Key-based interval splitting arithmetic coding (KSAC) has been proposed to improve the security of traditional arithmetic coding (AC). Chosen-plaintext attacks have been proposed for KSAC when the same key is used to encrypt different messages. In this paper, we consider a stronger version of KSAC, where different keys are used to encrypt different messages. We then use message indistinguishability to prove that this version of KSAC is insecure under ciphertext-only attacks, a weaker form of attack than chosen-plaintext attacks. Indistinguishability in the presence of an eavesdropper is a security definition equivalent to semantic security. We prove the insecurity over the alphabet {A,B} with pB=(1/2(1+2s)) and pA=1-pB where pA and pB are the probabilities of the source generating A and B, respectively, and s is the number of bits in each splitting key.
展开▼
机译：已经提出了基于密钥的间隔分裂算术编码（KSAC）以提高传统算术编码（AC）的安全性。当使用相同的密钥来加密不同的消息时，已经为KSAC提出了选择明文攻击。在本文中，我们考虑了更强大的KSAC版本，其中使用了不同的密钥来加密不同的消息。然后，我们使用消息不可区分性来证明此版本的KSAC在仅密文攻击下是不安全的，这是一种比选择明文攻击更弱的攻击形式。在存在窃听者的情况下，不可区分性是等同于语义安全性的安全性定义。我们证明了具有p B =（1/2（1 + 2s））和p A = 1-p 的字母{A，B}的不安全性B ，其中p A 和p B 分别是源生成A和B的源的概率，而s是每个拆分密钥中的位数。
展开▼
