The 5G transport network poses great challenges to traditional measurement technologies.
The typical application scenarios like eMBB, uRLLC and mMTC in the 5G era require large bandwidth, low delay, large connection, network slicing and high precision time synchronization for the transport network. Diversified service types and differentiated network demands put forward higher requirements for Service Level Agreement (SLA). Operators provide differentiated services for new 5G services to add network values. How to guarantee users’ service experience will be a big challenge for 5G transport networks.
In the traditional network O&M field, the OAM technology based on ITU-T Y 1731 and the TWAMP measurement technology based on RFC 5357 have many limitations in detecting the performance parameters such as frame loss rate, frame delay, frame delay change and throughput, because they use the indirect monitor mode of transmitting extratest packets.
First, the traditional indirect measurement mode cannot ensure that the packets are consistent with the real path. OAM packet loss or disorder may affect the statistical results and cannot truly reflect the network QoS. Alarms and faults are reported only when the QoS degradation exceeds the threshold, and services are switched over to the protection path.
Second, the traditional information collection period is in the level of minutes, and the millisecond-level delay change and low-probability packet loss problems cannot be effectively located. Network faults can only be located passively, so the locating period is long and difficult.As the leader of the 5G transport network solution, ZTE has been actively following and promoting the formulation of standards and specifications for the 5G transport network since 2015 and has made great achievements in the research of key technologies. To meet the requirements of 5G transport O&M, ZTE introduces the Inband OAM technology to solve the deficiencies of the current OAM technology. Based on real service packets, the technology can implement end-to-end and point-to-point monitor of service delay and packet loss. The “per-packet” measurement and nanosecond-level delay measurement for real services in the entire network guarantee user experience.
Inband OAM Improves Service Measurement of 5G Transport Networks
Based on the inband measurement principle, the Inband OAM technology provides the end-to-end and hop-by-hop performance measurement capability for the transport network service flow, so as to rapidly detect the faults related to network performance and perform precise delimitation and troubleshooting. No extra OAM test packet is needed. Measurement information (including measurement instructions and data) is carried in the monitor service packets to detect inband performance.
Measurement Mechanism and Principle
- Measurement object
The object to be detected is the service flow. The service flow can be flexibly defined in accordance with service features, including L2, L3 and L4 features of the service.To simplify the service flow identification information, the service feature information can be mapped into a Flow ID.The service flow can be identified in multiple ways, including the destination IP address, source IP address, and DSCP. The service flow is identified at the ingress of the PE.
2. Packet loss measurement
The Inband OAM refers to RFC 8321 (Alternate-Marking Method for Passive and Hybrid Performance Monitoring). It is a CAS measurement technology that marks (dyes) the features of the actual service flow and detects packet loss and delay on the feature fields.
The above figure shows the packet loss measurement principle.
Transmitting end: Color the detected service flow label field alternately at a certain period, collect statistics of the service flow performance in the period, and report the statistics to the management and control system.
Receiving end: Collect statistics on the performance of the dyed detected service flow feature field within the same period as the transmitting end and report the statistics to the management and control plane. The statistics time at the receiving end should be in 1~2 periods to ensure that the out-of-order packets can be counted correctly.
According to the information of the detected service flow reported by the transmitting and receiving ends, the management and control plane calculates the number of packets lost in the period i: PacketLoss[i] = Tx[i]–Rx[i].
3. Daily measurement
The Inband OAM delay measurement principle is shown in the above figure:
Transmitting end: In each measurement period, one of the packets of the detected service flow is dyed with delay within this period, and the ingress timestamp t1/t3 of the packet is recorded and reported to the management and control plane.
Receiving end: Record the egress timestamp t2/t4 of the packets dyed with delay in the detected service flow in the same period as the transmitting end, and report the timestamp to the management and control plane.
The management and control plane calculates the unidirectional delay in both directions of the service flow in the period i according to the information reported by the transmitting and receiving ends: Delay[i] (in) = t2-t1, Delay[i] (out) = t4-t3. The unidirectional delay requires that 1588v2 time synchronization be deployed between the transmitting end and the receiving end.
In the scenario where the detected service flow is on the same path in both directions, the management and control plane calculates the bidirectional delay of the service flow in the period i according to the information reported by the transmitting and receiving ends: Delay[i] (double) = (t2-t1) + (t4-t3).
Inband OAM Completes the Verification in Existing Networks and Starts Intelligent O&M
In September 2019, ZTE completed the verification of the new inband OAM measurement technology in the existing network, proving that the Inband OAM technology can provide innovative support for the intelligent management, control, operation and maintenance of the 5G transport.
In this verification, ZTE verifies the combination of the Inband OAM technology and the intelligent management and control system UME in the existing network to detect and identify network exceptions one by one, accurately detect the performance such as delay and packet loss of each service, and locate faults so as to support fast service recovery. During the verification process, the millisecond-level fiber coiling distance is added with various network fault simulations such as delay monitoring and 0.5%/1%/5% packet loss probability, and the fault location and fault recovery capabilities are tested. The test results show that the Inband OAM solution can accurately locate a total of 24 S1 and cross-ring X2 traffic based on the real-time awareness of faults, and visually display the packet loss/delay index to identify the network fault point. This indicates that the Inband OAM technology is mature.
As a world-leading supplier of integrated communications solutions, ZTE cooperates with various industries to continuously test and evaluate 5G O&M management technologies, actively promote the commercialization process of the Inband measurement technology, laying a solid technical foundation for 5G transport networks.