The rapid convergence of Internet of Things (IoT), edge computing, cloud services, industrial control systems, mobile devices, and intelligent applications has created heterogeneous ICT environments in which security, trust, interoperability, data integrity, communication efficiency, and resource utilization are tightly coupled. Conventional centralized security and analytics architectures are increasingly challenged by distributed attack surfaces, single points of failure, privacy constraints, non-uniform device capabilities, and high communication overhead. This paper surveys blockchain, artificial intelligence (AI), deep learning (DL), federated learning (FL), edge intelligence, peer-to-peer (P2P) communication, and intelligent resource optimization for heterogeneous ICT networks. The review is organized around four core capabilities: infrastructure protection, verifiable data integrity and trust, decentralized P2P communication, and adaptive resource optimization. The analysis identifies a persistent integration gap: most existing studies address one or two capabilities, whereas comparatively few provide an end-to-end architecture that jointly coordinates all four. To address this gap, a five-layer blockchain-enabled AI/DL framework is proposed, integrating heterogeneous infrastructure, secure P2P data exchange, a blockchain trust layer, AI/DL analytics, and security-aware resource optimization. The framework is intended as a research foundation for future empirical validation in realistic edge-cloud and cyber-physical environments.