5G Core Network Optimization

Abstract

The invention relates to a system that enables automatic horizontal and vertical scaling of virtual network functions (VNF) and container network functions (CNF) of 5G Core Network (5GC). The invention includes 5GC network functions (1) where 5G Core network functions AMF, SMF, UPF, PCF, UDM, AUSF are deployed as virtual / container functions; Kubernetes orchestration (2) which performs pod lifecycle management, service discovery, and load balancing; cloud infrastructure (3) which provides essential computing resources such as virtual machines, containers, storage, and networking; a network function metric collector (4) which collects metrics such as CPU usage, memory usage, network I / O, operations per second (TPS), number of active subscribers / session, response time, and error rate per network function instance; a load estimation engine (5) which estimates load per network function using LSTM / Prophet time series models; an anomaly detection module (6) which detects abnormal behavior using Isolation Forest / One-Class SVM; and predictions, anomalies,Scaling decision engine (7) that takes current status and policies as input and makes scaling (up / down), when (now / X minutes later), how much (add Y pod / Z CPU) decisions; horizontal / vertical scaling selector (8) that evaluates horizontal (multiply pods) and vertical (increase resources) scaling options; state-aware scaling orchestrator (9) that manages gentle scaling procedures for stateful network functions; inter-function dependency coordinator (10) that orchestrates chained scaling according to the network function dependency graph; automatic remediation action executor (11) that performs automatic actions after anomaly detection (restart pod, redirect traffic, increase resources, generate alarm); cost optimization engine (12) that evaluates the cloud cost model, performs spot sample suitability checks, performs inter-region cost comparisons and optimizes allocated capacity utilization.It includes a capacity planning advisor (13) that provides capacity planning recommendations with long-term (3-6 month) trend analysis, a policy engine (14) that manages scaling policies (minimum / maximum examples, scaling thresholds, waiting periods, cost limits, preferred regions), and a dashboard (15) that visualizes real-time network function status, scaling events timeline, forecast graphs, cost monitoring, SLA compliance metrics and anomaly alerts.