Tech-invite3GPPspaceIETFspace
21222324252627282931323334353637384‑5x

Content for  TR 22.832  Word version:  17.4.0

Top   Top   Up   Prev   Next
0…   4…   5…   6…

 

4  Overviewp. 10

4.1  Overview of use casesp. 10

This document describes vertical use cases of cyber-physical control applications that provide further potential service requirements in addition to the 5G service requirements in TS 22.261 and TS 22.104. The description is from a system's perspective at a summary level. It also provides clarification on vertical use cases in TS 22.104 where needed and leading to additional potential 5G service requirements.
Up

4.2  Discussion on network performance requirementsp. 10

4.2.1  KPIs from TS 22.104 and TS 22.261p. 10

The key performance requirements for cyber-physical control applications in vertical domains are specified in TS 22.104, including the new KPIs in addition to the usual KPIs (i.e., end-to-end latency, message size, service bit rate, and transfer interval):
Survival time - The maximum survival time indicates the time period the communication service may not meet the application's message delay requirement before there is an application layer failure such that the communication service is deemed to be in an unavailable state.
Communication service availability - This KPI indicates if the communication system works as contracted ("available"/"unavailable" state). The communication service is in the "available" state as long as the availability criteria for transmitted messages are met. The communication service is unavailable if the messages received at the target are impaired and/or untimely (e.g. update time > stipulated maximum), resulting in survival time being exceeded.
Communication service reliability - Mean time between failures is one of the typical indicators for communication service reliability. This KPI states the mean value of how long the communication service is available before it becomes unavailable.
Up

4.2.2  KPIs from TS 23.501p. 11

Meanwhile the 5G QoS characteristics are specified in TS 23.501 to describe the packet forwarding treatment that a QoS Flow receives edge-to-edge between the UE and the UPF. The most relevant performance characteristics to the identified key performance requirements are:
Packet Delay Budget - The Packet Delay Budget (PDB) defines an upper bound for the time that a packet may be delayed between the UE and the UPF that terminates the N6 interface. For a certain 5QI the value of the PDB is the same in UL and DL. In the case of 3GPP access, the PDB is used to support the configuration of scheduling and link layer functions (e.g. the setting of scheduling priority weights and HARQ target operating points).
Packet Error Rate - The Packet Error Rate (PER) defines an upper bound for the rate of PDUs (e.g. IP packets) that have been processed by the sender of a link layer protocol (e.g. RLC in RAN of a 3GPP access) but that are not successfully delivered by the corresponding receiver to the upper layer (e.g. PDCP in RAN of a 3GPP access) within the packet delay budget. Thus, the PER defines an upper bound for a rate of packet losses. The purpose of the PER is to allow for appropriate link layer protocol configurations (e.g. RLC and HARQ in RAN of a 3GPP access). For every 5QI the value of the PER is the same in UL and DL.
Up

4.2.3  Relationships between KPIs for 5G system designp. 11

The communication service reliability and communication service availability are complementary to packet error rate (PER). PER can be used to indicate the significance of individual packet losses which for many of the industrial applications differs from the significance of losing several consecutive packets (packet is 'lost' if it is not delivered intact within PDB). For example, loss of a single packet may only slightly reduce the quality of experience of an industrial application (e.g.. precision of a motion control application), while loss of several consecutive packets is considered as communication service unavailability potentially resulting in an emergency stop in the application. Packet error rate (PER) is directly related to communication service reliability and communication service availability only in the special case where 'failure' of the communication system is specified to be loss of a single packet ( e.g., survival time is zero and 1 message is contained in 1 packet) for periodic deterministic traffic.
Communication service availability and communication service reliability indicate the significance of how packet losses are distributed in time domain. For example, if 5G system design considers avoidance of multiple consecutive packet losses with higher priority than individual packet losses, this may result in a system design that improves the quality of experience of the industrial applications. Both communication service availability and communication service reliability are meaningful only when specified in context with survival time.
Communication service reliability can be quantified as the mean time between failures. Failure refers to the event when the communication service becomes 'unavailable' considering the application specific requirements. For many of the industrial applications, the communication service is considered unavailable if survival time is exceeded. For applications that have survival time equal to zero, any loss of packet triggers this unavailability, while for applications with non-zero survival time only two or more consecutive packet losses will trigger unavailability (depending on the agreed traffic periodicity and length of the survival time). The communication service is considered available again when it successfully delivers a packet, or the full set of packets constituting a message when message segmentation is utilized, within the delay constraints.
The communication service reliability and communication service availability can be considered in 5G system design e.g., by developing solutions that reduce the probability of exceeding the survival time.
Communication service availability can also be estimated from the mean time between failures (MTBF) and the mean time to repair (MTTR) of the communication service;
communication service availability ≈ MTBF / (MTBF + MTTR)
In this context, MTBF excludes downtime, as illustrated in Figure 4.2-1, while MTTR refers to the mean time until the communication service is available after a failure, i.e., until the next valid packet has been received .
Copy of original 3GPP image for 3GPP TS 22.832, Fig. 4.2-1: MTBF and MTTR
Figure 4.2-1: MTBF and MTTR
(⇒ copy of original 3GPP image)
Up
Survival time is another indicator that may be considered for 3GPP 5G system design to allow the performance requirements of cyber-physical control applications to be met. If survival time is assumed to be zero, the 5G system may over-provision the PER targets which may lead to significant reduction is system capacity and/or reduce the communication service reliability and communication service availability (e.g., in the case when there is resource conflict between two URLLC service flows) When message segmentation is utilized in the 5G system, survival time relates to the successful delivery of all packets comprising an application layer message rather than a single packet.
For many IIoT applications individual packet errors can be tolerated but exceeding survival time cannot. This allows that
target PER >1 - Communication Service Availability
One potential use of survival time could be to adjust PER if survival time is in jeopardy. For example, if packet errors are detected but survival time has not yet expired, steps could be taken to ensure delivery of subsequent packets within survival time.
The dependencies between communication service availability, communication service reliability and survival time might be considered in the 5G system to see if more efficient optimization can be achieved. One dependency to be considered is the number of packets lost during the survival time interval.
Up

4.3  Survival time vs. consecutive message lossp. 12

Survival time is defined in TS 22.104 as "the time that an application consuming a communication service may continue without an anticipated message". It has been identified as one influential quantity for periodic deterministic communication (Table 5.2-1 in TS 22.104). The survival time indicates to the communication service the time available to recover from message delivery failures. The survival time is expressed as a time period which, especially with cyclic traffic, accommodates the maximum number of consecutive incorrectly received or lost messages that can be tolerated without causing an application layer failure.
An example for periodic communication is given in Figure 4.3-1. The automation application delivers the messages to the ingress of the 5G system at a given transfer interval. When the messages are correctly received by the automation application at the egress of the 5G system, it is labelled as per message network status "UP" as well as communication service status "UP". Per message network status refers to the status as perceived by the application.
Incorrectly received or lost messages lead to the per message network status "DOWN". In practice if there is no message correctly received within the receiving window (e.g., based on the transfer interval and the latency), it will be considered as per message network down time. If that down time is within the limit of the pre-defined survival time such transmission errors can be compensated by the application. A communication service failure occurs when more consecutive messages are lost than the survival time allows, which also leads to a failure on the application layer. In an application layer failure situation, the application stops and then has to be restarted again after the communication service has recovered. This is represented by the application recovery time in Figure 4.3-1. This Figure illustrates the specific case when one message is contained in one packet.
Copy of original 3GPP image for 3GPP TS 22.832, Fig. 4.3-1: survival time vs. message loss
Figure 4.3-1: survival time vs. message loss
(⇒ copy of original 3GPP image)
Up
For cyber physical control applications, most messages will be small and may not need segmentation. In case messages are segmented into multiple packets, message delivery is only successful if all packets of the message are received within the PDB. Thus, it may not be sufficient to successfully deliver one packet after a packet failure to detect communication service availability, rather successful delivery of one full message is needed. It may be necessary to map the maximum allowed consecutive message loss to the maximum allowed consecutive packet loss in a 5G system, which depends on factors such as survival time, transfer interval and the message size.
Up

Up   Top   ToC