Cryptography and Trust in Distributed Systems: A Review from Lightweight and Quantum-Safe Cryptography to Blockchain, Privacy-Preserving Learning and Moving-Target Defence

DOI

https://www.doi.org/10.59256/indjcst.20260503049

Published In

Volume 5, Issue 3 (2026)

Published Online

September-December 2026

Pages

424-432

Article Type

Review Article

Check for updates

Distributed systems now link sensors, phones, edge servers, clouds and blockchains, and each link depends on cryptography and on some model of whom to trust. That foundation is under pressure from two sides. Constrained devices cannot always run standard public-key algorithms, and a large quantum computer would break the RSA and elliptic curve schemes that most systems still use. At the same time, attacks increasingly target the learning models, consensus protocols and network paths that sit above the cipher. The problem is that these topics are reviewed separately: cryptography surveys rarely discuss blockchain consensus or federated learning, and defence surveys rarely discuss the cost of new cryptographic standards. This review brings them into one framework of five layers: lightweight and classical cryptography, quantum-safe cryptography, decentralized trust through blockchain and consensus, privacy-preserving distributed learning, and adaptive defence through detection and moving-target techniques. It organizes 47 studies and standards and compares their methods and reported results. Estimates of the quantum resources needed to factor a 2048-bit RSA key fell from about one billion physical qubits in 2012 to fewer than one million in 2025. NIST has standardized lattice-based and hash-based replacements, but their lattice-based keys and signatures are about 12 to 72 times larger than elliptic curve equivalents. Permissioned blockchains reach more than 3,500 transactions per second against about 7 for Bitcoin, and federated learning cuts communication rounds by 10 to 100 times but can leak training data through shared gradients unless secure aggregation or differential privacy is added. Moving-target defence changes the attack surface over time, yet it is rarely combined with quantum-safe or privacy-preserving layers. Future work should design and test full stacks that combine post-quantum cryptography, lightweight consensus, private learning and adaptive defence on constrained devices.