In this blog we’ll drill into the service activation and turn-up use case with a specific focus on network slicing. To kick off, let’s briefly review what a network slice is and how it’s changed from 4G to 5G.
What is a Network Slice?
To paraphrase NGMN’s definition, “a Network Slice is a set of network functions, and resources to run these network functions, forming a complete instantiated logical network to meet certain performance levels required by specific types of services.” Each slice delivers a mix of performance capabilities along three core dimensions: capacity and throughput, referred to as enhanced Mobile Broadband (eMBB) in the 5G context; reliability and latency, referred to as ultra-reliable low latency communications (URLLC); and scale of connectivity, also called massive Machine Type Communications (mMTC).
Specific examples of service types and slice performance needs include:
Mission-critical services for factory automation (ultra-low latency and high reliability)
High-speed broadband for fixed wireless access (high capacity and throughput)
Massive IoT for remote metering (low power, high-scale device connectivity)
Autonomous vehicle systems (high mobility, high reliability)
You can think of a network slices as a set of network performance “knobs” which can be dialled up or down to tailor the network to the needs of a specific service. The figure below shows an example set of network performance characteristics and the relative weighting that might exist for three different slices:
5G brings dynamic, high-scale slicing
Network slices are formed from sub-networks including RAN, transport, and Core Networks. In 4G, slicing leverages APNs and bearer routing to logically segment traffic across the end-to-end network, routing it to a dedicated packet data network gateway (P-GW). While 4G may use QoS settings to prioritize different types of services over the RAN (e.g., QCI 1 for VoLTE traffic), slicing of the network occurs only in the Core Network via dedicated APNs for a specific service type. Slicing in 4G networks has been effectively deployed for applications such as VoLTE; however, each new slice introduces significant overhead limiting the number of slices that can effectively be managed. In addition to scalability limitations, 4G network slicing is largely static, due to the complexity of creating and deploying each slice.
5G introduces a true end-to-end model for network slicing that extends from the UE to the Core. The figure below depicts a simplified 5G service-based architecture with a completely new set of functions including the Access and Mobility Management Function (AMF), Session Management Function (SMF), and User Plane Function (UPF)—and many more. These new functions have been designed from the ground up to make 5G slicing highly scalable with dynamic resource allocations driven by network orchestration and automation systems.
5G slicing requires automated assurance
5G network slicing requires automation to fulfill its true potential. That means that operational testing and assurance to design, on-board, deploy, operate, and maintain slices also need to be automated. This includes a set of critical testing and assurance functions such as validating the performance of new slices, proactively monitoring service levels, and isolating faults.
In our early 5G engagements with tier-1 providers, we’re currently discussing a variety of network slicing strategies. Some providers want to offer a limited set of network slices for the foreseeable future. Others are looking at potentially launching hundreds of network slices in coming years, with the potential to offer thousands and—in the limit—one per subscriber! It’s clear that traditional manual techniques simply won’t be fast enough or economical enough to support the kind of scale our customers are envisioning.
Our recommended approach is to integrate automated assurance for slice activation as part of early 5G deployments.
Spirent’s approach to 5G slice activation
Our recommended approach is to integrate automated assurance for slice activation as part of early 5G deployments. Automated assurance for 5G must include virtual test agents (VTAs) that emulate small volumes of UEs and gNBs (i.e., the 5G RAN) to perform tests across the Core Network to the DN (see the figure below). Key performance metrics such as throughput, packet loss, latency, and jitter should be measured from the user’s perspective. Controller and Analytics components should enable integration of these VTAs with management and network orchestration (MANO), and automation platforms.
In the design phase of a new slice, VTAs may be used to help identify which services benefit from having their own slice and which specific configuration settings, such as QoS, lead to acceptable quality for customers. During network slice deployment, the Analytics, Controller, and VTA components of automated assurance enable performance tests to be automatically executed when a new network slice is activated, to assure it will perform when customers start using it. Any problems can be isolated and resolution guidance can be provided to orchestration and automation layers. Once the slice is operational, performance tests can be continuously performed to proactively assure that the specific performance capabilities of the slice are being met and that any related SLAs can be effectively managed.
By proactively validating performance during activation, providers can assure they’ll deliver on the performance promises they’ve made for the new slice. For mission-critical slices with SLAs that have financial penalties, meeting these promises is about more than just customer satisfaction and loyalty—it means savings of millions of dollars per year!
When can you expect this type of automated assurance to be commercially ready? One or two years from now? After all, 5G standards are just completed, right? Let’s just say you may be in for a pleasant surprise if you visit Spirent at tmforum’s Digital Transformation World in Nice on May 14th!