5G will deliver a wider broadband access, higher and faster mobility, unprecedented connectivity and coverage to massive matrix of personal devices, mobility based (vehicular/ commercial UAVs/ Drones) and Internet of Things / Everything (IoT / IoE) nodes with ensured availability and optimum reliability. Availability is targeted through multipronged Active/Passive attacks using DOS / DDOS (RF jamming / Bandwidth jamming with malformed packets) and reliability is targeted through exploiting existing architectural weaknesses of radio and data communication protocols (LTE-A / IP based Networking). In prevailing threat landscape, concepts of adoption and integration of Software Defined Networking (SDN) and Network Function Virtualization (NFV) in 5G networks will open gateways of new networking paradigms with a focus on centralized security dynamics (overall visibility and substantial controls). These network security dynamics in diversified use cases (Internet of Everything /smart grids) will provide an optimum centralized control with a global overview, vigilant monitoring as well as prompt reaction (detection and prevention) to these prevailing threats. In this purview of the 5G cellular security matrix with a centralized SDN and NFV, this document has reviewed on 5G cellular secure communication channels to unearth the effectiveness of conventional DOS/DDoS and Authentication attacks within environment of SDN and NFV. This research will also present proposed mitigation of DOS / DDOS and incorporation of advanced authentication mechanism for more reliable cellular communication networks.

Advanced Persistent Threats commonly use Domain Generation Algorithms to evade advanced detection methods to establish communication with their command and control servers. To overcome the DNS protocol legitimacy breach, several Domain Generation Algorithms detection methods have been proposed. To solve the problem of DGA malware detection amicably, Deep Learning based detection schemes attracted researchers’ interest recently. However, Deep Learning has already achieved optimal results and the gap is identified as fine-tuning of Deep Learning model hyper-parameters. The proposed solution is focusing on a model specific hyper-parameter known as the gradient optimizer. Gradient Optimisers are broadly categorised into Stochastic Gradient and Adaptive Moment based Gradient. Moment-based Gradient optimizer approaches are identified with suffering from weight decay and leading to poor generalization. Adaptive Moment (Adam) has improved with weight Decay as AdamW. To optimise moment-based gradient optimizers, Adam and AdamW are analyzed deeply. To optimize the functioning of AdamW, we present AdamW+, a novel solution for detecting DGA algorithms through re-implementing and nullifying the weight decay in AdamW. AdamW+ has been successfully implemented and shown promising results compared to Adam and AdamW optimizers in practice. AdamW+ preserved the properties of Adaptive Optimizer Adam while simplifying the weight decay implementation of AdamW. Empirical analysis has proved that AdamW+ has outperformed Adam and AdamW. The experimental result have substantiated that the proposed algorithm achieves the best accuracy result.

Centralized cloud platforms simplify application development and deployment, but they can also introduce vendor lock-in, recurring costs, privacy concerns, and dependence on third-party infrastructure. This paper presents dweb OS, a local-first peer-to-peer platform that transforms a personal machine into a self-hosted cloud node. Its three-layer architec- ture combines an Alpine Linux foundation; a HyperDHT and WebRTC networking layer for peer discovery, .dweb name resolution, direct connectivity, and relay fallback; and a multi- provider AI build engine supporting both local inference through Ollama and cloud-hosted models. A browser-based portal integrates service management, Git workflows, domain pub- lication, code generation, and deployment. In a single test environment, local cache resolution averaged 0.3 ms, while same-machine DHT, LAN WebRTC, and relay-assisted WAN resolution averaged 1.2 s, 2.8 s, and 4.5 s, respectively. The AI evaluation used five full-stack prompts executed five times per model and demonstrated that a locally hosted 7B coding model could generate runnable projects, with a trade-off in latency and success rate relative to the evaluated cloud models. Resource mea- surements showed a peak observed footprint of 280 MB of memory and 18% CPU utilization when all measured services were active. These results establish dweb OS as a functional pro- totype for sovereign development, private hosting, and decentralized application deployment, while identifying broader multi-network evaluation, stronger sandboxing, and persistent de- centralized availability as priorities for future work.