Network slicing’s primary benefit is its capacity to divide various business functions into separate, logical network segments. These segments, or network slices, can be customized and optimized for the specific needs of the supported business operation. This model allows for efficient and dependable service delivery because resources and functionalities can be accurately allocated and managed within each slice. As part of this ongoing series of 5G-related blog posts (https://www.vamsitalkstech.com/tag/5g/) where network slicing has also been previously discussed (https://www.vamsitalkstech.com/?s=slicing). Let us revisit a reference architecture for network slicing.
The Network Slicing Value Chain
Consider the illustrative example of a mobile operator leveraging network slicing to cater to diverse application demands. They might establish a dedicated network slice optimized for enhanced Mobile Broadband (eMBB) services, designed to deliver high data rates and seamless connectivity for bandwidth-intensive applications like video streaming and high-speed internet access. Simultaneously, another network slice could be provisioned to support massive Machine-Type Communications (mMTC), characterized by its ability to handle a vast number of low-power devices transmitting small amounts of data, ideal for applications such as smart city sensors and IoT deployments. Furthermore, a third network slice could be created for Ultra-Reliable Low Latency Communications (URLLC), engineered to provide extremely low latency and high reliability, essential for mission-critical applications like autonomous vehicles, industrial automation, and remote surgery.
Network slicing is evolving beyond basic network connectivity to capture a larger share of the value chain within its ecosystem. This progression is not a singular leap but a measured journey influenced by both market forces and technological advancements.
Key Concepts in Network Slicing
- The deployment of a cloud-native 5G core introduces critical considerations for network slicing. A key decision during slice planning involves determining whether mobility Network Functions (NFs) should be shared across multiple slices or dedicated to individual slices. While the absolute minimum requirement for a network slice often includes a dedicated User Plane Function (UPF), practical considerations based on the specific use case and billing requirements often necessitate dedicating other core network functions as well. These can include the Session Management Function (SMF), Policy Control Function (PCF), Charging Function (CHF), and NF Repository Function (NRF), among others, to ensure the required level of isolation and performance. The physical or logical placement of these NFs is another crucial aspect that must be carefully evaluated when designing a network slice, considering factors such as latency, resilience, and resource efficiency.
- Consequently, the technical specifications and solutions related to network slicing are currently distributed across various SDOs, leading to a degree of fragmentation. To realize the vision of a truly interoperable E2E network slicing solution, significant effort is required in terms of cross-SDO and open-source project cooperation and coordination. The concept of network slicing was initially highlighted as a key enabler for 5G capabilities and value creation in the NGMN 5G White Paper. Subsequently, it has become a fundamental feature specified by 3GPP to be supported by the 5G System. In 3GPP Release 15 specifications, the foundational system architecture, including dedicated identifiers to uniquely recognize each slice, has been standardized across both the core and Radio Access Network (RAN) domains. This provides a crucial framework upon which more advanced slicing capabilities are being built in subsequent releases.
- The transport network plays a vital role in delivering the performance guarantees promised by network slices, and its behavior, governed by SLAs, must be carefully planned. Different types of slices will impose distinct requirements on the transport network. For instance, ultra-reliable low-latency communication (URLLC) slices demand extremely low latency and high reliability. Other slices might prioritize high bandwidth and resilience, requiring features like fast reroute mechanisms and avoidance of shared risk link groups (SRLG). Some slices may necessitate end-to-end encrypted paths to meet stringent security requirements. Therefore, a comprehensive understanding of the specific needs of each slice is paramount for effective transport network planning, ensuring that the underlying transport infrastructure can adequately support the diverse performance and security demands of the different virtual networks.
Reference Architecture – Network Slicing in a 5G Network
Network slicing partitions physical network infrastructure into multiple isolated and virtualized logical networks, known as network slices. This division allows for the independent management and optimization of each slice, ensuring that diverse service requirements can be met with tailored network capabilities. Each slice operates as a self-contained logical network, benefiting from the efficiency of a shared underlying physical infrastructure while maintaining strict isolation and dedicated resources.
As shown above, the fundamental structure of network slice architecture comprises three interconnected layers that work in concert to deliver end-to-end network services:
- Radio Access Network (RAN): This layer is responsible for managing wireless connections between user equipment (UE) and the network. Within the context of network slicing, the RAN is partitioned to enforce distinct Quality of Service (QoS) levels for each slice. This partitioning allows for the prioritization of traffic and the allocation of radio resources according to the specific needs of each service. For instance, a slice dedicated to enhanced Mobile Broadband (eMBB) might prioritize high throughput and capacity, while a slice for Ultra-Reliable Low Latency Communication (URLLC) would focus on minimizing latency and ensuring high reliability. The RAN’s ability to create these separate partitions is crucial for delivering differentiated services over a shared radio spectrum.
- Transport Network: This layer provides the essential connectivity between the RAN and the core network. In a network slicing architecture, the transport network ensures secure and isolated data movement for each slice through the establishment of dedicated Virtual Private Network (VPN) pathways. These VPNs guarantee that traffic belonging to one slice remains segregated from other slices, enhancing security and preventing interference. The transport layer is responsible for maintaining the agreed-upon QoS parameters across the wired network infrastructure, ensuring that the performance characteristics defined for each slice are consistently delivered from the RAN to the core.
- Core Network: The core network serves as the central intelligence and management hub of the network. In a sliced architecture, the core network runs containerized applications, with each set of containers dedicated to managing the specific network functions required for a particular slice. This containerization allows for the deployment of customized network functionalities tailored to the unique demands of each service. For example, a slice designed for IoT applications might have lightweight and energy-efficient core network functions, while a slice for mission-critical services would incorporate redundant and highly available network elements. The core network is responsible for tasks such as session management, mobility management, and policy enforcement, all performed independently for each network slice.
To effectively manage and orchestrate these network slices, an overarching control framework is implemented. This framework provides essential functionalities, including:
- Resource Tracking and Billing: The control framework monitors the resource utilization of each network slice, enabling accurate tracking and billing based on the specific resources consumed. This ensures that service providers can effectively monetize their network slicing capabilities.
- Automated Slice Creation and Modification: The framework automates the entire lifecycle of network slices, from their initial creation and deployment to subsequent modifications and eventual termination. This automation streamlines network operations, reduces manual intervention, and allows for the rapid provisioning of new services.
- Service Level Management Across Domains: The control framework ensures that the agreed-upon Service Level Agreements (SLAs) for each slice are consistently maintained across all network domains, including the RAN, transport, and core. This end-to-end management of service levels is critical for guaranteeing the performance and reliability promised to customers.
- Coordinated Resource Allocation: The framework intelligently coordinates the allocation of network resources across different physical locations, including edge, regional, and national data centers. This coordination takes into account factors such as latency requirements and resource availability to optimize the performance and efficiency of each network slice. For latency-sensitive applications, resources might be allocated closer to the network edge, while other services might leverage regional or national data centers.
Each network slice operates autonomously with its own distinct set of characteristics, ensuring that it can meet the specific requirements of the services it supports:
- Performance Parameters: Each slice is configured with specific performance targets, such as bandwidth, latency, jitter, and packet loss rate, tailored to the needs of the applications running on that slice.
- Resource Allocation: Network resources, including radio spectrum, processing power, storage, and network capacity, are allocated to each slice based on its specific demands, ensuring that sufficient resources are available to meet its performance objectives.
- Security Policies: Each slice can have its own set of security policies and mechanisms, providing enhanced isolation and protection for sensitive data and applications. This allows for the implementation of stringent security measures for critical services without impacting other slices.
- Service Requirements: Each slice is designed to support specific service requirements, such as mobility management capabilities, quality of experience (QoE) parameters, and specific network functionalities needed by the applications it serves.
As is common with 5G deployments, the above architecture embraces a distributed processing model, strategically distributing computational tasks across different physical locations – edge, regional, and national – based on the specific latency requirements of the services and the availability of resources. This distributed approach optimizes performance by bringing processing closer to the user for latency-critical applications while leveraging centralized resources for other tasks. Furthermore, comprehensive management tools are employed to handle essential operational functions such as deployment, ongoing monitoring of performance and resource utilization, and efficient troubleshooting across all the virtualized and physical network segments, ensuring the overall health and efficiency of the network slice infrastructure. This design allows operators to run multiple distinct networks on shared infrastructure while maintaining separation and specific performance characteristics for each service type.
Commercially, the pace of network slicing adoption is contingent upon several factors. Local market demand dictates the immediate need and potential revenue from tailored network services. The maturity and engagement of the ecosystem, including device manufacturers, application developers, and vertical industry players, are crucial for creating and delivering end-to-end solutions. The ability to scale successful use cases from initial trials to widespread deployment is essential for realizing the full economic benefits. Furthermore, diverse regulatory landscapes across different regions can significantly impact the deployment and commercialization of network slicing.
Technically, the feasibility and timeline of network slicing depend on the readiness of underlying technologies. Device maturity, with features that can leverage network slicing capabilities, is a prerequisite. The availability of end-to-end solutions, integrating network, cloud, and edge components, is vital for delivering seamless and guaranteed service levels. Additionally, the maturity of local system integration capabilities, including the expertise to deploy, manage, and orchestrate network slices within existing infrastructure, plays a critical role in successful implementation.
Conclusion
By abstracting the complexities of the underlying physical network infrastructure, network slicing provides a highly flexible and programmable environment. This virtualized and software-defined approach allows operators to dynamically adapt and reconfigure network slices in response to evolving business needs, shifting traffic patterns, and the introduction of new technological advancements. This adaptability ensures that the network infrastructure remains agile and can readily accommodate future demands and innovations in the telecommunications landscape.
References
[1] GSMA E2E Network Slicing Architecture – https://www.gsma.com/newsroom/wp-content/uploads//NG.127-v2.0-3.pdf