An Enhanced Distribution Transforming Encoder (Dte) Of The Honey Encryption Scheme For Reinforcing Text-Based Encryption
Honey Encryption (HE) is a cryptosystem used as a reinforcement to the conventional encryption scheme to address brute-force attacks specifically in the context of password-based encryption systems. The HE scheme relies on a model called the Distribution Transforming Encoder (DTE), which focuses on...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/54016/1/ABIODUN%20ESTHER%20OMOLARA%27S%20THESIS.pdf http://eprints.usm.my/54016/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Honey Encryption (HE) is a cryptosystem used as a reinforcement to the conventional encryption scheme to address brute-force attacks specifically in the context of password-based encryption systems. The HE scheme relies on a model called the Distribution Transforming Encoder (DTE), which focuses on the use of deception as a key defensive approach in the design of primitives that facilitate information security by yielding plausible-looking but fake plaintext during decryption using an incorrect key. However, the concept of the HE scheme is limited by the bottleneck of applicability and thus fails to reach other real-world deployment use-cases. For instance, encoding a human-generated message such as email requires adapting the scheme to natural language which means re-designing the current deterministic encoder to generate plausible or realistic decoys message that can fool the attacker. This problem remains unsolved because of the few researches on enhancing the DTE fails to produce plausible decoy messages in human-language. Furthermore, they fail to introduce secrecy on the fake message as keywords or fragments of the underlying plaintext message are revealed during decryption, thus, enabling the system to a chosen-ciphertext attack where an attacker may use the results from prior decryption to inform their choices of which ciphertexts have decrypted. The two main contributions of this work are its responses to these two problems. A natural language-based encoder (NLBE) was developed and an approach for concealing the underlying plaintext and producing plausible decoy messages is presented. Experimental analysis using human simulators as the gold standard shows a 94% indistinguishability rate in the worst case where an unbounded adversary can explore the complete Oracle and a 100% indistinguishability rate when the keyspace is large enough. |
|---|