  • For C1, C3 : y is m; For C2, C4 : y is H(m); For C5 : y is r + m

A.Dutta Exact Security Analysis of HtM Construction

slide-15
SLIDE 15

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

PRF Attack Idea of C5 and C6

Attack Algorithm C5 : f (r) ⊕ g(r ⊕ m)

  • Choose (r1, m1), (r2, m2) s.t r1 + m1 = r2 + m2
  • Query Phase :

t1 ← (r1, m1), t2 ← (r2, m2), t3 ← (r1, m2), t4 ← (r2, m1)

  • Distinguishing Event : If

4

  • i=1

ti = 0, return 1.

A.Dutta Exact Security Analysis of HtM Construction

slide-16
SLIDE 16

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

PRF Attack Idea of C5 and C6

Attack Algorithm C5 : f (r) ⊕ g(r ⊕ m)

  • Choose (r1, m1), (r2, m2) s.t r1 + m1 = r2 + m2
  • Query Phase :

t1 ← (r1, m1), t2 ← (r2, m2), t3 ← (r1, m2), t4 ← (r2, m1)

  • Distinguishing Event : If

4

  • i=1

ti = 0, return 1. Attack Algorithm C6 : f (r) ⊕ g(r ⊕ H(m))

  • Query Phase :

t1 ← (r, m1), t2 ← (r, m2), . . . , t2n/2 ← (r, m2n/2)

  • If H(mi) = H(mj), query t′

i ← (r′, mi), t′ j ← (r′, mj), output 1

if t′

i = t′ j

  • Else, collision in g.

A.Dutta Exact Security Analysis of HtM Construction

slide-17
SLIDE 17

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Probabilistic PRF (pPRF)

Definition and Security Game Keyed function that takes two inputs (r, M) is indistinguishable from RF Adversary can only query the oracle with M. Goal is to distinguish the function from a RF; secure if it is hard

A.Dutta Exact Security Analysis of HtM Construction

slide-18
SLIDE 18

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Probabilistic PRF (pPRF)

Definition and Security Game Keyed function that takes two inputs (r, M) is indistinguishable from RF Adversary can only query the oracle with M. Goal is to distinguish the function from a RF; secure if it is hard pPRF Attack Algorithm of C1 : f (r) ⊕ m

  • Query Phase : t1 ← m1, t2 ← m1, . . . , t2n/2 ← m1
  • W.h.p ∃i, j ∈ {1, 2, . . . , 2n/2} s.t ri = rj
  • If ti = tj, return 1.

A.Dutta Exact Security Analysis of HtM Construction

slide-19
SLIDE 19

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Probabilistic PRF (pPRF)

Definition and Security Game Keyed function that takes two inputs (r, M) is indistinguishable from RF Adversary can only query the oracle with M. Goal is to distinguish the function from a RF; secure if it is hard pPRF Attack Algorithm of C1 : f (r) ⊕ m

  • Query Phase : t1 ← m1, t2 ← m1, . . . , t2n/2 ← m1
  • W.h.p ∃i, j ∈ {1, 2, . . . , 2n/2} s.t ri = rj
  • If ti = tj, return 1.

pPRF Attack for C2, C3, C4 is same as that of C1

A.Dutta Exact Security Analysis of HtM Construction

slide-20
SLIDE 20

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Attack Idea of C5

r m t f g – n – n – n – n –

n

C5 : t := f(r) ⊕ g(r ⊕ m)

m0 m1 (m0 ⊕ m1 = δ)

. . . . . .

ri rj rk rl, ri ⊕ rk = δ ti = f(ri) ⊕ g(m0 ⊕ ri) tj = f(rj) ⊕ g(m1 ⊕ rj) tk = f(rk) ⊕ g(m0 ⊕ rk) tl = f(rl) ⊕ g(m1 ⊕ rl)

Figure 0.1: Distinguishing Event : If ti ⊕ tj ⊕ tk ⊕ tl = 0, output 1.

A.Dutta Exact Security Analysis of HtM Construction

slide-21
SLIDE 21

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Attack Idea of C6

r m t f H g – n – ℓ – n – n – n –

n

C6 : t := f(r) ⊕ g(r ⊕ H(m))

m0 m1 (h(m0) ⊕ h(m1) = δ)

. . . . . .

ri rj rk rl, ri ⊕ rk = δ ti = f(ri) ⊕ g(ri ⊕ h(m0)) tj = f(rj) ⊕ g(rj ⊕ h(m1))) tk = f(rk) ⊕ g(rk ⊕ h(mk)) tl = f(rl) ⊕ g(rl ⊕ h(ml))

Figure 0.1: Distinguishing Event : If ti ⊕ tj ⊕ tk ⊕ tl = 0, output 1. A.Dutta Exact Security Analysis of HtM Construction

slide-22
SLIDE 22

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Forging Idea of C1-C6

Forging C1 : f (r) ⊕ m

  • Query Phase : t ← (r, m).
  • Forge : (r, m ⊕ δ, t ⊕ δ).

A.Dutta Exact Security Analysis of HtM Construction

slide-23
SLIDE 23

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Forging Idea of C1-C6

Forging C1 : f (r) ⊕ m

  • Query Phase : t ← (r, m).
  • Forge : (r, m ⊕ δ, t ⊕ δ).

Forging C2 : f (r) ⊕ H(m)

  • Query Phase :

t1 ← (r1, m1), t2 ← (r2, m2), . . . , t2n/2 ← (r2n/2, m2n/2).

  • W.h.p i, j ∈ {1, 2, . . . , 2n/2} such that ri = rj. It leaks

H(mi) ⊕ H(mj) = δ.

  • Query t ← (r, mi).
  • Forge : (r, mj, t ⊕ δ).

A.Dutta Exact Security Analysis of HtM Construction

slide-24
SLIDE 24

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Forging Idea of C1-C6

Forging C1 : f (r) ⊕ m

  • Query Phase : t ← (r, m).
  • Forge : (r, m ⊕ δ, t ⊕ δ).

Forging C2 : f (r) ⊕ H(m)

  • Query Phase :

t1 ← (r1, m1), t2 ← (r2, m2), . . . , t2n/2 ← (r2n/2, m2n/2).

  • W.h.p i, j ∈ {1, 2, . . . , 2n/2} such that ri = rj. It leaks

H(mi) ⊕ H(mj) = δ.

  • Query t ← (r, mi).
  • Forge : (r, mj, t ⊕ δ).

Forging attack of C3, C4 is same as that of C2

A.Dutta Exact Security Analysis of HtM Construction

slide-25
SLIDE 25

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Forging Idea of C5

r m t f g – n – n – n – n –

n

C5 : t := f(r) ⊕ g(r ⊕ m)

m0 m1 (m0 ⊕ m1 = δ)

. . . . . .

ri rj rk ri ⊕ rk = δ ti = f(ri) ⊕ g(m0 ⊕ ri) tj = f(rj) ⊕ g(m1 ⊕ rj) tk = f(rk) ⊕ g(m0 ⊕ rk) Forge: (rk, m1, ti ⊕ tj ⊕ tk) (∵ ti ⊕ tj ⊕ tk = f(rk) ⊕ g(m1 ⊕ rk)) A.Dutta Exact Security Analysis of HtM Construction

slide-26
SLIDE 26

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Forging Idea of C6

r m t f H g – n – ℓ – n – n – n –

n

C6 : t := f(r) ⊕ g(r ⊕ H(m))

m0 m1 (h(m0) ⊕ h(m1) = δ)

. . . . . .

ri rj rk rl, ri ⊕ rk = δ ti = f(ri) ⊕ g(h(m0) ⊕ ri) tj = f(rj) ⊕ g(h(m1) ⊕ rj) tk = f(rk) ⊕ g(h(m0) ⊕ rk) Forge: (rk, m1, ti ⊕ tj ⊕ tk) (∵ ti ⊕ tj ⊕ tk = f(rk) ⊕ g(h(m1) ⊕ rk)) A.Dutta Exact Security Analysis of HtM Construction

slide-27
SLIDE 27

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Alternating Cycle

A transcript τ := {(x1, y1), (x2, y2), . . . , (xq, yq)} has an alternating-cycle in τ of length k (k is even and ≥ 2), if we have k pairwise distinct indices i1, i2, . . . , ik such that xi1 = xi2, yi2 = yi3, xi3 = xi4, . . . , xik1 = xik, yik = yi1.

A.Dutta Exact Security Analysis of HtM Construction

slide-28
SLIDE 28

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Alternating Cycle

A transcript τ := {(x1, y1), (x2, y2), . . . , (xq, yq)} has an alternating-cycle in τ of length k (k is even and ≥ 2), if we have k pairwise distinct indices i1, i2, . . . , ik such that xi1 = xi2, yi2 = yi3, xi3 = xi4, . . . , xik1 = xik, yik = yi1.

(x1, y1) (x2, y2) (x3, y3) (x4, y4)

Figure: Alternating Cycle of length 4. Red line indicates first coordinate

  • matches. Green line indicates second coordinates matches

A.Dutta Exact Security Analysis of HtM Construction

slide-29
SLIDE 29

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Benes Butterfly Result

Theorem (Benes-Butterfly (AV’96)) Let f and g be two n-bit independent and uniformly distributed random functions. Let us consider a transcript τ = {(xi, yi, ti)1≤i≤q} which does not contain any alternating

  • cycle. Then

Pr[f (xi) ⊕ g(yi) = ti, 1 ≤ i ≤ q] = 1 2nq . Proof Sketch : If there is no alternating cycle in τ = {(xi, yi)1≤i≤q} then from each of q many equations, we get at least one uniform random variable

A.Dutta Exact Security Analysis of HtM Construction

slide-30
SLIDE 30

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Advantage of C5 and C6

Theorem Advpprf

C5/C6(q, t) ≤ Advprf fk1 (q, t) + Advprf fk2 (q, t) + q4 23n .

A.Dutta Exact Security Analysis of HtM Construction

slide-31
SLIDE 31

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Advantage of C5 and C6

Theorem Advpprf

C5/C6(q, t) ≤ Advprf fk1 (q, t) + Advprf fk2 (q, t) + q4 23n .

  • Bad Transcript : Alternating cycle on (r, r ⊕ m)/(r, r ⊕ h(m)).

A.Dutta Exact Security Analysis of HtM Construction

slide-32
SLIDE 32

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Advantage of C5 and C6

Theorem Advpprf

C5/C6(q, t) ≤ Advprf fk1 (q, t) + Advprf fk2 (q, t) + q4 23n .

  • Bad Transcript : Alternating cycle on (r, r ⊕ m)/(r, r ⊕ h(m)).
  • No bad event ⇒ No alternating cycle in the transcript.

A.Dutta Exact Security Analysis of HtM Construction

slide-33
SLIDE 33

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

pPRF Advantage of C5 and C6

Theorem Advpprf

C5/C6(q, t) ≤ Advprf fk1 (q, t) + Advprf fk2 (q, t) + q4 23n .

  • Bad Transcript : Alternating cycle on (r, r ⊕ m)/(r, r ⊕ h(m)).
  • No bad event ⇒ No alternating cycle in the transcript.
  • Probability of bad event :

q4 23n

A.Dutta Exact Security Analysis of HtM Construction

slide-34
SLIDE 34

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

SUF Advantage of C5 and C6

Theorem (SUF Advantage of C5) Advsuf

C5 (q, q′, t) ≤ Advprf fk1 (q + q′, t) + Advprf fk2 (q + q′, t) + q3 22n + q′ 2n .

A.Dutta Exact Security Analysis of HtM Construction

slide-35
SLIDE 35

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

SUF Advantage of C5 and C6

Theorem (SUF Advantage of C5) Advsuf

C5 (q, q′, t) ≤ Advprf fk1 (q + q′, t) + Advprf fk2 (q + q′, t) + q3 22n + q′ 2n .

  • Bad Transcript : Alternating cycle on (r, r ⊕ m) after making

signing and verfication queries.

A.Dutta Exact Security Analysis of HtM Construction

slide-36
SLIDE 36

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

SUF Advantage of C5 and C6

Theorem (SUF Advantage of C5) Advsuf

C5 (q, q′, t) ≤ Advprf fk1 (q + q′, t) + Advprf fk2 (q + q′, t) + q3 22n + q′ 2n .

  • Bad Transcript : Alternating cycle on (r, r ⊕ m) after making

signing and verfication queries.

  • Good Transcript ⇒ No Alternating Cycle in the transcript.
  • Probability of Bad Transcript :

q3 22n .

A.Dutta Exact Security Analysis of HtM Construction

slide-37
SLIDE 37

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Proof Idea of SUF Advantage of C5 and C6

Theorem (SUF Advantage of C6) Advsuf

C6(q, q′, ℓ, t) ≤ Advprf fk1 (q+q′, t′)+Advprf fk2 (q+q′, t′)+ q4 23n + 10q′ 2n ,

where t = t′ + O(qTh)

A.Dutta Exact Security Analysis of HtM Construction

slide-38
SLIDE 38

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Proof Idea of SUF Advantage of C5 and C6

Theorem (SUF Advantage of C6) Advsuf

C6(q, q′, ℓ, t) ≤ Advprf fk1 (q+q′, t′)+Advprf fk2 (q+q′, t′)+ q4 23n + 10q′ 2n ,

where t = t′ + O(qTh)

  • Bad Transcript : Alternating cycle on (r, r ⊕ h(m)) after making

signing and verfication queries.

A.Dutta Exact Security Analysis of HtM Construction

slide-39
SLIDE 39

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Proof Idea of SUF Advantage of C5 and C6

Theorem (SUF Advantage of C6) Advsuf

C6(q, q′, ℓ, t) ≤ Advprf fk1 (q+q′, t′)+Advprf fk2 (q+q′, t′)+ q4 23n + 10q′ 2n ,

where t = t′ + O(qTh)

  • Bad Transcript : Alternating cycle on (r, r ⊕ h(m)) after making

signing and verfication queries.

  • Good Transcript ⇒ No Alternating cycle.

A.Dutta Exact Security Analysis of HtM Construction

slide-40
SLIDE 40

Message Authentication Code HtM Construction Contributions Conclusion Attack Idea Proof Idea

Proof Idea of SUF Advantage of C5 and C6

Theorem (SUF Advantage of C6) Advsuf

C6(q, q′, ℓ, t) ≤ Advprf fk1 (q+q′, t′)+Advprf fk2 (q+q′, t′)+ q4 23n + 10q′ 2n ,

where t = t′ + O(qTh)

  • Bad Transcript : Alternating cycle on (r, r ⊕ h(m)) after making

signing and verfication queries.

  • Good Transcript ⇒ No Alternating cycle.
  • Probability of Bad Transcript :

q4 23n as (we need one more point)

A.Dutta Exact Security Analysis of HtM Construction

slide-41
SLIDE 41

Message Authentication Code HtM Construction Contributions Conclusion

Summary

  • Tight Security Analysis of HtM Probabilistic MAC.
  • Tight Security Analysis of EHtM.
  • Impossibility result on Probabilistic MAC.

A.Dutta Exact Security Analysis of HtM Construction

slide-42
SLIDE 42

Message Authentication Code HtM Construction Contributions Conclusion

Summary

  • Tight Security Analysis of HtM Probabilistic MAC.
  • Tight Security Analysis of EHtM.
  • Impossibility result on Probabilistic MAC.

Thank You

A.Dutta Exact Security Analysis of HtM Construction