As mobile networks grow and expand from core to edge, and as 5G implementations begin, virtualized infrastructures are becoming increasingly critical in CSPs’ deployments. Unlike generic IT workloads, NFV has stringent KPIs, such as deterministic performance, high throughput, or low latency, which often require using the underlying infrastructure in inflexible and inefficient ways to achieve them. For instance, dataplane VNFs keep servers where they run constantly at a high power-up state, as if they are always operating for peak demand. Similarly, operators deploy critical functions typically in isolation, reserving upfront a large portion of server resources to prevent contention from other services and avoid Service Level Objective (SLO) violations. However, this approach leaves a significant portion of the infrastructure unutilized. Intracom Telecom’s NFV Resource Intelligence uses AI to achieve autonomous service assurance, addressing the challenges mentioned above. It automatically determines the ideal distribution and configuration of resources in closed-loop
forms and dynamically, under any traffic or colocation condition. This ensures that SLOs are always maintained and resources are used cost-efficiently, solving a highly complex challenge that goes beyond human expertise. Intracom Telecom’s NFV Resource Intelligence guarantees optimal execution of virtualized Network Services and optimal utilization of the infrastructure where they are running.
Highlights
Reduces energy consumption od DPDK-based packet processing VNFs by throttling their power according to their actual traffic load
Throttles power consumption of DPDK-based packet processing VNFs according to their actual traffic load.
Carefully slices and isolates critical hardware resources to protect high-priority services from “noisy neighbors” using advanced hardware technologies.
Automatically discovers optimal resource configurations that deliver certain performance levels for one or more VNFs, specified by the user.
Enables denser VNF placement (colocation) without introducing contention and SLO violations, increasing infrastructure utilization.
Exposes rich telemetry to the user via customized dashboards, delivering maximum observability for both the VNFs and the platform.
Supports multiple types of VNFs: KVM Virtual Machines, native Linux applications, Docker containers, Kubernetes pods.
Leverages AI and closed-loop control to realize autonomous service assurance.
Energy optimization for user-plane network functions
5G’s UPF network functions have high performance requirements, which are usually met by frameworks like DPDK. However, these frameworks require servers to run at high power states even during periods of low traffic. NFV-RI™ uses AI-driven mechanisms to manage the power of user-plane network functions dynamically, while ensuring zero packet drops, resulting in significant energy savings. In a PoC, NFV-RI™ achieved a 14% reduction in power consumption over 24 hours for a vEPC node prototype and 17-35% daily power savings for 5G UPFs.
Intent-based 5G Core slicing
Allocating resources to key network functions is a challenging task for CSPs when deploying multiple network slices with differentiated performance characteristics. NFV-RI™ simplifies this process by offering AI-based workflows that automatically deliver customized performance for 5G Core slices, using the least amount of resources. In a PoC, NFV-RI™ delivered the intended latency and packet drop objectives for two colocated 5G UPFs that were servicing subscriber groups of different priorities, resulting in energy savings ranging from 16% to 43%.
