[Intern] 12/09/2022 Accelerating O-RAN fronthaul with DPDK - Shahaf Shuler & Dotan Levi, NVIDIA.

# [Intern] 12/09/2022 Accelerating O-RAN fronthaul with DPDK - Shahaf Shuler & Dotan Levi, NVIDIA. ###### tags: `BMW-Lab`, `Intern` :::success **Goal:** To summary the Accelerating O-RAN fronthaul with DPDK - Shahaf Shuler & Dotan Levi, NVIDIA. ::: :::success **References** - [Accelerating O-RAN fronthaul](https://youtu.be/uY574MQxrTc) ::: ## Background ## 5G Fronthaul Fronthaul is defined as the fiber-based connection in RAN infrastructure between the Baseband Unit (BBU) and Remote Radio Head (RRH). Fronthaul originated with LTE networks when operators first moved their radios closer to the antennas. This new link was established to supplement the backhaul connection between the BBU and the central network core.In a 5G network, flexible fronthaul configurations have become an essential ingredient for balancing the latency, throughput and reliability demands of advanced 5G applications.Next-generation RAN is resulting in increased fronthaul fiber deployment and a greater reliance on multiplexing, virtualization and split fronthaul architecture. This has placed eCPRI fronthaul among the most important 5G technologies in operator surveys. ![](https://imgur.com/49NUPSe.png) ## PTP Hardware Clock PTP Hardware Clock (PHC) subsystem infrastructure and the uses of the PHC as an interface to both software-based and hardware-based servos that provide synchronization interfaces required to meet ITU-T network limits for 5G Radio Access Network (RAN).It also uses hardware timestamping. ## PHC The phc container is responsible for synchronizing the two available clocks in a cluster node, typically these are the PHC and the system clocks. This program is used when hardware time stamping is configured. It usually synchronizes the system clock from the PTP hardware clock on the defined network interface controller (NIC). ![](https://imgur.com/Hsm4Uyo.png) ## eCPRI protocol evolved Common Public Radio Interface (eCPRI) is a protocol, which will be used in fronthaul transport network. It will be included in standard Ethernet frames and UDP frames. ## Time Division Multiplexing Time Division Multiplexing (TDM) is a technique for the serial transmission of user data over a common medium such as a coaxial cable. At a time, only one user's data are transmitted serially in a time slot. TDM allows each user to use the entire system bandwidth. ## Time-synchronization technology in mobile communication Synchronization technology is essential for data transmission between communication systems and has been introduced into many telecommunication operators’ networks. Conventionally, network carriers offering services such as telephones and leased lines have established synchronous networks and have synchronized the clock frequencies of the devices, thereby multiplexing and separating data and providing high-quality services. Synchronization technology has been important even for packet-based asynchronous networks and is required for efficient data communication for fourth-generation (4G) and 5G mobile networks. Thus, synchronization technology is important in mobile communication as well as in fixed communication and is one of the basic technologies in realizing the network services of telecommunication operators. Synchronization technology is mainly classified as frequency synchronization and phase/time synchronization. The state in which the clock frequencies of different systems match is called frequency synchronization, and the state in which the timings between the clocks agree is called phase synchronization. In particular, when the clock timing is synchronized with Coordinated Universal Time (UTC),*1 that state is defined as time synchronization. Time-synchronization technology for communication services is currently used for time-division multiplex communication based on Long-Term Evolution (LTE), and fixing the time synchronization between UTC and mobile base stations contributes to improving the utilization efficiency of the frequency band at the base stations. In the 5G era, more efficient use of bandwidth is necessary to handle the increasing amounts of data traffic, and advanced communication methods for improving communication quality are required in order to address the diversification of end applications. High-precision time synchronization for temporal coordination and control between base stations is also required. Representative example technologies are dual connectivity (i.e., communicating by using multiple base stations) and carrier aggregation (i.e., bundling a large number of carrier frequencies as one channel). Together with the standardization of such mobile technologies, standardization of the synchronization technologies that support them is under discussion. *1 UTC: Time managed to continuously maintain the difference between global time (based on the earth’s rotation) and international atomic time (based on the time standards of the standards agencies of each country) within 0.9 seconds by inserting leap seconds, etc. ## Technology for achieving high-precision time/frequency synchronization Precision Time Protocol (PTP)*2 is a protocol for achieving time synchronization based on time-stamp information stamped on packets. PTP has attracted attention in recent years because time synchronization with UTC is possible with microsecond- to nanosecond-order precision. Therefore, standardization of PTP as a candidate protocol for time synchronization of future 5G mobile networks is underway. PTP generally uses UTC information received from a global navigation satellite system (GNSS), typified by the GPS (Global Positioning System), as the time reference. It is possible to transmit and synchronize that time information—with the installation site of the GNSS antenna as the reference point—by utilizing the network. Frequency synchronization, represented by Synchronous Ethernet (Sync-E), is also commonly used as a backup for time synchronization. If PTP packets are missing (due to network congestion, etc.), and time synchronization is not possible, time synchronization can be maintained for a certain period of time by applying frequency synchronization. *2 PTP: A protocol to calculate the time shift of a slave (subordinate device) relative to a master (superior device) based on time-stamp information and round-trip-delay information stamped in a dedicated time-synchronization packet and synchronize the calculated time with the master clock. ## Relationships between standardization organizations in terms of time/frequency-synchronization technology Standardization bodies are addressing various issues related to a frequency-, phase-, and time-synchronization technologies including PTP, and the relationships between these organizations in terms of these technologies. ## Requirements concerning time and frequency synchronization for 5G mobile communication At the ITU-T SG15 meeting in February 2018, a technical report (GSTR-TN5G) on transport networks for 5G was given consent. That report also includes network architecture and accuracy requirements concerning time- and frequency-synchronization technologies. The discussion topics concerning general network configurations and standardization for time and frequency synchronization are shown in Fig. 2. The time-synchronization network is configured with a primary reference time clock (PRTC)—a time-reference device that receives time from a GNSS, a telecom-boundary clock (T-BC)*3 for receiving, synchronizing, and transmitting time information from the PRTC, and a telecom-time slave clock (T-TSC)*4 for supplying the time from the T-BC to the end application. The frequency-synchronization network supporting time synchronization consists of a primary reference clock (PRC)—a frequency-reference device—and a synchronous Ethernet equipment clock (EEC), which synchronizes and distributes frequency by Sync-E. In addition, the synchronization-network management system monitors and controls these time- and frequency-synchronization devices. Accuracy is cited as the main requirement of time synchronization. In addition to the absolute time error defined in terms of error with respect to UTC, in recent years, relative time error between base stations is also being considered under the assumption of 5G applications. The time-synchronization accuracy requirements of major mobile applications being discussed and regulated by the 3GPP and IEEE . Up to now, standardization has been advanced with the aim of achieving absolute time-synchronization accuracy of 3 μs for TD-LTE (time-division duplexing LTE). However, in recent years, higher precision (namely, a minimum of 65 ns) is being required as a relative time error for carrier aggregation and other technologies. Therefore, to ensure high-precision and stable operation of various synchronous devices for providing time information to base stations, high reliability as well as management capabilities are required. In the following section, standardization trends concerning these topics are introduced. ## Conclusion 5G will unleash unparalleled opportunities for mobile network operators (MNOs) and telecom OEMs. However, the immense possibilities and opportunities require timing synchronization capabilities that were previously built into the circuit-switched equipment used in 3G and 4G networks. To support fully packet-based voice and other real-time applications, 5G networks will need packet synchronization with very high accuracy. In addition, other infrastructure applications including transmit diversity on MIMO antennas or certain carrier aggregation implementations also need time synchronization. The 5G standard for these applications has rigorous requirements to support end-to-end latency and increased antenna and base station density. Accurate packet timing synchronization is critical to enabling a quality experience for real-time applications. This is driving the need for specialized network adapter cards that can synchronize timing with nanosecond accuracy for the most demanding real-time applications and services.