Simultaneously Generating Secret and Private Keys in a Cooperative Pairwise-Independent Network

Peng Xu, Zhiguo Ding, Xuchu Dai, George K. Karagiannidis

Research output: Contribution to journalArticlepeer-review

18 Scopus citations

Abstract

This paper studies the problem of simultaneously generating a secret key (SK) and a private key (PK) between Alice and Bob, in a cooperative pairwise-independent network (PIN) with two relays. In the PIN, the pairwise source observed by every pair of terminals is independent of those sources observed by any other pairs. The SK needs to be protected from Eve, while the PK needs to be protected not only from Eve but also from the two relays. Two cooperative SK-PK generation algorithms are proposed: both of them first generate common randomness, based on the well-established pairwise key generation technique and the application of the one-time pad; but then, the two algorithms utilize the XOR operation and a specific random-binning-based SK-PK codebook to generate the expected keys, respectively. The achievable SK-PK rate regions of both the two proposed algorithms are analyzed. Of particular interest is the second algorithm with random-bing based codebook, whose achievable key rate region is demonstrated to be exactly the same as the derived outer bound, a crucial step for establishing the key capacity of this PIN model. Finally, the two proposed SK-PK generation algorithms are extended to a cooperative wireless network, where the correlated source observations are obtained from estimating wireless channels during a training phase.

Original languageBritish English
Article number7378467
Pages (from-to)1139-1150
Number of pages12
JournalIEEE Transactions on Information Forensics and Security
Volume11
Issue number6
DOIs
StatePublished - Jun 2016

Keywords

  • Cooperative PIN model
  • Information-theoretic security
  • key capacity region
  • private key
  • secret key

Fingerprint

Dive into the research topics of 'Simultaneously Generating Secret and Private Keys in a Cooperative Pairwise-Independent Network'. Together they form a unique fingerprint.

Cite this