As enterprises weigh the integration of 5G technologies into their corporate network environments, they face new questions about how best to ensure consistent quality of service (QoS) for essential business applications across the 5G RAN and traditional corporate LAN. The need for uniform and deterministic connectivity across both network technologies becomes an absolute must.
Traditional QoS within corporate networks, such as the use of traffic flow identification, classification, DSCP assignments, and queuing/dropping techniques, have long been used as the de facto method of controlling and prioritizing the convergence of voice, video, and data.
Modern 5G technologies have now taken portions of what's great about QoS and additional built-in improvements to codify what's generally referred to as 5G network slicing. This is a set of standards formulated by the 3GPP, the primary standards body responsible for advancing the architecture of 5G.
A new approach to enterprise 5G QoS, called “microslicing,” takes the 5G network slicing concept one step further. Emerging as a popular approach to helping organizations guarantee service quality and enforcement from the cellular radio access network (RAN) across the entire enterprise IT infrastructure, Microslicing is widely believed to be a vital technique in simplifying the integration of 5G technology within enterprise networks.
What is microslicing?
Microslicing is an evolutionary technology that extends QoS on the private 5G radio access network (RAN) across the existing enterprise infrastructure with the goal of achieving consistent end-to-end performance for critical applications that simply can’t tolerate erratic packet transmission across a private cellular wireless medium.
Microslicing is defined as a set of network functions within a 5G LAN that allows various QoS metrics and traffic thresholds to be automatically defined and enforced on discrete application flows or device groups. This creates a cohesive, end-to-end logical network that maps QoS from the ingress cellular RAN through the entire existing enterprise L2/L3 network.
A single microslice can be configured to meet an application’s network requirements from a service level agreement (SLA) perspective. An SLA of a microslice includes a range of network performance requirements for different traffic traversing the 5G RAN. Some of the configurable options include:
- Flow priority
- Packet delay budget (PDB)
- Packet error rate (PER)
Within the microslicing framework, the control plane continuously monitors and adjusts traffic transmissions as network congestion occurs to meet the established QoS criteria. These processes and functions guarantee that higher-priority traffic is given preference and special treatment over lower-priority flows in real-time as defined by network administrators.
Slicing things up
Microslicing relies on the telecommunications concept of bearer channels. Bearer channels are logical paths across the network that are solely responsible for user and device data transport.
A microslice can be configured as one of three bearer types depending on the performance needs of the application data being transported across the bearer channel. The three types of microslice bearer channels are:
Guaranteed bit rate (GBR) – used to set traffic policy for device groups and applications that are bit rate sensitive. GBR defines the minimum bit rate value allocated to a bearer. Bit rates higher than the GBR are possible if sufficient network resources are available.
Non-guaranteed bit rate (Non-GBR) - used to set traffic policy for device groups and applications that are not bit rate sensitive.
Delay-Critical Guaranteed bit rate (GBR) – for services that require both GBR and end-to-end network latency as low as 5 milliseconds. These bearer channels commonly cater to industrial internet of things (IIoT) implementations across a 5G LAN.
While choosing a bearer channel creates the framework for a microslice, it must still be defined and prioritized using network-based SLA parameters. In other words, once you’ve set a bearer type, such as GBR or non-GBR QoS class options must be chosen.
Microslicing uses the QoS Class Identifier (QCI) standard outlined by the 3GPP standards organization. This is done to simplify the process of setting the appropriate network quality parameters used in packet scheduling across the 5G RAN.
5G network slicing vs. 5G LAN microslicing
5G LAN microslicing and 5G network slicing, a 3GPP standard feature of the 5G technology widely used by mobile operators, are distinctly different. However, the distinctions between them are not well understood.
In the world of mobile network operators (MNOs), network slicing is an architecture that allows multiple virtual networks to be created over a common public or shared physical infrastructure using overlay techniques. From a 5G operator’s perspective, a network slice is an independent, end-to-end logical network that runs on a shared physical infrastructure that is capable of delivering an agreed-upon network service quality.
The customizable network slicing service options include transport speed, quality, latency, security, and reliability. These service levels are delivered based on service quality requirements negotiated between the mobile operator and the business customer. But this leaves enterprises with little or no control of the service quality.
With network slicing, the 5G operator is solely responsible for managing and orchestrating network resources to meet the customer SLA. From the carrier perspective, network slicing is a way to gain additional revenue opportunities for customers that wish to pay more for improved 5G service experiences.
In most cases, 5G operators create and manage network slices on a customer-by-customer basis. This means that all customer traffic traversing a public 5G network slice will receive identical application service levels regardless of the type of application or service that is being transmitted across the RAN.
In contrast, 5G LAN microslicing gives IT administrators complete control over end-to-end QoS, with the ability to create virtual microslices at a per-application level. MicroSlicing is one of the only ways to provide strict, SLA-backed traffic flow guarantees for each application across a wireless medium that can be directly integrated with the existing enterprise QoS framework already in place.
To accomplish granular control over microslice data flows, administrators can organize their RAN into application and device groups.
- Device groups – Administrators can place 5G endpoints that connect to the RAN via SIM card or eSIM into groups to logically segment them by device type. In many cases, the type of device often dictates what apps/services are being operated on a regular basis. This helps to coordinate which devices are assigned to a specific microslice.
- Applications – Applications can be configured so that the control plane can identify the data flow being transmitted across the RAN. Once the application is identified, the flow is assigned to the appropriate microslice and associated network performance SLAs.
In turn, explicit SLAs can be created that operate from one end of the 5G LAN to the other. Once data leaves the radio network and flows into the corporate LAN, the traffic flows can be automatically translated or mapped to the requisite enterprise VLAN ID or QoS policy settings that enterprise-grade routers and switches understand.
Undoubtedly, as organizations embrace 5G LAN technology to improve wireless coverage, eliminate mobility problems, and gain new levels of deterministic performance across both the RAN and LAN, new microslicing technology will play an essential role in making cellular technology a reality within the enterprise.
Srinivasan Balasubramanian is a Distinguished Member of the technical staff, Office of CTO, at Celona.