A service-level agreement (SLA) is an agreement between the provider and the user of services that details the expected quality of service (QoS), responsibilities, and so on. Service-level specifications (SLSs) are the technical parts of an SLA. SLSs are needed to bring dependability and assurance to the communication, which is a prerequisite for achieving a reliable production system.
SLSs for industrial settings differ from those commonly used in consumer mobile networks due to the higher demands for dependability and assurance, and they often vary even between different manufacturing operations.
The aspects that both OT and ICT players must consider before drawing up SLSs are presented in this topic. The aspects discussed is limited to SLSs used within the scope of 5G non-public networks (NPNs).
SLSs for industrial applications place a strong emphasis on ensuring service reliability. Both the IT and OT domains in an industrial network require varying services. Table 1 shows examples of ways in which many common requirements for industrial network services are met.
Table 1: Common requirements of industrial network services
| Requirements | Examples |
| Types of services | · Communication service |
| Managed service components | · Ownership, relationship, API, software, hardware, certification etc.
· Reports and their frequency |
| Service Availability | · Platforms
· Service components · Spectrum availability · Managed service components · Fault tolerance · Optional/backup connections to a public land mobile network (PLMN) service for voice and data, e.g., especially for standalone NPN (SNPN) [20] |
| Responsibilities shared by OT and ICT | · Roles and responsibilities
· Components · API, certificate authorities, credentials, shared keys (encryption), device registration and management · Access policies · QoS (implications for 5GS in an OT environment) and exposed network capabilities [12] · Tolerance level of individual networks, platforms and services · Failure recovery time for each service |
To maintain productivity within manufacturing operations, the SLSs need to be assured. Network slicing can contribute to this assurance, as specific SLSs must only be met within specific slices. This enables different SLSs to be applied within the same network. QoS monitoring functions can be used in the 5G system for SLS assurance purposes. These functions can cover both forwarding-plane and control-plane services as well as non-communication services. More about QoS can be found in the topic “5G QoS for Industrial Automation.” Prediction mechanisms can also be utilized to support SLSs assurance.
SLSs for factory and process automation should specify the OT partner’s requirements, which then serve as input for designing the 5G system. They can also serve as a guide for evaluating and approving system performance during deployment and operation. A well-defined SLS should be meaningful, understandable, achievable, measurable, and mutually acceptable [1].
Some important attributes that can be considered in SLSs are: system availability, cell availability, service area, QoS monitoring, communication service availability, communication service reliability, end-to-end latency, transfer interval, survival time, message size, user-experienced data rate, mobility, positioning service, user plane transport protocol, and clock synchronization. More details about these attributes and an outline of an SLS template using them can be found in the white paper “Service-Level Specifications (SLSs) for 5G Technology-Enabled Connected lndustries.”
The attributes and requirements defined in an SLS can vary depending on the use case. Therefore, use cases with similar service requirements or features should be grouped together so they can use the same SLS.
Further information about SLSs in industrial 5G, as well as some examples of SLSs, can be found in the White Paper “Service-Level Specifications (SLSs) for 5G Technology-Enabled Connected lndustries.”
[1] Sturm, W. Morris, and M. Jander, “Foundations of Service Level Management”, Sams Professional Series. SAMS, 2000.
