Anonymity-preserving Public-Key Encryption Markulf Kohlweiss Ueli Maurer, Cristina Onete, Björn Tackmann, and Daniele Venturi PETS 2013 Context: Encryption and Anonymity Public-key encryption Short but eventful history, late 70s, 80s. Security usually defined using Games: IND-CPA, IND-CCA, … Anonymity Shorter eventful history, early 90s. Anonymity is arguably a more high-level property What if used together? Key privacy, robust encryption, formal analysis of onions Games prone to require iterations to find “right” notion PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 2 What is Anonymous Encryption? [PH08] Sender Anonymity Receiver Anonymity Anonymity not created, but preserved PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 3 Our contribution PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 4 Chosen Ciphertext Attack Security (IND-CCA) Dec c Dec(c) Bit b m0, m1 Enc(mb) bit d Challenger pk d = b? PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 5 Key Privacy (IK-CCA) [BBDP01] Dec0 Dec1 c Dec0 (c) c Dec (c) 1 Bit b m Enc(pkb; m) bit d Challenger pk0, pk1 d = b? PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 6 Weak Robustness (WROB) [ABN10] Dec m, i, j Challenger pk1, ..., pkn c Enc(pki, m) ≠ Dec(skj, c) ? ┴ PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 7 Constructive Cryptography [MR11] Resources (existing/assumed, desired): Available to everyone, including adversary/simulator through interfaces Converters: Transform existing into desired resources Two interfaces, inner and outer Protocol: composition of many converters, one for each user Security: Correctness: without Eve the protocol works correctly Security: when Simulator connected, no-one can distinguish between assumed and desired worlds. PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 8 Confidential Receiver-Anonymous Channel PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 9 PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 10 Constructing the Channel from Broadcast Existing Resources (pki) n x B1 m m m B2 m ┴ … Bn PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 11 Constructing the Channel from Broadcast Existing Resources n x (pki) Encryption scheme that is: B1 m m* … Bj m Converters m*, j … Bn PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 12 IND-CCA IK-CCA WROB Simulation (intuition) Existing world … (c, i) … … c Desired world B1 … Bj (m, i) Bi |m| Bn c m, i c Key-Generation: generate n keypairs (for each Bi), one separate (sk, pk) Ciphertext generation: get |m|, encrypt 0|m| under pk to get c PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 13 … Bj Bi Bn m, i D … B1 Simulation (intuition) Existing world … c* (c, i) (c*, j) … c … Bj Bi |m| Bn (c*, j) m* (m, i) Dec(c*) … Desired world B1 D … (m*, j) … B1 Bj Bi Bn (c*, j) m* Ciphertext delivery: deliver c* to Bj: • if c* not seen before decrypt under skj and inject message m* into network PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 14 Simulation (intuition) Assumed world … (c, i) c … c Desired world B1 … If i = i* Bj (m, i) m Bi Bn (c, i*) D m=Dec(c) (c, i*) … Trial Delivery |m| (H, i*) … … Bj Bi Bn (c, i*) H <-> m m Ciphertext delivery: deliver c to Bj: • if c seen before deliver corresponding msg. to correct receiver Intuition: this is where we need WROB – wrong receiver outputs error PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 15 B1 (More) Results in a Nutshell WROB sufficient SROB leads to a tighter reduction WROB necessary without WROB, achieve anonymity with erroneous transmission Impossibility: SROB does not construct better resource Constructive aspects: Model network with single sender, many receivers PK settings: use uni-directional authenticated channels Trial deliveries prevent better anonymity PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 16 Results in Picture Game-based analysis PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 17 Constructive result Strong Robustness (SROB) Dec c, i, j pk1, ..., pkn PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 18 Challenger both ┴ ≠ Dec(ski, c) ┴ ≠ Dec(skj, c) PETS 2013 | Markulf Kohlweiss | Anonymity-preserving PKE | Slide 19
© Copyright 2026 Paperzz