Why O-RAN sync monitoring needs AI/ML

Infineon / Mitsubishi / Fuji / Semikron / Eupec / IXYS

Why O-RAN sync monitoring needs AI/ML

Posted Date: 2023-09-07

As Open RAN deployments collect tempo and correct timing, synchronization has emerged as a crucial step towards constant efficiency. It's a advanced activity, with units requiring time synchronization to UTC reference by way of IEEE PTP/SyncE-based boundary clocks, slave clocks, PRTC clocks with GNSS receivers, in addition to correct, real-time monitoring in assist of buyer SLAs. As hundreds of Open RAN units are deployed, monitoring the community for related failures and taking applicable motion would require AI/ML developments. Early classes from the sector reveal a path ahead.

Correct and dependable synchronization has lengthy been required to keep up crucial telecom community capabilities, corresponding to cell operations, cell protection and effectivity, exact handover between cell towers, and so forth. Telecom service suppliers can implement varied strategies to fulfill stringent section and time synchronization necessities. Every technique’s intent is to make sure synchronization of all nodes to the first reference time clock (PRTC) supply utilizing the International Positioning System (GPS) / International Navigation Satellite tv for pc System (GNSS). GNSS refers to a constellation of satellites offering alerts from area that transmit positioning and timing knowledge to GNSS receivers for time reference.

The placement of the sync supply could, nonetheless, fluctuate relying on the community topology, price, and utility. Usually, telecom community nodes are synchronized utilizing Precision Timing Protocol (PTP) and Synchronous Ethernet (SyncE) applied sciences utilized in IEEE 1588 PTP Grandmaster (PRTC supply), boundary clocks, clear clocks, and slave clocks inside the community.

Open RAN is an industrywide initiative to develop open interfaces, disaggregate RAN into its particular person community parts, and allow interoperability amongst {hardware} suppliers. The standardization of Open RAN is supposed to present operators the flexibleness to combine and match radio community elements for greatest efficiency. This flexibility is just not with out complexity, additional rising the criticality of community timing perform synchronization throughout elements. Particularly, given synchronization purposes are distributed throughout the front-haul (FH) and mid-haul (MH) community, together with the distributed unit (DU) and radio unit (RU), you will need to synchronize the purposes needing timing on the DU and RU to keep up superb sector/cell key efficiency indicators (KPIs).

In a legacy radio entry community (RAN), all finish purposes needing synchronization reside inside a single eNodeB. In 3GPP break up architectures (Open RAN), disaggregated purposes want timing distributed over the community, together with DU, RU, and centralized unit (CU). That timing distribution will increase the operator’s burden of managing the synchronization networks.

Eyes on KPIs
With the evolution of IEEE PTP, the radio community’s dependency on GNSS techniques has considerably decreased. GNSS is, nonetheless, nonetheless wanted because the PRTC as outlined in ITU-T G.8272 and ITU-T G.8272.1 specs, relying on the kind of deployment. That is true for Open RAN DU, which is a logical node internet hosting RLC/MAC/high-PHY layers based mostly on a lower-layer purposeful break up, and RU, which is a logical node internet hosting Low-PHY layer and RF processing based mostly on a lower-layer purposeful break up (Determine 1). IEEE 1588 grandmaster clocks (PRTC clock supply) are positioned on the central knowledge heart inside the community to supply the clock to DUs and RUs in front-haul or mid-haul networks.

Why O-RAN sync monitoring needs AI/ML

Determine 1. In a break up 7.2 O-RAN structure, low-PHY features reside within the radio items whereas high-PHY features reside within the distribution unit.

Each PRTC clock system with an built-in GNSS receiver or an intermediate boundary clock supply that gives the time to the downstream community is inclined to potential sync outages.

Relying on the place the GNSS supply is put in or positioned, the affect on all the community varies. For instance, an O-RAN Alliance-compliant community with the DU performing as PTP grandmaster sourcing the clock to RUs in a front-haul community will see any GPS outages on the DU in RAN LLS-C1 configuration affect the DU and related RUs.

In distinction, sourcing the GPS on the RU, which has an built-in GPS receiver and RU performing as PRTC sync supply for the purposes, would solely see an outage affect at that RU. This additionally has implications for the PTP/SyncE KPIs distributed over the community from the PRTC supply.

All timing and sync configurations and provisioning for a cloud-based deployment occur over a community. Outages are rising, and networks must be monitored intently to take corrective actions. This necessitates a central, real-time, correct monitoring mechanism over the cloud community to retrieve the timing KPIs, construct intelligence to estimate the outages based mostly on key metrics, and take reactive and corrective steps to keep away from impacting cell KPIs and enhance cell availability.

Root causes of sync-based outages
As Open RAN networks proliferate, operators want to grasp the first causes of sync outages, all of which require real-time monitoring to detect. In our work in Japan and different international locations, we've recognized:

  • GNSS sign jamming, each intentional and unintentional
  • GNSS sign spoofing, together with of a number of constellations and on a number of frequency bands
  • Sign blockages and multipath errors brought on by tree canopies in rural areas or giant glass buildings in city areas
  • Ionospheric results and geographical points
  • {Hardware} faults connecting GNSS receivers, corresponding to guarantee arrestor points as a result of lighting or cable faults
  • Poor climate circumstances
  • Leap second warnings, together with the necessity to inform all purposes on upcoming leap second additions/deletions
  • PTP packet drops, clock high quality degradations, and clock commercials
  • SyncE clock high quality degradations

There are additionally safety issues to think about. Fashionable networks present important infrastructure for enterprise, mission, and society-critical purposes that are of nationwide concern. Between July 2020 and June 2021, the telecom trade was probably the most focused trade with regard to GNSS safety threats, with 40% of assaults versus 10% for the next-highest trade vertical (Supply: EUROCONTROL EVAIR).

Detect and analyze timing and sync failures
Primarily based on our expertise, community operators can take the next steps to provoke mitigating actions for timing and synchronization failures, together with these associated to GNSS Safety threats:

  • Develop a mechanism for the O-Cloud interface to detect and analyze GNSS sign jamming, or interference or spoofing circumstances earlier than the circumstances deteriorate.
  • Analyze unhealthy climate circumstances that may intervene with GNSS alerts based mostly on GNSS outage historical past detected on one DU. This knowledge can be utilized to foretell and anticipate how these circumstances would possibly affect neighboring and co-located DUs, permitting an operator to extra shortly provoke corrective actions.
  • Equally, analyze GNSS error circumstances and predict how these circumstances would possibly doubtlessly affect neighboring and co-located DUs to provoke corrective actions.
  • Detect and Analyze packet timing sign failure (PTSF) circumstances and provoke mitigative actions. Predict PTSF circumstances, and take corrective actions corresponding to transferring to an alternate clock supply earlier than ready for the faults to really occur and affect the community.

This isn't meant to be an exhaustive checklist. These mechanisms must be reported over the O-Cloud interface and herald vital contributions to WG6 O-Cloud API spec for the synchronization airplane (S-plane) as the present datasets aren't ample.

AI and ML for timing and synchronization
A multi-layered strategy to integrating synthetic intelligence (AI) and machine studying (ML) into timing and synchronization ought to be thought-about for exact, efficient, optimized dealing with of timing and synchronization occasions.

For instance, an answer to timing outages might be divided into a number of layers and leverage AI/ML mechanisms to do native studying per DU/RU, a set of DUs/RUs pertaining to that CU, or a bunch of O-CUs (see Determine 2). It wants strategies to outline the interactions between these a number of layers in real-time and non-real-time based mostly on occasion, alarm, or error sort. In apply an operator would:

  • Outline detection and mitigation algorithms for native sync failures inside DU/RUs.
  • Analyze info publicity to rApps/xApps to make clever mitigation actions and insurance policies at Non-RT-RIC (SMO) or Close to RT-RIC.
  • Construct automation and AI/ML-based algorithms for sync failure detection and analytics as a part of the 5G core or developed packet core (EPC).

Determine 2. This Excessive-level structure can leverage AI and ML for timing and synchronization in Open RAN networks.

Open RAN brings the promise of elevated income streams, decrease capex, and TCO, and variety benefits to operators. Reaping these advantages will rely partly on leveraging AI/ML applied sciences in timing and synchronization options to carry extra effectivity and optimization to managing sync networks in Open RAN and assist operators enhance KPIs and the client expertise.