Above MEC V2X business analysis (1) (ETSI GS MEC 030 V3.1.1 (2023-03))
Brief description: V2X business information service can predict QoS to help vehicles make decisions. One is for characteristic routes, that is, journey-specific predictive QoS; the other is multi-domain end-to-end prediction, which can predict vehicle workshop, vehicle to facility Between QoS.
5.4.5 VIS and its role in generating journey-specific predictive QoS notifications
Accurate and timely predictions of the radio environment at locations planned to be visited by vehicles can trigger, modify or postpone: i) application of certain V2X functions; and/or ii) content delivery/download of software packages.
However, for V2X system scenarios characterized by high mobility and dynamic topology (as shown in Figure 5.4.5-1), the accuracy and timeliness of information (such as radio network, location information, etc.) may be affected by the following factors Impact:
- Environmental situations, such as the occurrence of network congestion events when many vehicles try to provide radio measurements to eNB/gNB connected to the MEC host;
- · Deployment density of the cellular network, and the capabilities of the deployed MEC infrastructure.
An example of the impact of the above constraints on system performance is the vehicle planning trajectory from location A to location B and the associated MEC application, prior to the vehicle passing, needs to be informed of the wireless condition "route" in advance, and then a decision can be made. Decisions may include enabling/disabling autonomous driving features, downloading infotainment content, scheduling software/firmware over-the-air updates (SOTA/FOTA), etc. To address these challenges, VIS services can help implement a framework for collaboratively acquiring, segmenting, and distributing information for efficient, journey-specific QoS prediction. That is, VIS services can be utilized to identify spatial/temporal correlations between radio quality data collected by different vehicles in a V2X system and a specific vehicle's planned trip to better predict communication network quality along a given route.
5.4.6 VIS and its role in multi-domain end-to-end predictive QoS notification
In addition to the data collected by different vehicles, VIS can also support predictive QoS notifications in multi-domain MEC deployments by collecting and distributing external Prediction Functions (PF) located in different network domains. Obtaining analysis information from external PFs such as Network Data Analytics Function (NWDAF) in the 3GPP network domain helps to provide more accurate QoS predictions for service consumers. At the 5G Automotive Association (5GAA) event, several scenarios were investigated where a service consumer might benefit from information related to QoS prediction of an E2E user plane link between two specific endpoints. These scenarios range from teleoperation driving (ToD), to collaborative operation scenarios and use cases, such as lane merging, platooning, etc., as well as HD map collection/sharing and in-vehicle infotainment, as described in [i.30] and [i.31].
What is VIS end-to-end QoS prediction?
** is information related to QoS prediction of an E2E user plane link between two specific endpoints.
So what can these two specific endpoints be?
** V2N2V is the vehicle-to-network and then to the vehicle end-to-end, and the two endpoints are the vehicle and the vehicle.
** V2N2I means vehicle to network to road facilities, and the two endpoints refer to Vehicle to Infrastructure.
What is the connection between V2N2V and V2N2I ? How many network domains are there?
Among them, the QoS prediction of the E2E user plane link needs to consider at most 5 different network domains. In Figure 5.4.6-1 (V2N2V), these domains are specifically: the 3GPP network domains of the two operators (MNO A and MNO B), the data networks (DNA) and (DN B) where the MEC host is located, and The Internet's IP control network.
those domains are specifically: the 3GPP network domains of the two operators (MNO A an MNO B), the Data Networks (DN A) and (DN B) where the MEC host is located, as well as the IP controlled network that is interconnecting the networks of the PLMN A and PLMN B.
In Figure 5.4.6-2 (V2N2I case), the network domain used for the E2E user plane link is replaced by 4 different domains: 3GPP network domain where the operator provides 5G connectivity to the vehicle (MNO A of vehicle 1) , the DN where the MEC host is located (the DNA of vehicle 1), the interconnection domain and the third-party domain where the road infrastructure is located.
In figure 5.4.6-2 (case V2N2I), those network domains for E2E user plane link are instead 4 distinct domains: the 3GPP network domain of the operator that provides 5G connectivity to the vehicle (MNO A for vehicle 1), the DN (DN A for vehicle 1) where the MEC host is located, the interconnection domain and the 3rd party domain where the road infrastructure is located.
How to predict the QoS of each network domain?
- 3GPP domain VIS can cooperate with the analysis service provided by NWDAF according to clause 6.9 of ETSI TS 123 288 [ i.27 ]. Provides predictions such as QoS sustainability analysis for 3GPP network domains ( e.g. at radio access network - RAN level and core network segment between UE and user plane functions - UPF nodes).
- DN in the IP interconnection domain or service provider domain (including MEC hosting) , domain-specific PF can be deployed to provide analysis information to users who invoke services through VIS. These DN domain-specific predictions may be provided by NWDAF as described in DN performance analysis described in clause 6.14 of ETSI TS 123 288 [i.27].
What information is the predicted QoS?
3GPP system domain-specific predictions may include analysis of user plane performance (i.e. average/maximum traffic rate, average/maximum packet delay, average packet loss rate ), provided to service consumers in the form of statistics or predictions.
In the considered multi-domain scenario, VIS can aggregate these information to provide end-to-end service quality prediction. According to clause 5.2.6.16 of ETSI TS 123 502 [i.28], these user plane performance analyzes can be obtained directly from NWDAF or through the Network Exposure Function (NEF).
Note: In some cases, V2X MEC applications communicating with VIS can also generate predictions for domains where PF functionality is not available based on the collected information . VIS can aggregate analytics from external PFs such as NWDAF or other analytics providers to provide (predict/estimate) QoS covering the E2E user plane link between two V2X application instances.
5.4.7 Interaction between VIS and non-conversational V2X services
Some V2X services are non-session, and these services involve broadcasting information related to events or warnings about potential hazards (eg road status announcements, warning reminders). This means that a single V2X message needs to be delivered to a group of locally related UEs or a group of locally related V2X application servers.
One possible deployment method is based on using a V2X message distribution server, which is responsible for receiving V2X messages from non-conversational V2X services and distributing them to the relevant target entities. Message Broker is usually deployed as a V2X message distribution server, and Message Broker can be based on different application layer protocols.
5.4.8 VIS and information provision of V2X message distribution server
V2X MEC applications that implement non-session-based V2X services require very low latency. MEC applications may need to directly interact with the corresponding V2X message distribution server to avoid any additional delay introduced by VIS. Therefore, MEC applications should provide provisioning information about which V2X message distribution servers are available, their characteristics, and how to connect with them.
Such information can be provided by the VIS when V2X message distribution servers are registered to the MEC platform, for example when these V2X message distribution servers are implemented as service-generating MEC applications. Each V2X message distribution server can manage the exchange of V2X messages in a given geographical area based on a specific application layer protocol (eg MQTT, AMQP) . The MEC application can provide some criteria for selecting only relevant V2X message distribution servers according to the application layer protocols supported by the MEC application or the geographical area of interest of the MEC application. The way the V2X message distribution server registers to the MEC platform (e.g., as a service producing MEC application) and the interaction between the VIS and the V2X message distribution server is beyond the scope of this paper.
Reference：
ETSI GS MEC 030 V3.1.1 (2023-03)