We propose a quantum -resistant key exchange protocol based on hard problems of lattices using MaTRU cryptosystem as an underlying scheme. A key exchange protocol based on NTRU cryptosystem given by Lie et al is not secure against man-in-the-middle (MITM) attack. To remove this failure and provide a secure protocol, our protocol uses a trusted third party (TTP). Additionally, our protocol is better than NTRU-KE on efficiency and security point of view. In this paper, we propose key exchange protocol with TTP and without TTP, and describe the advantages and disadvantages of both schemes. KEYWORDS closest vector problem, lattice based cryptosystems, MaTRU cryptosystem, NTRU cryptosystem, public key cryptosystems, shortest vectors problem
INTRODUCTIONKey exchange protocol is one of the important cryptographic techniques in exchanging secret keys. Without key exchanging, we can not go for secure communication over insecure channel using symmetric key cryptosystem. There may be manual key distributions for symmetric key cryptosystem. In 1976, Diffie and Hellman proposed a key exchange protocol (DH protocol) in their seminal paper "New Directions in Cryptography". 1 The DH-protocol is based on discrete logarithm problem (DLP). Since then, many key exchange protocols are proposed based on different algebraic structures. Most of them are the Diffie-Hellman type schemes. The Diffie-Hellman type means that the schemes are basically the same as the Diffie-Hellman scheme but defined over different groups. Some of the Diffie-Hellman type schemes are the ECDH protocol (elliptic curve-based DH protocol), 2,3 XTR DH protocol (DH protocol based on XTR cryptosystem 4 ), etc. But, P.W. Shor proposed an algorithm 5,6 in which he showed that if quantum computers become reality, then we can solve discrete logarithmic problems and can break all the Diffie-Hellman type key exchange protocols. In this way, we need a quantum-resistant key exchange protocol. To fulfill this requirement, several key exchange schemes 7-12 are proposed by distinguished researchers using learning with error (LWE) problem (introduced by Regev 13 ) and the ring LWE (RLWE) problem, 14 secure against quantum computers (LWE problem is as hard to solve as the worst case lattice problems). Since, hard problems of lattices still remain secure against quantum computers, 15 a lattice-based cryptosystem, NTRU cryptosystem, 16 which is also based on hard problems of lattices (shortest vector problem [SVP] and closest vector problem [CVP]), is considered to be secure in the world of quantum computers, 17 and a strong candidate of post quantum cryptography. Hence, to complete the ciphersuit of NTRU cryptosystem, someone should have to think about the key exchange protocol based on NTRU scheme, which was missing till 2013; and then, by motivating this fact, Lie et al 18 proposed a key exchange protocol based on the NTRU cryptosystem. Lie et al neither used any authentication technique in their key exchange algorithm nor any trusted third party (TTP). Because of this, their ...