Current Journal of Applied Science and Technology

Blockchain systems are expected to keep operating correctly under faults and attacks, to protect information about their users and to accommodate growing demand. These expectations, termed here resilience, privacy and scalability, are seldom examined together. The widely cited scalability trilemma frames design as a contest among decentralisation, security and scalability, leaves privacy implicit and folds resilience into a broad notion of security. This structured critical narrative review examines whether the three properties have been treated jointly or independently, what evidence supports claims about their interaction, and where independent treatment has produced unanticipated weaknesses. Peer-reviewed literature published between January 2008 and September 2026 was identified through multidisciplinary scholarly indexes, discipline-specific repositories and citation tracing, and was appraised qualitatively for the explicitness of threat models, the reporting of effects on non-target properties, the realism of evaluation and reproducibility. The review proposes a taxonomy that separates five treatment modes (independent, constrained, co-designed, composed and post hoc audited) from four interaction types (synergy, antagonism, conditional interaction and displacement of risk to another layer or actor). The coupling between resilience and scalability is the most rigorously formalised, particularly for proof-of-work parameters, Byzantine fault-tolerant committees and sharding under adaptive adversaries. The coupling between privacy and scalability is well characterised at the level of cryptographic cost but only weakly at the level of deployed systems. Joint treatment of all three properties remains rare, and much of the clearest cross-property evidence comes from attack and measurement studies of payment-channel networks and peer-to-peer layers, which show that mechanisms introduced for one property shifted risk onto another. Evaluation practice is still dominated by fault-free throughput benchmarks, inconsistent privacy metrics and threat models that differ from one property to the next. Joint treatment appears most necessary where mechanisms share a resource or trust assumption, such as validator committees, trusted hardware, transaction sequencers, routing topology and network metadata. The evidence supports a conditional, layer-specific account of trade-offs rather than a universal trilemma, and progress will depend on common adversary models, failure-inclusive benchmarks and composable definitions that span all three properties.