Mobile Networks: Tutorial
Time sensitive networking is a catch all term which is focused on taking an inherently non-deterministic Ethernet network and making it deterministic with minimum latency. Ethernet’s lack of determinism comes about because it is a shared medium – traffic from different sources shares the same switching and routing resources on the way to its destination, much like the cars driving through a city use the same roads. When traffic arrives at a device – e.g. a switch or a router, it has to wait behind other traffic scheduled to go out on the same port. How long it has to wait depends on the length of the queue, which will vary depending on the amount of traffic in the system. If the queue becomes too large, buffers will overflow and traffic will be dropped.
Until recently, the primary users of Ethernet for time critical operations were the telecommunications industry and in particular the mobile backhaul. It is now becoming critical for a growing list of applications – for example, audio-visual broadcast (AVB), the automotive industry where Ethernet is replacing heavy and expensive cable harnesses, industrial automation where Ethernet is replacing more traditional interconnect technology and the so called Internet of Things (IoT) where potentially thousands of sensors are interconnected providing data on a myriad of things.
To achieve this, additional Ethernet standards have been developed to address problems in key areas. These are minimising latency – getting packets from one end of the network to the other in the minimum possible time, allowing some packets to pre-empt others by jumping the queue – even over packets that have already started being transmitted, guaranteeing that certain traffic can be delivered at specific times by clearing the line – also known as time aware shaping, guaranteeing bandwidth – known as stream reservation and underlying much of this, synchronization – specifically through developing a new profile for PTP, known as IEEE 802.1AS.