Volume 13 - Issue 4
Efficient Lattice based (H)IB-DRE
- Kunwar Singh
National Institute of Technology, Tiruchirappalli, TamilNadu, India
- C. Pandu Rangan
Indian Institute of Technology, Chennai, TamilNadu, India
- Ilsun You
Kookmin University, Seoul, Republic of Korea
isyou@kookmin.ac.kr
- Amalan Joseph Antony
National Institute of Technology, Tiruchirappalli, TamilNadu, India
- SK Karthika
National Institute of Technology, Tiruchirappalli, TamilNadu, India
- Jiyoon Kim
Gyeongsang National University, Jinju, Gyeongsangnam-do, Republic of Korea
jykim92@gnu.ac.kr
Keywords: Lattice, identity based cryptosystem, Dual receiver encryption
Abstract
In CCS’04, Diament etc al. presented special kind of public key encryption called Dual Receiver
Encryption (DRE). In DRE, sender encrypts the plaintext using public keys of two independent receivers.
Both independent receivers can decrypt the ciphertext into the same plaintext using their
own private keys. This new cryptography primitive has applications in construction of complete nonmalleability
and plaintext-awareness public key encryption, for a secure management of data that
is to be disseminated to distributed processors, for ubiquitous and mobile computing applications.
Daode Zhang et. al. constructed lattice-based IB-DRE scheme which is secure against stronger security
notion i.e. adaptive-ID. We present hierarchical identity based dual encryption (HIB-DRE)
scheme under LWE assumption. To the best of our knowledge, this gives the first provably secure
HIB-DRE scheme in the lattice based setting. Independent work by Naccache[1] and Chatterjee -
Sarkar[2] presented a variant of Waters’s identity based encryption scheme to reduce Public Parameters.
They have considered an identity of l-bits as l0 chunks where size of each chunk is l=l0. This
reduces the Public Parameter (PP) size from l to l0 nm matrices. This idea was named as blocking
technique[3]. We have used blocking technique to reduce the size of PP. Daode Zhang et. al.[4]
presented adaptive secure IB-DRE scheme. This scheme[4] contains l +1, nm matrices as PPs,
where l denotes the number of the bits in identity. Using blocking technique we have reduced the
size of PP by around factor b. Because of this the size of prime number q in field Zq is increased
by 2b which results in increase in computation cost. We have shown that compared to Daode Zhang
et. al.scheme the size of the PP can be decreased approximately by 80% and the time complexity is
increased by only 1:40% for a suitably chosen b.