The rise of Open RAN, where multi-vendor radio systems comprising centralized, distributed and radio units are made interoperable, is now decidedly mainstream. Following early commercial deployments from greenfield operators, Open RAN is scaling out in brownfield networks. In addition to the pre-deployment interoperability work, there’s also a need for tight focus on test and measurement for timing and synchronization given the architectural shifts brought about by radio access network (RAN) disaggregation.
During the recent Test and Measurement Forum (available here on demand), VIAVI Solutions Senior Manager Reza Vaez-Ghaemi took a deep dive into test and measurement for timing and synchronization in increasingly dynamic, disaggregated wireless networks. Citing the relevant (and extensive) associated standards from 3GPP and ITU-T, Vaez-Ghaemi noted the importance of conformance to standards across networking domains.
Specific to Open RAN, Vaez-Ghaemi said, “O-RAN introduced four lower-layer split (LLS) synchronization configuration modes (LL1-C1 [through] LLS-C4), added requirements for time/phase and frequency budgets, node behavior guidelines and detailed error budget analysis aspects. It also brought in an Open Fronthaul Management plane that defines synchronization attributes that further help more smoothly interoperate solutions from different vendors.”
In addition to conformance testing for open distributed units and open radio units, an end-to-end view comes into focus with interoperability testing for fronthaul and midhaul, as well as test requirements for transport networks in fronthaul, midhaul and backhaul.
In terms of relevant synchronization metrics, Time Alignment Error (TAE) includes relative and absolute figures; time error (TE) is defined as the time differences at a user network interface (UNI) compared to another UNI or based on a Primary Reference Time Clock (PRTC). To put that in context, MIMO diversity transmissions at each frequency need a relative TAE of 65 nanoseconds and 5G intra- and inter-band, non-contiguous carrier aggregation needs a relative TAE of 3 microseconds.
“The more we get into these advanced wireless services and capabilities, the synchronization requirements increase significantly,” Vaez-Ghaemi said. “We see all these requirements becoming tighter and tighter.”
In terms of the entire timing and synchronization chain, and how to test across it, the workflow is as follows:
- GNSS/GPS installation verification, including antenna location, receiver functionality, cabling integrity, satellite availability and signal strength. Then ensure accurate antenna cable delay is measured and respective bias value is entered into the attached PTP primary.
- Fronthaul network synchronization testing, including verification of GPS, connection and configuration of sync parameters, and synchronization applications—emulate Precision Time Protocol (PTP) secondary and perform time error measurements, then emulate Synchronous Ethernet (SyncE) and validate frequency offset and Ethernet Synchronization Message Channel (ESMC).
- Midhaul/backhaul synchronization testing wherein operators may deploy (Assisted) Partial Timing Support (APTS), emulate Precision Time Protocol (PTP) secondary, and perform two-way packet selected time error and floor packet percentile measurements.
- SyncE testing, including functionality and configuration of Ethernet Equipment Clock (EEC) in the synchronization path, SyncE Wander Analysis to measure dynamic frequency error in the network, and use the Synchronization Status Message (SSM) to help establish whether the device under test is transmitting the proper quality level
- Finally, over the air (OTA) synchronization measurement to verify that the radio sync conforms to the appropriate synchronized mode and delivers the expected outcome in terms of frequency and time error versus GPS and time error versus GPS.
“At the end,” Vaez-Ghaemi said, “this whole thing is about the air interface. That measurement measures basically time/phase and frequency against the GPS or another reference…[It] should be then within that limit 3GPP defined both for frequency and time error.”
