ByRichard Bradbury, BBC Research & Development
First published Oct. 2022, in HighlightsIssue 05
In Release 17, 3GPP has defined a new reference architecture for collecting UE data and exposing it to downstream consumers. As well as supporting generic UE data, such as service experience, the abstract building blocks are intended to be reusable in any number of different data domains, the first of which is 5G Media Streaming.
In TS23.501 and TS23.288, SA2 has defined a Network Data Analytics Function (NWDAF) which analyses data collected from Network Functions and UEs in the 5G System, and publishes the results to subscribing data analytics consumers. Network Functions supporting the data analytics architecture implement an event exposure service-based API. The NWDAF subscribes to uniquely identified events published by Network Functions in order to acquire the data that it then goes on to analyse and expose to its subscribers. These may, in turn, use the analysed data to modify the operation and/or configuration of the 5G System, potentially in real time. In this way, a closed loop can be formed to continuously optimise a 5G System.
Reference architecture for UE data collection, reporting and exposure
To support the collection of data from UEs, SA2’s architecture envisages an Application Function that receives data reports from a Data Collection Client via conventional data plane PDU Sessions and exposes events to the NWDAF using the Naf_EventExposure service-based interface defined by SA2 and specified by CT3 in TS29.517. Recognising its domain expertise in media-related metrics, SA2 turned to SA4 for help in defining the Data Collection AF and the related architecture for UE data collection, reporting and exposure. This is documented by SA4 in TS26.531 and is summarised in the figure below. Detailed procedures for provisioning the Data Collection AF and reporting UE data to it, as well as definitions of the corresponding service-based interfaces, are specified by SA4 in the companion TS26.532.
As well as supporting the (currently small) repertoire of generic UE-related events defined by SA2, SA4 decided to expand the scope of the UE data collection, reporting and exposure system in two different directions, as highlighted by the second figure:
- It was recognised that Application Server (AS) instances can be an additional source of UE data. Application logs contain valuable data about UE activity that can be analysed to extract patterns and trends.
- It was recognised that third party Application Service Providers (ASPs) are valid consumers of UE events from the Data Collection AF in addition to the NWDAF. An ASP’s Event Consumer AF may wish to perform its own analysis of UE data in order to optimise its applications.
In order to support this extended architecture, additional data protection mechanisms were deemed necessary by SA4. When the collection of UE data is provisioned by an ASP at the Data Collection AF, a number of data processing instructions can be specified to limit the UE data exposed to event consumers. These instructions are expressed in the form of Data Access Profiles as follows:
- For a particular event type, the exact parameters to be collected can be limited by each Data Access Profile. This permits compliance with one of the key principles of data protection legislation that only data necessary for specific purposes should be collected.
- In addition, each metric of collected UE data can be summarised along the axes of time, user and/or location using an aggregation function. For example, rather than exposing events detailing the service experienced by individual UEs, a particular Data Access Profile may expose only maximum, minimum and mean average values aggregated over five-minute intervals.
- Multiple Data Access Profiles can be provisioned for a given event type to vary the data restrictions imposed on different event consumers. When more than one Data Access Profile is provisioned, the Data Collection AF selects one based on local policy when it receives a new subscription request from an event consumer.
- As part of the authorisation procedure for event consumers, the Data Collection AF may also collaborate with an external Authorization AS, following a similar message exchange pattern to OAuth.
Instantiation of the reference architecture
The reference architecture for UE data collection, reporting and exposure defined in TS26.531 is abstract in the sense that it is intended to be instantiated inside the architecture of other data domains. One such instantiation has been defined by SA4 in Release17 for the 5G Media Streaming (5GMS) data domain. Summarised in the figure below, the detailed definition can be found in TS26.501.
Here, the Data Collection Client is realised by the existing UE-side Media Session Handler component. This uses existing consumption reporting and QoE metrics reporting mechanisms to report collected UE data to a Data Collection AF realised by the existing 5GMS Application Function. Thus, consumption and QoE metrics can be exposed as events to downstream event exposure subscribers such as a 5GMS Application Provider’s Event Consumer AF. Similarly, invocations of standard media streaming features such as dynamic QoS policies and network assistance that are already logged by the 5GMS AF can also be exposed as events. Finally, the media streaming application logs collected by the 5GMS Application Server can be passed on to the Data Collection AF for exposure as events. In all, five new event types are defined in TS26.532 for these 5GMS features, and the corresponding data types for exposing them to event consumers have been added to TS29.517 by CT3.
Future work
SA4 expects to instantiate the UE data collection, reporting and exposure architecture in other data domains in subsequent 3GPP releases. Hence, there may be more than one concrete instantiation of the Data Collection AF in a given 5G System deployment, one for each data domain. And there may be multiple logical Data Collection Clients running on a UE.
Meanwhile, in Release18, SA2 is studying the feasibility of exposing the data analytics output of the NWDAF back to UEs in order to drive feedback loops for Artificial Intelligence and Machine Learning applications.
Further reading:
TS23.288 | Architecture enhancements for 5G System (5GS) to support network data analytics services |
TS23.501 | System architecture for the 5G System (5GS) |
TS26.501 | 5G Media Streaming (5GMS); General description and architecture |
TS29.517 | 5G System; Application Function Event Exposure Service; Stage 3 |
TS26.531 | Data Collection and Reporting; General Description and Architecture |
TS26.532 | Data Collection and Reporting; Protocols and Formats |