Network functions virtualization (NFV)-based 5G networks have already caught the fancy of the telco cloud service provider, enticed by the promise of high quality of service (QoS) levels.
These providers can expect high speed/throughput, high reliability, low latency, increased capacity, availability and connectivity, and dynamic bandwidth allocation from the 5G cloud radio access networks (RAN) and NFV core networks. These are the QoS attributes that will be consumed by, and will characterize, the 5G services that they serve:
- eMBB (enhanced mobile broadband) – this requires high capacity, high speed and dynamic bandwidth allocation. While high speed helps to upload and download video-based content faster and in gigabyte sizes, bandwidth can be allocated on demand for ultra high-definition (UHD) video, virtual reality, augmented reality, etc.
- uRLLC (ultra-reliable and low-latency communications) – these services require high reliability, high availability and low latency. High reliability supports mission-critical services such as connected robotic factories and remote surgery, while low latency makes delay-critical services such as autonomous cars and remote-controlled drones a reality.
- mMTC (massive machine-type communications) – high capacity, high speed, dynamic bandwidth allocation characterize this class of services. It is most suited for internet of things (IoT) services like smart cities, serving billions of ‘things’ and ‘devices’ up to a density one million per square kilometer.
How will QoS improve with 5G?
Here is a quick round-up on the promised values of the 5G QoS parameters:
- Speed/throughput of 10Gbps, and a target of 20Gbps
- End-to-end latency reduced to a range of 1-10 milliseconds and to 0.5 milliseconds for the physical network
- Network availability stretched to 100%
- Reliability parameters like block error ratio for 5G expected to be 0.00001 in a 1 millisecond period, compared to 0.01 in 4G
- Jitter brought down to 10-100 microseconds
- Bandwidth from 100Kbps for small sensor devices to several hundred megabits per second for industrial robotic cameras
To achieve these ambitious QoS values there is tremendous work being done on 5G MIMO antennas, 5G Cloud RAN and the NFV core network. However, as we know from experience, physical radio access and air interface pose the real challenge and will continue to remain an obstacle for faster propagation and transmission. Efforts are being made to reduce packet size, transmission time and propagation time to as low as possible.
As exabyte-scale data generated by the 5G services traverses the 5G network and/or its complementary 4G and Wi-Fi networks, it needs to be monitored in real time for associated QoS guarantees at all points of time. A technique which is unique to 5G helps in achieving the QoS guarantees to a large extent: slicing of 5G network functions (RAN and core) so that network and service resources can be dedicated and negotiated (scaling in, scaling out) dynamically.
Assurance must cover 5G network slices
Even as these QoS challenges are being worked on, the responsibility lies with the 5G CSP to assure latency, reliability and connectivity on a 24/7 basis to prevent any degradation of these critical QoS parameters. Assurance systems are needed to extend assurance to 5G slices and sub-slices, not just to the entire network. A sliced, secure always-available network with built-in redundancy will allow seamless data handovers between 5G, LTE and WiFi networks, thus assuring mobility continuity, coordination and availability.
It is important to note that many 5G services, if not offered with guaranteed QoS, will land the telco cloud service providers in serious commercial and legal problems. Imagine a drone dropping from the sky, a driverless car veering off the motorway, a stalled robot in the production line or a video frame freezing in the middle of a remote surgery.
Telco cloud service providers are conscious that delay-critical services are highly dependent on the 5G network QoS and the management of its network slices. They know that it requires redoubled effort on the 5G CSP’s part to analyze and action network/service data quickly to keep capacity, latency, reliability, availability and connectivity under close surveillance and offer proactive remediation.
Through the use of key assurance techniques such as proactive closed-loop automation, service assurance and service orchestration as well as automated root-cause analysis among others, telco cloud service providers can begin to offer the challenging QoS values that 5G promises.