While developing our value proposition and our SAAS & AI-based solution, we pondered a lot about which consumer or industrial use cases (or innovative service) are required now or will soon require ultra-low latency (below 15 milliseconds) combine with very high reliability.
New categories of applications require high network reliability and very low latency communication to be able to achieve tasks in network. According to , the emerging 5G is designed to support enhanced Mobile Broadband (eMMB), Ultra-Reliable Low Latency Communication (URLLC) and massive Machine Type Communication (mMTC). URLLC communication ensure at the same time a very high network reliability and low latency communication, that claims most of network applications. Using the 5G communication, mobile provider network will offer high performance networks to their customer and will meets applications requirements.
The table below shows some URLLC use cases with their requirements in terms of latency and reliability compiled by our team  through academic research .
If we zoom in on Factory Automation, this use case often describes the industry 4.0 paradigm which enables interconnection and communication between machines, devices, sensors and people. High reliability and a guaranty of low latency is often required in an industrial manufacturing . The figure below shows the scenario of a factory automation use case where a robotic hand sends latest data on its action and context and in return receives command(s) from remote server control unit on the next actions to perform. This approach provides increased production flexibility compared to a pre-programmed robot that only perform repetitive batch manufacturing tasks. However, for this to result in high production quality, flawless connectivity with sustained low-latency is required.
Maintaining 5G QoS such as sustained ultra-low latency, in order fulfill B2B customer SLA is a puzzle game for the provider. It is not possible anymore to just “throw extra bandwidth” to the problem and hope it will fix itself. This strategy, proven useful in the last 30 years with IP networks, will not be sufficient with ultra-low latency (<15ms) and very high reliability (>99,999%) requirements. For us, the most efficient and effective way to ensure that QoS can be maintained is by using advanced analytics and performing the real time network monitoring.
By advanced analytics, we envision the following features (all in real-time as dealing with low-latency cannot wait):
- Real-time AI-based anomaly detection of ultra-low latency metric per application or network site
- Real-time decomposition analysis of ultra-low latency into its sub-components such as propagation delay, transmission delay, jitter, etc. We will soon publish a specific blog post on this topic.
- Real-time AI-based prediction & trend analysis of end-to-end latency and reliability
- Real-time comparative analysis between 5G networks, sites, frequency, network protocols, B2B use case, etc.
We have started to implement the above functions and would be happy to perform a live demo of our latest prototype.
- Njakarison Menja Randriamasinoro, Low Latency communication, Monitoring, Measurement and estimation, LatenceTech WhitePaper, 2020.
- X. Jiang, H. Shokri-Ghadikolaei, G. Fodor, E. Modiano, Z. Pang, M. Zorzi, and C. Fischione.
Low-latency networking: Where latency lurks and how to tame it. Proceedings
of the IEEE, 107(2):280-306, 2019.