5G network slicing is essentially an E2E concept that connects user equipment to tenant-specific applications. E2E network slicing consists of RAN slices, core slices and transport slices. Each network domain has its own slice controller, which helps to meet the E2E SLA of the entire 5G network.

The figure below shows an example of an end-to-end 5G network slice, where there are 5 network slices (namely NS1 to NS5), and each color represents a different SLA. Each end-to-end network slice supports specific customers or “tenants” and specific services, where slices can be shared to host multiple services. Soft/hard slice operators need to ensure that the traffic on one slice does not interfere with the performance of another slice, which means that traffic isolation mechanisms are required, which are usually divided into hard isolation mechanisms and soft isolation mechanisms, that is, “hard slices and soft slices” . Soft slicing means sharing the physical infrastructure, creating logical segmentation between customers, with a lower level of traffic isolation. According to this definition, soft slicing is not a new concept from the perspective of IP/MPLS network. A traditional L3VPN can be seen as an implementation example of soft slicing in an MPLS network, a VPN can be thought of as a series of tunnels connecting to customer sites, each site may have different QoS processing, and all traffic to and from the site is internal to the customer. In VPN services, service providers configure routing policies on shared physical infrastructure to ensure that each customer’s traffic is logically differentiated. In this sense, network slicing at the MPLS level can be achieved using: • Virtual Routing and Forwarding (VRF), which enables multiple routing contexts over a shared MPLS transport network. • Virtual Interface (VSI), which supports multiple switching environments on the same shared infrastructure. Each physical router is capable of hosting multiple VRFs and multiple VSIs, effectively dividing it into multiple routing and switching environments that can be assigned to different tenants/services. Hard slicing refers to providing dedicated resources for specific network slicing instances. The main difference between “hard slicing” and “soft slicing” is that hard slicing dedicates network resources to one slice, while soft slicing allows the use of shared resources. Allocating dedicated, unshared resources to each network slice instance guarantees the performance, availability, and reliability required by each application or customer. However, if these resources are not fully used, they cannot be used for other slices. Therefore, hard slicing may not be a cost-effective option. Soft slicing allows controllable overbooking of transport resources, allowing network resources to be more economically used for high-capacity applications with looser constraints. Key technologies for 5G network slicing: OTN and FlexE >OTN OTN has always played an important role in optical networks. OTN’s transparent transmission, complete OAM, protection and other functions can meet the requirements of new services for service quality. Furthermore, OTN provides an efficient way to multiplex different services into optical paths,

How to use OTN to carry FlexE? Today, a multitude of services and applications are generated with bursty, unpredictable traffic patterns, with widely varying and more stringent requirements for bandwidth and data transfer performance. OTN is also an important part of the 5G network system. In order to realize customized services and fine-grained management and control of network resources, and meet the higher performance requirements of 5G services, network resources need to be allocated according to SLAs, so as to implement network slicing for OTN. OTN transparently encapsulates each client payload into a container for transport across the optical network while preserving the client’s local structure, timing information, and management information. OTN’s enhanced multiplexing capabilities allow different traffic types to be transported on a single optical transport unit frame, including Ethernet, storage and digital video, and SONET/SDH. Since OTN is a fully transparent protocol, it can be easily adapted to existing services. >FlexE Ethernet is a statistical multiplexing technology. The services carried by it can share the bandwidth of the entire interface, which can maximize bandwidth utilization and facilitate deployment. However, this feature also prevents strict isolation between carried services. It is impossible to provide resource reservation with certain bandwidth for different services, so that Ethernet cannot meet the requirements of the traditional transmission network built with SDH/OTN technology. However, as the traditional SDH protocol stops developing on STM-64 (40G), OTN technology is more focused on the development of large bandwidth, and there is insufficient support for the transmission requirements between the branch headquarters of the enterprise network with low granularity and low bandwidth requirements, resulting in There is a gap between the traditional transmission leased line service requirements and the technological evolution.

FlexE is a technology developed on the basis of Ethernet technology to meet the requirements of high-speed transmission and flexible bandwidth configuration. The popularity of FlexE has prompted some service providers to consider using it to separate traffic at the physical layer, and implement slicing and strict isolation at the physical layer through strict time-division multiplexing channelization technology. Now, Ethernet-based transports can separate different types of services within the same transport path or port.

For example, a 100G FlexE can divide the capacity into 6 slices, corresponding to 6 FlexE clients, and each slice carries a type of service. FlexE can provide a high degree of isolation of network resources, ensuring that the traffic from one FlexE client will not affect the traffic of other FlexE clients. By encapsulating more slices on FlexE, the wavelength capacity can be optimized. When FlexE is coupled with advanced Layer 2 and Layer 3 technologies on the Overlay, traffic engineering and QoS technologies can be used to achieve further optimization. The example below demonstrates how a 100G FlexE instance on one wavelength can host 4 network slices while maintaining a balance between hard isolation and statistical multiplexing.

Summary OTN (ITU-T G.709) standards and equipment have long been dedicated to providing absolute transmission guarantees in the network. OTN is based on time-division multiplexing (TDM), which enables zero-packet loss and low-latency packet transmission. However, OTN does not support statistical multiplexing and cannot utilize capacity as efficiently as Ethernet. As a new-generation network hard slicing basic technology, FlexE can meet the requirements of 5G service bearing for large network bandwidth, slicing, and service physical isolation, and provides technical support for operators to build a network infrastructure suitable for the long-term development of 5G services. Therefore, the three major operators have clearly defined FlexE as the basic bearer technology in the future, and written it into the technical specifications for the construction of the next-generation bearer network. Although FlexE can currently meet the requirements of large-bandwidth and coarse-grained backbone networks, it is still in the process of further developing requirements for low-speed Ethernet interfaces and fine-grained slices.