Telecommunication networks are fundamentally about sending and receiving signals accurately. In order for the network to accurately juggle the complex transmission and reception of data from many devices in various locations, accurate timing and synchronization is absolutely crucial. When timing and sync is off, network performance and user experience can collapse.
“Timing alignment is fundamental in telecom,” said Kishan Shenoi, timing and synchronization architect for Intel, during a session at ATIS’ recent Workshop on Synchronization and Timing Systems (WSTS).
In early wireless networks, Shenoi said, synchronization was often referred to under the umbrella of frequency or time “accuracy”, with a focus on the impact on handovers. After all, the ability to move and maintain a connection is one of the defining and most important characteristic of a cellular network. Base stations have to be both synchronized relative to moving user equipment, accounting for Doppler shift, as well as coordinate with their neighbor base stations in order to facilitate smooth handoffs. Cellular networks also require synchronization to an external or absolute source of time across all network nodes, in order to avoid interference and to be able to successfully leverage the advanced RF features mentioned earlier.
This was all a lot simpler in previous generations of cellular technologies, when 1) Frequency Division Duplexing (FDD) was the primary mode in 3GPP systems, in which different frequencies are used for uplink and downlink communications, and 2) RF transmission and reception was less complex, relying on lower frequencies, single spectrum bands rather than carrier aggregation, and single-input-single-output or lower-order multiple-input-multiple-output antennas as opposed to modern massive MIMO systems.
Per Lindgren, group CTO and head of sync at Net Insight, which provides a network-based timing distribution solution, said that in 5G midband TDD systems, the accurate synchronization of both time and phase is much more complex. The move to widespread deployment of TDD spectrum for 5G by most major carriers “was the big shift, and I don’t think people realized what a big shift that was,” he reflected.
As systems have migrated to TDD and 5G, with increasingly advanced radio features, the range for allowable timing alignment error in cellular systems has tightened over the years. In CDMA systems, base stations had to maintain a TAE less than 10 microseconds. In modern LTE and 5G systems, Shenoi explained, TDD requirements for TAE can be less than 3 microseconds, and for some services in 5G clusters, the timing alignment error must be less than 130 nanoseconds. Meanwhile, timing source servers have been more widely distributed throughout the network to compensate for things like the fact that most operators lease at least some portion of their transport network and therefore do not control the end-to-end network.
Carrier aggregation, geolocation, Multimedia Broadcast Multicast Services (MBMS, or eMBMS/5G Broadcast), Inter Cell Interference Coordination (ICIC or enhanced ICIC) are all examples of cellular network features or services which deal with signal processing and must by necessity be properly synchronized in order for them to function as they should.
“The degradation with lack of synchronization is dramatic,” Shenoi warned.
Open RAN introduces additional timing and synchronization complexity by way of the disaggregation of RAN functions across fronthaul and midhaul connections in addition to the typical backhaul. In fact, in O-RAN, synchronization is such a vital concept that it is considered a separate plane: There is a data plane, control plane and the synchronization plane or S-plane. “It has been elevated, it’s not just an afterthought any more,” Shenoi said.
This blog continues my “Are Your Servers in Sync—is Network Timing the Synchronized Swimming of the Digital Age?” earlier blog about Timing and Synchronization, in which I made a real-life comparison between synchronized swimming in the Olympics and the importance of timing and synchronization in data networks. I also briefly mentioned new applications such as online gaming and VR as the applications needing such.
It all comes down to the accuracy of synchronized devices to ensure that whatever the data in question, devices are all in sync. There is a need for accurate and precise time measurements for any distributed system. PTP (Precision Time Protocol – IEEE v2) offers a solution to precisely synchronize clocks in such systems. These could be radios in mobile networks or the various database transactions within data center environments that I also touched on in my previous blog.
In a nutshell, it is all about having a Stratum clock, which is responsible for current, accurate time and sending a time reference to any device on the network that asks for it. If you’re unfamiliar, a Stratum clock is a Primary Reference Source, which could be a clock system employing direct control from Coordinated Universal time (UTC) frequency and time services such as a Global Positioning System (GPS). There are different levels of Stratum reference sources, labeled “0” to “4”, based on accuracy and connectivity. This reference time is transmitted to other devices on the network via a data packet known as the PTP sync message, allowing them to update their clocks. Essentially, one device on the network acts as a timekeeper for all other devices, which synchronize to this central device to achieve accurate time measurements. In other words, PTP deals with network latency by providing devices on a network with more precise and synchronized timing information. The nanosecond level accuracy provided by PTP helps to minimize the effects of network latency by reducing delays caused by differences in time measurements between devices.
Ethernet has long been the technology of choice both in the traditional computer IT (Information Technology) and automation OT (Operation Technology) networks. This open standard allows a variety of devices to be quickly and simply connected and exchange data. Ethernet, a 50-year-old technology, was not originally designed to meet the requirements of automation technology. More precisely, requirements around guaranteed real-time communication have led to various bus systems in the OT sector evolving over time, using Ethernet on a physical level while implementing proprietary real-time protocols on top. OT networks handling time-critical data traffic are separated from IT networks today due to the critical data traffic such networks need to handle. In the future, Industry 4.0 applications will require increasingly more consistent Ethernet networks, and Time Sensitive Networking (TSN) will allow the convergence of such IT and OT networks.
Link to California Triangle
Figure 1. IT and OT Separated Networks Pyramid Model
For effective real-time communication, it’s crucial to ensure consistent cycle times and minimize any fluctuation in these times across a variety of applications. This is particularly important in sectors like automation, which involves driving, controlling, snsing, and processing technologies. Data must be transferred in less than one millisecond in these contexts, necessitating guaranteed low latencies to maintain optimal performance.
Time Sensitive Networking (TSN) comprises Ethernet sub-standards, established by the IEEE 802.1 TSN Task Group, aimed at bridging Information Technology (IT) and Operational Technology (OT) through the enhancement and extension of current Ethernet standards. TSN aims to standardize features on OSI Layer 2 so that different protocols can share the same infrastructure. One prerequisite is that all network equipment has the same understanding of time. Thus, it is essential for all switches and end systems in the network to be time-synchronized.
In addition to the general IEEE specification, the TSN Task Group adopted a special profile that stipulates the use of IEEE specifications in conjunction with IEEE 802.1Q, and some elements needed to be modified, and as such the IEEE 802.1AS-rev. was created. A second core functionality that deals with transmitting critical and non-critical data traffic has been modified to handle buffer mechanisms in Ethernet switches, since a low-priority Ethernet data packet could delay those critical high-priority data packets. Hence, a new prioritization mechanism has been introduced to allow and regulate their coexistence. Additional traffic shaping or scheduling mechanisms can be implemented depending on the application requirement. These are the IEEE 802.1Qav Credit-Based Shaper and IEEE 802.1Qav Time Aware Scheduler.
Today, Ethernet networks for verticals such as manufacturing are based on a hierarchical pyramid model, which separates IT from OT. The IT part covers office communication with end devices such as printers and computers, while the OT part is based on systems, machines, and software used for process control and automation.
These two areas fundamentally differ in how they communicate, as IT depends more on bandwidth, while for OT, high availability is vital. As such, data traffic at the IT level is classified as non-critical, while data traffic at the OT level is classified as (time) critical.
As a result, IT uses communication standards such as Ethernet with TCP/IP, while OT uses various bus systems that meet requirements for guaranteed latency times. It is fair to say that with the above in mind, each control vendor usually promotes a specific fieldbus system, which means from the user perspective that selecting the controller dictates the selection of the bus type.[JH4] [MH5]
Previously, there were no connections between IT and OT. But now, as we transition towards Industry 4.0, the continuous flow of data has become crucial. Enterprises in certain sectors need reliable communication to gather operational data, access systems remotely, and connect machines to cloud services.
Industrial automation is currently being transformed to accommodate the demands for more adaptable and smart manufacturing, a change driven by the concepts of Industry 4.0 and the Internet of Things (IoT). This shift involves the development of intelligent production systems where components, machines, and entire factories are interconnected, allowing them to share information continuously. The aim is to enhance and streamline processes through automation.
For the benefit of integration, the traditional automation pyramid is transforming into a unified network, converging the IT and OT functions, including the critical applications such as sensors, directly into a converged Ethernet infrastructure, which can transport critical data simultaneously with non-critical data, without adverse reciprocal effects. Ethernet must cater to such environments, and as such traffic classification and the delivery of critical traffic must be assured with support for TSN on a converged Ethernet infrastructure.
Figure 2. Convergence on the IT and OT Networks
In the future, we will see more and more use cases and applications that need low latency and reliable delivery and, hence, need to be synchronized. As history has shown, this demand started from the Telco segment, moved into online gaming and VR (Virtual Reality), and later data-center infrastructure. Now we see the need arising in the digitalization of industry based on Industry 4.0. Different real-time requirements demand different approaches. TSN lays the foundation to meet these requirements and fulfill various latency, jitter, and reliability requirements.
Are you interested in learning more about Timing Synchronization for Mobile Networks? Contact us today to secure an expert consultation!