5G Network Slicing using FlexE at the Optical Layer
By Tim Warland | 830 words | 3.5 minute read
5G networks will introduce numerous technologies, including new radio (5GNR) with beamforming capabilities for higher density per tower, new frequency spectrum allowing higher data-rates, and network slicing to guaranty end-to-end performance. Of these, network slicing provides a mechanism for network operators and carriers to monetize their existing optical infrastructure. This post provides an overview of edge computing for low latency but focuses on network slicing and the underlying protocol that makes it successful and governs its implementation.
Applications such as autonomous vehicles, remote surgery and immersive mobile VR and AR rely on the enhanced performance and low latency delivered by 5G networks. One achieves Low latency through a combination of mobile-edge computing (MEC, also known as Multi-access edge computing) and network slicing. Network slicing will have a profound effect on the way carriers and network operators manage and monetize their networks and on the end-user experience.
Mobile edge computing processes information local to the radio-head and is beneficial for V2X (vehicle to anything) communications and other IoT applications. Edge computing allows data to be processed closer to where it is created instead of sending it across long routes to data centres or the cloud. By running applications closer to where one generates the data, you improve responsivity. A relevant example exists with autonomous vehicles: Multiple vehicles communicate their location and movement with the MEC, and the MEC can manage traffic flow. There’s no reason to transport this data to the cloud – the MEC provides quicker reaction time. Similarly, AR & VR applications can use edge computing as a local video cache, so that image movement appears more fluid to the user.
When communications need to traverse beyond the local region, this is an opportunity for network slicing to achieve low latency to access cloud or other remote computing applications. Edge compute nodes and cloud infrastructure are connected using optical fibre for high reliability. Signals on the fibre travel close to the speed of light and the IEEE continues to evolve the volume of data that are carried on a given optical link. But it’s the FlexE specification that allows network slicing to achieve the quality of service (QoS) objectives.
Network slicing provides different classes of service over the same optical connection. For example, low-latency and time-sensitive information can be prioritized for transmission over a shared optical medium. Consider a surgery application where a doctor located in Kingston Ontario is operating remotely on a patient located in Ottawa, a distance of 200km. Optical latency accounts for about 5ms, but in a traditional network, these time-critical packets could get stalled behind other data (like the cat videos everyone is streaming). Because all data is treated equally, this can result in excess delay and a loss of fidelity during operation – the doctor wouldn’t get the right tactile response. Network slicing that uses the FlexE protocol can be implemented to achieve low latency by prioritizing this traffic. It would provide high responsivity; thereby, it would make the remote surgery a practical solution.
The FlexE specification allows multiple channels to be bonded together to create higher bandwidth connections. This innovation is relevant to clients who need to dynamically increase and decrease bandwidth, for example, to broadcast a sporting event. Sporting events need massive bandwidth to support broadcast and other services, but they only need them for a couple of hours a week. With network slicing, the network operator can assign (and bill) for the bandwidth when it’s required and re-assign that bandwidth when it’s not. This capability makes for a more efficient network without having to over-provision the amount of bandwidth available in the region. Furthermore, this means that you can monetize network slices, and sell network slices as “Network Slice as a Service.”
The concept of Network Slicing can be implemented based on the (OSI) Layer 1 Flexible Ethernet (FlexE) specification developed by the Optical Internetworking Forum (OIF). FlexE provides deterministic transport flow control to map packet services onto optical transmission paths. The FlexE gearbox creates fixed bandwidth connectivity with the ability to scale bandwidth up and down as required for each network slice. It essentially acts as a clutch mechanism between the IP slices and the optical wavelength.
FlexE for network slicing is based on a concept I presented earlier this decade to support flexible mapping for Ethernet. I attended the OIF 3Q2019 Plenary in Montreal, Canada and introduced an amendment to make the FlexE specification more robust. Considering that FlexE is a protocol that allows the delivery of low-latency services like remote surgery, these newly created links must be robust.
It’s magnificent to see the industry take a specification that was designed to connect multiple router ports and make amendments to enable network slicing using FlexE. It’s been over a decade since I made significant contributions to the industry standards that govern our communications infrastructure. It’s important to note that these contributions have not only changed the way we communicate today but set the path for how networks will operate in the future. And, it’s rewarding to continue to contribute to the industry.
Tim Warland is an innovative and conscientious risk-taker with 20+ years in high-tech management and engineering. He has developed, marketed and sold products in high-risk, high-reward B2B markets. He has also contributed to the development of international standards, defined product/market requirements, engaged lead-customers, managed product development and supported customer education and bring-up for two generations of Ethernet products supporting 10G and 100G leading to high return on investment. He is Canada’s #1 leader in Ethernet physical layer products.
Tim has been providing consulting services since 2003. He specializes in supporting complex technology products through market research, marketing collateral (white papers, blogs, application notes, press releases, etc), and sales training. He helps client organizations articulate value proposition and product differentiation. Currently, Tim works with Invest Ottawa as Manager, Venture Programs, on the Next Generation Network Program and the ENCQOR 5G testbed.