SPN

What Is SPN?

SPN stands for Service Packet Network, which is a next-generation transport network architecture that builds upon, improves, and pioneers packet transport network (PTN) technology. Initially introduced by China Mobile, Huawei, and other leading equipment vendors to meet the demands of 5G transport, this innovative system centers on slicing Ethernet (SE) technology. Its key technical benefits include delivering low latency, high bandwidth, extremely precise synchronization, and flexible management and control capabilities. Moreover, it seamlessly integrates with the Ethernet ecosystem while boasting cost-effectiveness and simplified deployment processes.

Why Do We Need SPN?

With the emergence of new and diverse services in the 5G and cloud era, there is a growing demand from various industries, services, and users for higher bandwidth, lower latency, increased reliability, and other performance metrics in transport networks. These networks must support a range of functionalities, including flexible traffic scheduling, networking protection, as well as management and control, while ensuring performance in terms of bandwidth, latency, synchronization, and reliability. However, traditional PTN solutions fall short in meeting the transport requirements of 5G services.

Utilizing the Ethernet transmission architecture, SPN inherits, enhances, and innovates the functionalities of the PTN transport solution. SPN integrates a lightweight TDM layer into the Ethernet physical (PHY) layer, enabling packet devices to offer network slicing capabilities, such as hard isolation and deterministic low-latency, without requiring modifications to existing packet technologies. SPN has garnered significant attention and industry support globally due to its support for packet and TDM convergence, low latency, network slicing, and compatibility with existing ecosystems, leading to substantial cost reductions. The G.mtn project was officially approved by the ITU-T in October 2018, signaling the maturation of slicing Ethernet (SE) technology driven by SPN, and endorsing the core concept and technical architecture of SE.

SPN enhances and innovates PTN in various ways.

• SPN utilizes SR-TP for establishing manageable and controllable Layer 3 (L3) tunnels and for constructing an end-to-end (E2E) L3 service deployment model. It introduces an L3 control plane characterized by centralized management and control. Through the Software-Defined Networking (SDN) platform, which integrates management and control functions, SPN logically abstracts and virtualizes the physical resources of network elements (NEs), including forwarding, computing, and storage resources, to dynamically create a virtual network as needed. Consequently, SPN is capable of presenting a network model with multiple networking architectures within a single physical network, providing users with an open, flexible, and efficient network operating system.

• SPN introduces FlexE interfaces to decouple service rates from physical channel rates based on Ethernet. This allows for flexible definition of physical interface rates, eliminating the requirement for them to be equal to client service rates.

• SPN introduces Four-Level Pulse Amplitude Modulation (PAM4) for Ethernet gray light and Wavelength Division Multiplexing (WDM) to decrease the transmission expenses of metropolitan and local networks.

The Advantages of SPN

SPN Architecture

Similar to the ITU-T network model, the SPN architecture is structured in layers. Illustrated in the diagram below, the SPN architecture primarily comprises the slicing packet layer (SPL), SCL, and slicing transport layer (STL). It also encompasses two planes: the management/control plane and the time/clock synchronization plane.

• The SPL is responsible for addressing, forwarding, and encapsulating transport pipes for various services such as IP, Ethernet, and constant bit rate (CBR). It supports service types like L2VPN, L3VPN, and CBR transparent transmission, utilizing multiple addressing mechanisms including IP, MPLS, 802.1Q, and physical interface protocols. Through these mechanisms, the SPL facilitates service mapping, service identification, traffic distribution, and ensures Quality of Service (QoS) for different services.

• The SCL ensures end-to-end channelized hard isolation for network services and slices. Leveraging innovative SE technology, it handles timeslot processing for Ethernet physical interfaces and FlexE groups, offering Ethernet-based virtual network connection capabilities. Additionally, it provides low-latency and hard-isolation slicing channels for multi-service transport based on Layer 1 protocols. Moreover, the SCL supports OAM and protection functions for SE paths, enabling performance monitoring and fault rectification across the network.

• The STL facilitates efficient transmission of large-bandwidth data using IEEE 802.3 Ethernet physical-layer technologies and Optical Internetworking Forum (OIF) FlexE technologies. The OIF FlexE PHY layer supports various high-speed Ethernet interfaces such as 50GE, 100GE, 200GE, and 400GE, enabling the construction of low-cost, high-bandwidth networks. Additionally, it supports mainstream networking modes with single hops of up to 80 km.

• The management/control plane provides SDN management and control capabilities, allowing flexible configuration of services and network resources, as well as automatic and intelligent network Operations and Maintenance (O&M) capabilities.

• The time/clock synchronization plane facilitates the deployment of high-precision clock sources on core nodes and provides IEEE 1588v2-based high-precision time synchronization transport capabilities. This ensures the fulfillment of requirements for basic and collaborative 5G services requiring high-precision time synchronization.

SPN architecture

Ultra-High Bandwidth Transmission

As the demand for bandwidth continues to grow, transport networks are in urgent need of new technologies. One such technology is SPN, which integrates high-speed Ethernet interface technology with DWDM multi-wavelength technology to provide ultra-large bandwidth.

Ethernet is a widely adopted interface technology in the communications industry. With the advancement of the Internet, high-speed Ethernet interfaces have made significant progress in recent years to accommodate the increasing bandwidth demands of services. FlexE, aiming to achieve flexible rate adaptation, separates the MAC and PHY layers by introducing a FlexE shim layer based on IEEE 802.3 standards. A FlexE interface can serve as a new high-speed Ethernet interface, such as 50GE, 100GE, 200GE, or 400GE, facilitating cost-effective and high-bandwidth network deployment while meeting the demands for increased bandwidth.

High-speed Ethernet interfaces encompass the following essential technologies:

• Forward Error Correction (FEC): enables long-distance transmission utilizing the well-established KP4 FEC.

• PAM4: doubles the data rate without altering the baud rate, thereby lowering the expense of optical interfaces.

SPN integrates FlexE bonding and DWDM for flexible expansion and segmentation of transport network bandwidth. FlexE aggregates multiple optical interfaces to deliver high-speed Ethernet interfaces at a cost-effective rate. The maximum bandwidth capacity of a FlexE link depends on the number of interfaces aggregated into a FlexE group. For instance, four 200GE Ethernet interfaces can be aggregated to yield an 800GE bandwidth. This implies that FlexE aggregation enables seamless bandwidth expansion without requiring service adjustments — the greater the number of interfaces that can be bundled into a FlexE group on a device, the higher the network scalability. Moreover, in addition to furnishing high bandwidth for individual fibers, FlexE and DWDM facilitate smooth, on-demand bandwidth expansion when operating with DWDM channels.

Low-Latency and Reliable Transport

Ultra-low latency is critical for 5G services, with eMBB services requiring a maximum E2E latency of 10 ms, and URLLC services demanding even lower E2E latency. However, traditional packet networks often introduce queuing in the outbound direction, resulting in high latency that can reach tens of microseconds (μs). During network congestion, latency may even exceed 10 ms, which is unacceptable for latency-sensitive 5G services.

FlexE packet slicing addresses this challenge by replacing the traditional packet-by-packet transmission mechanism with a strict TDM scheduling mechanism. Unlike traditional interface scheduling, which prioritizes packets based on packet priorities, FlexE packet slicing ensures that channels occupy exclusive bandwidth, thus preventing low-priority long packets from blocking high-priority short packets and avoiding service interference. This timeslot-based scheduling approach allows for dedicated bandwidth allocation and enhances service security by strictly isolating services from each other.

Defined by OIF, FlexE offers interface-level slicing capabilities. Leveraging FlexE, G.mtn (defined by the ITU-T) enhances OAM and cross-connect features as part of SPN's ongoing development. G.mtn supports both interface- and channel-level slicing capabilities.

Channelization of FlexE packet slicing

Efficient and Flexible Slices

With the advent of diverse services in the 5G and cloud era, various industries, services, and users impose different service performance requirements on networks. As transport networks are tasked with accommodating services from numerous industries, the need for network slicing arises to isolate emerging industries, thereby reducing the impact of service deployment on the entire network and minimizing trial-and-error costs. Although FlexE, developed by OIF, offers a mechanism for logical slice isolation based on Ethernet physical interfaces, it lacks SLA isolation for end-to-end (E2E) services. To address this limitation, MTN is introduced within the SPN technical architecture. MTN slicing, built on native Ethernet and extended for compatibility with the current IEEE-defined Ethernet ecosystem, facilitates Ethernet Layer 1 networking. This approach not only ensures deterministic low latency and hard pipe isolation capabilities but also eliminates the need for packet buffering and forwarding at Layers 2 or 3.

MTN incorporates the following core technologies:

• Ethernet 64B/66B block-based timeslot cross-connect: This technology ensures low latency and transparent transmission, while also providing robust hard isolation.

• On-demand end-to-end (E2E) Operations, Administration, and Maintenance (OAM): By employing idle block replacement within the IEEE 802.3 block extension, MTN enables OAM and protection for slicing Ethernet (SE) paths, thereby supporting E2E Ethernet Layer 1 networking.

• Transparent mapping of Ethernet services: Through a transcoding mechanism, MTN seamlessly maps various services onto SE paths for transmission, ensuring efficient and transparent delivery.

SPN offers both robust hard slicing capabilities, utilizing MTN TDM channels, and adaptable soft slicing capabilities, employing Segment Routing (SR) over Ethernet packet switching channels and Quality of Service (QoS) technology. Hard slicing ensures low latency and bandwidth guarantees for critical services like private lines and Ultra-Reliable Low Latency Communications (URLLC), while soft slicing delivers high bandwidth and differentiated Service Level Agreement (SLA) assurances for packet services such as Enhanced Mobile Broadband (eMBB).

Outlined in the diagram below, slices are implemented as follows:

• Default slice: DSCP priorities are adjustable to prioritize high-priority packet forwarding. This slice type is suitable for users with minimal slicing requirements.

• Packet+MTN interface slices (shared TDM slices): Different users of Layer 3 Virtual Private Network (L3VPN) slices are isolated by Virtual Private Networks (VPNs) and provided bandwidth through Committed Information Rate (CIR). This slice type caters to users with moderate slicing needs.

• MTN channel slices: End-to-end hard slicing channels with dedicated resources offer TDM isolation, ultra-low latency, and bandwidth assurances. This slice type is tailored for users with extensive slicing requirements.

Slicing implementation

Optimized SDN-based Centralized Management and Control

SPN utilizes the SDN paradigm to achieve open, flexible, efficient, and automated network Operations and Maintenance (O&M). Apart from facilitating automatic service provisioning, SPN can continuously monitor network conditions to dynamically optimize network performance. The diagram below illustrates SPN's progression towards intelligent management and control, within a unified platform that integrates management, control, and analytics functionalities.

Intelligent management and control platform

In an SPN environment, management, control, and analysis functionalities are executed as follows:

• Management: An SDN controller collaborates with network devices to automate service deployment and optimization processes. Node and adjacency labels are configured on the network devices to generate forwarding entries. Subsequently, label information is disseminated and delegated to the controller, enabling it to autonomously establish Segment Routing (SR) tunnels. Upon gathering route topology information and labels for each node and adjacency, the controller performs global path computation and delivers an SR label stack to the ingress node. Upon receiving the stack, the ingress node encapsulates and forwards the packets accordingly.

• Control: SR-TP rerouting safeguards the network against multiple points of failure, ensuring uninterrupted service availability and facilitating flexible and reliable connections. Previously, both working and protection paths followed a linear configuration, resulting in service disruptions if both paths failed simultaneously. This necessitated rapid fault identification and service restoration, incurring substantial maintenance expenses. Moreover, manual route adjustments during the repair process led to inefficiencies. NCE offers real-time path control capabilities for SR-TP tunnels, enabling rerouting during SR-TP tunnel path computation and protection. These capabilities ensure that, in the event of concurrent failure of both working and protection paths, the control plane initiates rerouting to calculate a new optimal path. Consequently, sub-second service restoration is achieved, guaranteeing continuous service operation. For additional details on SR-TP, refer to the SR-TP tunnel documentation.

• Analysis: Traditional TP OAM relies on out-of-band measurement, which can be conducted either indirectly or directly. Indirect measurement utilizes simulated service data, potentially resulting in inaccuracies as the constructed service path may not accurately reflect real-world conditions. Direct measurement involves inserting measurement packets into service flows for tunnel-level measurement, but it lacks support for service-level measurement and suffers from low measurement precision due to packet interval transmission. Both methods have limitations. In-situ Flow Information Telemetry (IFIT) addresses these shortcomings by directly measuring real packets, allowing for accurate reflection of the actual path and latency information. IFIT can also measure individual packets, enabling precise detection of minor packet loss and facilitating visualization of service-level SLA compliance.

The Application of SPN

Smart Mining Service

As the coal mining industry transitions towards greater intelligence, coal mine networks face increasingly demanding performance requirements. Typically, coal mines operate multiple underground networks to support production services such as remote monitoring and remote operation. These services require hard isolation to guarantee independent bandwidth for each service. Additionally, the challenging underground environment makes device maintenance difficult, necessitating highly reliable underground networks. Network faults can significantly impact operations, requiring time-consuming troubleshooting procedures that hinder production efficiency. Moreover, the concurrent operation of multiple underground networks with numerous maintenance points increases network construction costs. Current networks also struggle to provide sufficient bandwidth to support the widespread deployment of high-definition (HD) cameras. To address these challenges, SPN offers an intelligent slicing solution with fixed-mobile convergence (FMC). This solution ensures high bandwidth, low latency, and high reliability for coal mine networks.

• 50GE equipment is deployed underground to establish an access ring, with FlexE slicing applied to 50GE links to construct an ultra-broadband infrastructure network.

• Core SPN devices situated above ground in the mining area connect to the core network and system platforms via various interfaces. Specifically, these SPN devices are deployed within the multi-access edge computing (MEC) equipment room and utilize diverse physical interfaces to link to the User-Network Interfaces (UNIs) of the MEC node, security supervision system, remote control system, and video surveillance system.

• Apart from their connection to the MEC node, a pair of SPN devices (core nodes 1 and 2 in the mining area as depicted in the accompanying diagram) also establish connections to the access rings above and below ground, serving as common aggregation nodes within the metropolitan SPN transport network.

• Base station Baseband Units (BBUs) and SPN devices share deployment within the same equipment room, resulting in reduced investment in equipment rooms.

Smart mining solution

Underground surveillance, AI-based intelligent identification, underground video communication, and underground remote control services — whether utilizing wired or wireless communication — can all be accommodated on the SPN, enabling intelligent slicing with fixed-mobile convergence (FMC).

• Both wired and wireless services can be supported by different slices on the same link, ensuring stringent isolation between the services.

• The video surveillance, remote control, and security monitoring systems are segregated into distinct slices, ensuring robust isolation. In the event of congestion in one slice, services in other slices remain unaffected.

• Each network slice's bandwidth can be dynamically adjusted without impacting existing services, allowing for flexible scaling in or out as needed.

Intelligent slicing featuring FMC

SR rerouting guarantees continuous service availability. With SR-TP tunnels, NCE delivers real-time path management, including path computation and rerouting in response to faults. In the event of simultaneous failures at multiple points, the control plane initiates rerouting to establish a new optimal path swiftly, enabling rapid service restoration and self-healing.

Rerouting for fast service self-healing

Ultra-HD Video Service

As the demand for high-quality video content, including ultra-HD technologies like 4K, 8K, and VR/AR, continues to rise, network quality becomes increasingly crucial. While carriers possess significant network infrastructure advantages, enabling them to offer superior video services compared to Internet Over the Top (OTT) providers, traditional network architectures primarily cater to High-Speed Internet (HSI) services. Unlike HSI services, big video services exhibit high concurrency and necessitate substantial bandwidth, low packet loss rates, and minimal latency. This presents carriers with bandwidth and latency challenges as they strive to meet the demands of these evolving services.

To tackle these challenges, carriers can leverage SPN to implement network slicing based on FlexE hard pipes. Each network slice carries its own set of services, ensuring isolation from other services. For instance, carriers can dedicate a separate network slice to IPTV services to prevent network congestion from impacting service quality, thus achieving superior performance compared to OTT services. The utilization of SPN in ultra-HD video services is illustrated in the accompanying diagram. To monitor video traffic performance, including metrics like latency, bandwidth, and jitter, IFIT can be employed. Additionally, IFIT can periodically gather and analyze network statistics to facilitate adjustments and ensure service performance.

Application of SPN in the ultra-HD video service