Name: Liu Jiamu
Student ID: 19011210553
The information comes from the Internet, edited and organized by myself
Embedded Bull Introduction: Time Sensitive Network The goal is to realize the simultaneous transmission of real-time critical data flows and ordinary data flows in the same network with good compatibility. To achieve the integration of these two services, devices in time-sensitive networks are required to accurately control schedules and achieve the low latency and low jitter required by real-time critical services. In addition, if the intricate business flows of various types of equipment can be transmitted on the same network, this means the reduction of dedicated network connections, simplifying the deployment process of system equipment, and at the same time reducing the size and cost of system equipment.
Embed Niu Nose: Time-sensitive network? TSN deterministic transmission
Embed Niu Nose: Time-sensitive network related issues that have been raised in recent years
Embed Niu Text ：
1. Question
The forwarding mode usually adopted by traditional Ethernet is "Best Effort", but this forwarding mode often lacks certainty. When a data packet arrives at the sending port and is ready to be sent, the sending end forwards it on a first-in-first-out basis. However, when a sending port has multiple data to send at the same time, the data will be queued and the waiting time will be Determined by multiple factors such as queue length and data sending speed. If the traffic in the network is too large, congestion or packet loss will occur, the queuing time will become unpredictable, and certainty cannot be guaranteed. This will lead to the standardization of traffic scheduling, time synchronization, traffic monitoring, and fault tolerance mechanisms. etc. questions.
With sufficient bandwidth, this best-effort Ethernet can be adapted to most current situations, but in some application areas this uncertainty is intolerable, such as telemedicine. Or network-assisted autonomous driving. In these security or life-critical network applications, uncertainty in a certain information transmission may bring irreparable consequences.
At this time, establishing a reliable transmission mechanism has become the primary issue facing technicians.
In order to ensure the deterministic behavior of some of the more important controlled physical systems, the real-time network needs to have deterministic and low network delay and delay variation (jitter). Traditionally, fieldbuses have been used for this purpose, but due to various factors such as bus design, cost, volume, weight, etc., time-sensitive networks began to be proposed.
Time Sensitive Networking (TSN) is based on standard Ethernet. Communication traffic on standard Ethernet (such as audio and video streams) can share the physical network with high-priority deterministic information flows (such as motion control). Different services have different latency requirements, especially in downlink service areas that require deterministic transmission, which are particularly sensitive to latency and jitter requirements.
The goal of time-sensitive networks is to achieve synchronous transmission of real-time critical data flows and ordinary data flows in the same network with good compatibility. To achieve the integration of these two services, devices in time-sensitive networks are required to accurately control schedules and achieve the low latency and low jitter required by real-time critical services. In addition, if the intricate business flows of various types of equipment can be transmitted on the same network, this means the reduction of dedicated network connections, simplifying the deployment process of system equipment, and at the same time reducing the size and cost of system equipment.
TSN does not cover the entire network, but only the definition of the MAC layer and the process of processing data frames.
2. Content History
AVB - Ethernet Audio Video Bridging technology (Ethernet Audio Video Bridging) is a set of real-time audio and video based on the new Ethernet architecture that the IEEE 802.1 task group began to develop in 2005. Transport protocol set. It effectively solves the problems of timing, low latency and traffic shaping of data in Ethernet transmission. At the same time, it maintains 100% backward compatibility with traditional Ethernet, and is a next-generation network audio and video real-time transmission technology with great development potential.
These include:
1. 802.1AS: Precision Time Protocol (PTP)
2. 802.1Qat: Stream Reservation Protocol (SRP) )
3. 802.1Qav: Queuing and Forwarding Protocol (Qav)
4. 802.1BA: Audio Video Bridging Systems
p>5. 1722: Audio/Video Bridging Transport Protocol (AVBTP)
6. 1733: Real-Time Transport Protocol (RTP)< /p>
7. 1722.1: Responsible for device search, enumeration, connection management, and mutual control between 1722-based devices.
AVB can transmit not only audio but also video. When used for audio transmission, in a 1G network, AVB will automatically use 750M of bandwidth to transmit bidirectional 420-channel high-quality, uncompressed professional audio through the bandwidth reservation protocol. The remaining 250M bandwidth can still transmit some non-real-time network data. When used for video transmission, the reserved bandwidth can be adjusted according to specific applications. For example: 750M bandwidth can easily transmit high-definition full HD visually lossless video signals. And can be routed arbitrarily in the AVB network.
The IEEE 802.1 Task Group officially changed the name of AVB to TSN – Time Sensitive Network in November 2012. In other words, AVB is just one application in TSN.
The first application is our professional audio and video (Pro AV). The emphasis in this application area is on the master clock frequency. In other words, all audio and video network nodes must follow the time synchronization mechanism.
The second application is in the field of automotive control. Most current automotive control systems are very complex. For example: brakes, engines, suspensions, etc. use CAN bus. The lights, doors, remote control, etc. use LIN system. Entertainment systems are even more diverse, including current in-vehicle networks such as FlexRay and MOST. In fact, all the above systems can be unified managed using TSN that supports low latency and has a real-time transmission mechanism. It can reduce the cost and complexity of adding network functions to automobiles and professional A/V equipment.
The third application is in the field of commercial electronics. For example, if you are sitting at home, you can connect to any electronic device at home through wireless WIFI and browse any audio and video materials in real time.
The last application is also the most widespread application in the future. All industrial fields that require real-time monitoring or real-time feedback require TSN networks. For example: robotics industry, deep-sea oil drilling and banking industry, etc. TSN can also be used for data transfer between servers that support big data. The global industry has entered the era of Internet of Things (IoT). There is no doubt that TSN is the best way to improve the interconnection efficiency of the Internet of Things.
3. Research status and hot spots
TSN is being widely adopted in critical small closed automotive and industrial networks to establish reliable ULL end-to-end connections. However, the key TSN limitations focus precisely on closed networks, such as vehicular networks and small-scale robotic networks. Network applications running in robotic and vehicular networks often involve significant interaction with external non-TSN networks. Robotics and vehicular networking applications need to be tightly integrated with mobility handlers via external networks. If advanced network features (such as mobility) are not properly supported in the external network, the benefits of TSN are essentially limited to small closed networks. Therefore, smooth interoperability between TSN and different external networks is essential for TSN operation in heterogeneous network scenarios.
Ideally, the connectivity between TSN and non-TSN networks should be able to accommodate similar characteristics as TSN to ensure overall end-to-end connectivity requirements in heterogeneous deployments.
V2X communication: Lee and Park proposed iTSN, a new method of interconnecting large TSN networks for large-scale applications. The iTSN approach utilizes wireless protocols such as IEEE 802.11p for Internet between different TSN networks. In particular, sharing global timing and synchronization information across interconnected networks is important for establishing a public timing platform to support TSN features in external networks. The iTSN approach thus enables, for example, vehicular networks to send safety-critical information to control nodes, such as roadside units (RSUs), with microsecond latency in heterogeneous deployments. By employing this reliable interconnect technology, vehicle braking safety distances can be achieved in much shorter (microsecond) time spans than currently feasible in the millisecond range. Overall, TSN and interconnection technologies such as iTSN can create a communication platform for safe autonomous driving systems.
Network modeling: Although TSN standards have received great attention in automotive driving networks, a major challenge in network deployment is how to manage the complexity of the network. With the advancement of technology, the automotive industry has put forward more requirements for the existing in-vehicle network infrastructure. As the number of sensors in the vehicle network increases, the increasing connections between sensors should be met accordingly in network planning and bandwidth requirements. However, dynamic changes in in-vehicle control system network requirements may require more extensive network infrastructure, resulting in higher expenses.
Hardware and software design: The design of hardware and software components to support TSN functions, such as scheduling, preemption and time-triggered event generation in TSN nodes, requires a lot of engineering and development work. Hardware implementations are highly efficient in terms of computational resource utilization and execution latency, but result in rigid architectures that are difficult to adapt to new application requirements. Software implementations, on the other hand, can flexibly adapt to new application requirements but may overload the CPU due to the softwareization of network functions, such as time-triggered scheduling and hardware virtualization.
Summary and lessons learned: To date, most research on TSN has focused on in-vehicle networks that are independent and isolated from external networks. Another limitation in the field of TSN research is the lack of simulation frameworks that encompass large-scale heterogeneous network architectures. Valid use cases including local and external network interactions (e.g. car driving) should be created and considered in the baseline evaluation. Currently, the general use cases in most TSN research are in-vehicle networks supporting in-vehicle sensor connectivity and audio/video transmission for infotainment. Future customized TSN simulation frameworks should be based on networks that support next-generation applications with localized and external network interactions, such as car driving. Similarly, SDN-based TSN management can leverage a layered controller design to extend management from localized networks such as vehicle networks to external networks such as vehicle-to-any (V2X) networks .
4. Next research trends
TSN network infrastructure and protocols must support limited end-to-end latency and reliability to support basic functions related to critical applications in IoT, medicine, automotive driving, and smart homes . TSN-based solutions for meeting these application requirements result in complex network infrastructure supporting various protocols. Therefore, a simplified TSN network management mechanism is crucial to reduce complexity while meeting the key requirements of ULL applications.
Therefore, reliable, secure and low-latency communication between multiple TSN networks is critical to support a wide range of future applications. The lack of TSN standards for connecting and communicating with external TSN and non-TSN networks hinders research activities in interoperable networks and needs to be addressed urgently. In summary, we identify the following major future design requirements for TSN research:
① Support a variety of applications from time-sensitive to delay-tolerant applications with traffic scheduling capabilities.
② Connections between multiple closed TSN architectures.
③ Flexible and dynamic priority allocation to ensure limited end-to-end delay for lower priority traffic.
④ Use SDN to centrally manage TSN functions from a global network perspective.
⑤ Achieve efficient timing information sharing and accurate clock design through self-estimation and local clock deviation correction.
⑥ Computationally efficient hardware and software design.
1. Transmission of low-priority data in TSN
TSN nodes preempt ongoing low-priority frame transmission and are used to send incoming high-priority frames to ensure high priority The absolute minimum TSN node transmission delay for frames. Depending on the intensity of high-priority traffic, low-priority frames can be preempted multiple times. As a result, since preemption events directly depend on high-priority traffic intensity, the end-to-end delay characteristics of low-priority traffic cannot be guaranteed. If the intensity of high-priority services is significantly higher than that of low-priority services, the end-to-end delay of low-priority services can be greatly increased. Typically, low-priority traffic carries latency-sensitive data, which is not as critical as high-priority traffic data, but should still be delivered within the worst-case deadline. In the current state of the art, there are no research mechanisms or standards to ensure the worst-case end-to-end latency of low-priority traffic under preemption.
Therefore, future research needs to develop new mechanisms to ensure bounded worst-case delays for low-priority traffic in TSN networks.
2. Development of Wireless TSN
p>In order to wirelessly connect industrial equipment (industrial sensors/actuators) to TSN networks, 5G is a very suitable solution. Compared with 4G, 5G's new features, especially the radio access network (RAN), provide better reliability and transmission latency. Moreover, the new 5G system architecture allows for flexible deployment. Therefore, 5G can realize TSN networks that are not restricted by cable installation.