An overview of the major TSN standards(1)
Fancy Wang
Helping Global Enterprises Optimize Network Performance | Ethernet Card & Switch Solutions
Fancy Wang 1009 2020
The following content comes from network information
The key TSN features that provide guaranteed message delivery timing are time synchronization and traffic scheduling. They are addressed by the 802.1AS and 802.1Qbv standards, respectively. All devices participating in the TSN network are synchronized to a global time and are aware of a network schedule that dictates when prioritized messages will be forwarded from each switch. TSN makes use of multiple queues per port at the egress of the switch, where messages are held until a gate opens (at a time slot specified by the schedule) to release queued messages for transmission. The timed release of messages ensures that delays in the network can be deterministically predicted and managed. This allows for the convergence of critical traffic and noncritical traffic on the same network.
The preemption feature defined in the TSN 802.1Qbu standard can be used to increase the efficiency of bandwidth use for noncritical messages. In highly converged networks, it could be the case that large low-priority frames are delayed by higher-priority traffic on the network and dropped. Preemption enables the transmission of large frames to be interrupted, sent in smaller fragments and reassembled at the next link. This maximizes bandwidth utilization for all traffic types on the TSN network. Another important benefit of message preemption is the reduction of transmission latency for so-called Express traffic, which can preempt regular (lower-priority) Ethernet packets. Especially on lower-speed networks (e.g., 10 or 100 megabits per second (Mbps)) carrying large regular Ethernet packets up to 1,500 bytes and more, the latency reduction for Express traffic can be useful for building converged networks.
TSN provides a standard method for achieving seamless redundancy for industrial communication over Ethernet. The feature allows for the simultaneous transmission of duplicate message copies across different paths in the network. The first message copy to be received in time without error is processed, while the other copies are discarded. This adds another layer of determinism to the delivery of critical messages in converged networks.
A crucial feature of TSN is the support for open, vendor-independent network configuration. This is achieved through the standardization in IEEE of YANG models for various TSN standards. These can be configured over the NETCONF protocol using encoding formats such as XML or JSON. YANG models for bridging, traffic scheduling, frame preemption, seamless redundancy, and policing ensure that configuration of key TSN features is done according to standard methods. This allows TSN networks to be composed of any standard compliant device from any vendor and can be configured by any standard compliant network configuration software.
TSN Profiles
TSN is essentially a toolbox of features that address various needs such as reliability, bounded low latency, time synchronization, and resource management. These capabilities are realized through various TSN specifications (e.g., IEEE 802.1AS-Rev, IEEE 802.1Qbv, etc.), and customers can choose the relevant standards to implement based on their specific application needs.
Profile standards are being specified for some of them to describe which TSN standards to use and how. A TSN profile selects features, options, configurations, and protocols to build a bridged network for the given TSN application. Table 5-3 shows a list of select TSN profiles that are currently being defined.
802.1AS/AS-Rev
Enhanced Generic Precise Timing Protocol: Timing and synchronization are vital mechanisms for achieving deterministic communication. 802.1AS is a profile of the IEEE 1588 PTP (Precision Time Protocol) synchronization protocol that enables synchronization compatibility between different TSN devices (Figure 5-1). This lays the foundation for the scheduling of traffic through each participating network device. 802.1AS-Rev is being defined to add support for fault tolerance and multiple active synchronization masters (Figure 5-2). Multiple clock-masters for redundancy enable high availability of TSN networks – in cases when a grandmaster becomes faulty, system elements such as end nodes and bridges are still able to remain synchronized by obtaining the timing information from the redundant grandmasters. 802.1AS-Rev is also a profile of the IEEE 1588 PTP synchronization protocol.
Open image in new window
Figure 5-1
802.1AS operation3
Open image in new window
Figure 5-2
802.1AS-Rev operation4
802.1Qbv
Time-Aware Shaper: Scheduling of traffic is a core concept in TSN. Based on the shared global time provided by 802.1AS, a schedule is created and distributed between participating network devices. 802.1Qbv defines the mechanisms for controlling the flow of queued traffic through gates at the egress of a TSN switch (Figure 5-3). Frames are assigned to queues based on Quality of Service (QoS) priority. The transmission of messages from these queues is executed during scheduled time windows. Other queues will typically be blocked from transmission during these time windows, therefore removing the chance of scheduled traffic being impeded by nonscheduled traffic. In other words, there is a gate in front of each queue which opens at a specific point of time which is reserved for that queue. This means that the delay through each switch is deterministic and that message latency through a network of TSN-enabled components can be guaranteed. The IEEE 802.1Qbv standard defines up to eight queues per port for forwarding traffic. The scheduler is designed to separate the communication on the Ethernet network into fixed length, repeating time cycles.
Figure 5-3 shows an example with four queues, with a cycle time of td and guard band of tg. At time t0, the time-critical data queue, Queue 3 is open. Once that frame is transmitted, the best effort Queues 0, 1, and 2 are opened. Before the end of the cycle, at time t0-tg, all the non-time-critical data is blocked, so that the port is free to transmit the time-critical data at the start of the next cycle. This is essentially a time-division multiple access .