WO2011050540A1

WO2011050540A1 - Method in a wireless communication system for determining quality of service fulfilment

Info

Publication number: WO2011050540A1
Application number: PCT/CN2009/074736
Authority: WO
Inventors: Peter Legg; Kai Zhang; David Soldani
Original assignee: Huawei Technologies Co.,Ltd.
Priority date: 2009-10-31
Filing date: 2009-10-31
Publication date: 2011-05-05
Also published as: EP2417796A4; CN102396259A; EP2417796A1

Abstract

The present invention relates to a method in a wireless communication system for determining quality of service (QoS) fulfilment of one or more data flows between a first communication node (N1) and a second communication node (N2) in a wireless communication system, wherein said one or more data flows are transmitted on at least one radio link between said first (N1) and second (N2) communication nodes, and each data flow is associated with one or more quality of service (QoS) attributes, said method comprising the steps of: measuring at least one measurement value for each one of said one or more quality of service (QoS) attributes; comparing said at least one measurement value for each one of said one or more quality of service (QoS) attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and determining the quality of service (QoS) fulfilment of said one or more data flows based on said one or more satisfied or non-satisfied representations. The invention also relates to a computer program, a method in a communication node, and a communication node apparatus thereof.

Description

METHOD IN A WIRELESS COMMUNICATION SYSTEM FOR DETERMINING

QUALITY OF SERVICE FULFILMENT

Technical Field

The present invention relates to a method in a wireless communication system for determining quality of service fulfilment, or more specifically to a method according to the preamble of claim 1. The invention also relates to a computer program, a method in a communication node, and a communication node apparatus thereof. Background of the Invention

Network management of a Radio Access Network (RAN) according to 3GPP standard TS32.101 involves three principle components:

• A Network Element (NE) - an eNB in a Long Term Evolution (LTE) communication system,

· An Element Manager (EM) - a management entity that manages one or more NEs, and

• A Network Manager (NM) - a management entity that usually manages one or more EMs, but a direct management interface to the NE is also possible. Figure 1 and 2 shows two possible network management configurations which are supported by the 3GPP standard, i.e. System Context A and System Context B, respectively. In both contexts an Interface-N (Itf-N) exists, and 3GPP has defined a number of management services which can run across said interface. The Itf-N is either:

• The management interface between an EM and a NM - called System Context A; or · The EM functionality may in certain cases reside in the NE in which case this

interface is directly between the NE and NM - called System Context B.

Further, 3GPP defines some Integration Reference Points (IRPs) over the Itf-N, e.g. Entry Point (EP) IRP, Notification IRP, etc. 3GPP publishes a number of IRP specifications, each relating to a set of operations and notifications for a specific telecom management domain such as alarm management, configuration management, etc. In Figures 1 and 2, the IRP

Manager represents the management entity, whilst the IRP Agent undertakes tasks on behalf of the IRP Manager. Interface IRPs also contain definitions of supported Information Object Classes (IOCs) which describe the information that can be passed/used in management interfaces. Also, of relevance to this disclosure is a Network Resource Model (NRM) which is an information service describing IOCs representing the manageable aspects of network resources, e.g. a Radio Network Controller (RNC) or a base station, such as a NodeB or eNB.

Furthermore, the interface between an EM and a NE is called Interface-S (Itf-S). This interface is not subject to standardization. The interface between the NM and the NE in System Context B is sometimes called the direct interface.

The Performance Management (PM) IRP allows the NM to assess performance of different aspects of the NE using performance measurements that are taken by the NE and passed over the Itf-N to the NM. In the 3GPP LTE system performance measurements taken by the eNB are captured in the specification 3GPP TS32.425. Examples of measurements are: attempted outgoing intra-frequency handovers, in session activity time for a User Equipment (UE), average Downlink (DL) cell packet bit rate, and DL packet drop rate.

TS32.425 specifies in detail how each of these measurements should be taken. Many of the measurements involve a simple accumulative counting of an event or a metric, and often the terminology "performance counters" is used synomonously with "performance

measurements". In TS32.425 a number of counters reflect the Quality of Service (QoS) within a cell. Below is a list from the table of contents:

4.4 Cell level radio bearer QoS related measurements 30

4.4.1 Cell PDCP SDU bit-rate 30

4.4.1.1 Average DL cell PDCP SDU bit-rate 30

4.4.1.2 Average UL cell PDCP SDU bit-rate 31

4.4.1.3 Maximum DL cell PDCP SDU bit-rate 31

4.4.1.4Maximum UL cell PDCP SDU bit-rate 32

4.4.1.5 Average DL cell control plane PDCP SDU bit-rate 32

4.4.1.6 Average UL cell control plane PDCP SDU bit-rate 32

4.4.2 Active UEs 33

4.4.2.1 Average number of active UEs on the DL 33 4.4.2.2Average number of active UEs on the UL 33

4.4.3 Packet Delay and Drop Rate 33

4.4.3.1 Average DL PDCP SDU delay 33

4.4.3.2DL PDCP SDU drop rate 34

4.4.4 Packet loss rate 34

4.4.4.1DL PDCP SDU air interface loss rate 34

4.4.4.2UL PDCP SDU loss rate 34

4.4.5 IP Latency measurements 35

4.4.5. IIP Latency in DL, SAE Bearer level 35

However, the specified measurements only express the QoS aggregated across the users in a cell. So, e.g. the "DL PDCP SDU air interface loss rate " is calculated by counting how many packets that are delivered on the DL and how many are lost over the air interface, taking the UE population of the cell as a whole. Therefore, it is not known from these measurements whether the losses, e.g. are isolated to one or two UEs who would then have poor QoS, or if they are spread evenly across the cell population who could then all have satisfactory QoS. One of the objects of the present invention is to address this limitation of prior art.

According to another prior art solution individual user's Asynchronous Transfer Mode (ATM) cell loss rates are derived from an aggregated cell loss measurement. This prior art solution is limited to only one aspect of QoS, namely the cell loss rate.

According to yet another prior art solution measurements of different Service Level

Agreements (SLA) of IP flows are defined for an end to end IP connection. This prior art solution is limited to these QoS aspects individually, and no attempt is made to combine the SLA satisfaction levels for a single end device.

Summary of the Invention

One object of the present invention is to provide a solution for determining the QoS fulfilment of one or more data flows between a first and second node in a wireless communication system. Another object of the invention is to provide a solution which wholly or in part solves the problems and limitations of prior art. Yet another object of the invention is to provide a solution which makes it possible for a mobile operator to tune assessment parameters relating to QoS fulfilment for data flows and/or mobile stations.

According to one aspect of the invention, the objects are achieved with a method in a wireless communication system for determining quality of service (QoS) fulfilment of one or more data flows between a first communication node (Nl) and a second communication node (N2) in a wireless communication system, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication nodes, and each data flow is associated with one or more quality of service (QoS) attributes, said method comprising the steps of:

- measuring at least one measurement value for each one of said one or more quality of service (QoS) attributes;

- comparing said at least one measurement value for each one of said one or more quality of service (QoS) attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and

- determining the quality of service (QoS) fulfilment of said one or more data flows based on said one or more satisfied or non- satisfied representations.

Embodiments of the method in a wireless communication system above are disclosed in dependent claims 2-14.

According to another aspect of the invention, the objects are also achieved with a method in a first communication node for determining quality of service (QoS) fulfilment of one or more data flows between said first communication node (Nl) and a second communication node (N2) in a wireless communication system, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication nodes, and each data flow is associated with one or more quality of service (QoS) attributes, said method comprising the steps of:

- comparing said at least one measurement value for each one of said one or more quality of service (QoS) attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and - determining the quality of service (QoS) fulfilment of said one or more data flows based on said one or more satisfied or non- satisfied representations.

Embodiments of the method in a first communication apparatus above are disclosed in dependent claims 16-28.

The method in the first communication node can be implemented in a computer program which when run in a computer causes said computer to execute the method in the first communication node.

According to yet another aspect of the invention the objects are also achieved with a first communication node apparatus arranged for wireless communication with a second communication node (N2) apparatus in a wireless communication system, said wireless communication system employing one or more data flows for communication between said first (Nl) and second (N2) communication node apparatuses, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication node apparatuses, and each data flow is associated with one or more quality of service (QoS) attributes, wherein said first communication node (Nl) apparatus is configured for:

The communication node apparatus according to the invention may also be configured according to method claims 16-28.

The present invention exploits a hierarchical approach for determining the QoS fulfilment of one or more data flows belonging to e.g. a mobile station, and hence the invention provides the means to perform performance measurements of the QoS fulfilment of individual mobile stations. These performance measurements can e.g. be passed to an Operations and

Maintenance (OAM) system for further processing. The invention makes it possible for a mobile operator to adjust parameters relating to QoS attributes that govern the criteria for QoS satisfaction.

As mentioned, QoS satisfaction can be assessed for individual data flows taking into account multiple QoS aspects compared to prior art which only addresses individual QoS aspects. Measurements can then be made on the satisfaction of the data flows of a given QoS class for mobile stations individually; such measurements can indicate what fraction of the user population are satisfied with the QoS of a given QoS class (carrying, for example, conversational speech traffic).

The QoS satisfaction can also be assessed for individual mobile stations (considering all data flows of each mobile station), thus providing an operator the means to measure the fraction of the user population that are satisfied with the quality of all their traffic streams according to embodiments of the invention, while prior art performance measurements aggregate QoS measurements for a cell without taking into account the end user aspect.

Therefore, the operator is able to tune satisfaction thresholds for each of the QoS aspects of each data flow; QoS aspects may be nulled completely if desired so they do not influence the satisfaction of a data flow. This gives the operator control over how data flow satisfaction is determined, allowing adaptation to different types of traffic, different grades of subscribers, subscriber feedback to the operator, etc. Brief Description of the Drawings

The appended drawings are intended to clarify and explain different embodiments and aspects of the present invention in which:

Figure 1 shows System Context A;

Figure 2 shows System Context B;

- Figure 3 shows an embodiment of the present invention;

Figure 4 shows another embodiment of the invention;

Figure 5 shows a flowchart of a method according to the invention;

Figure 6 shows a flowchart of an embodiment of a method according to the invention; Figure 7 illustrates communication between a first node and a second node in a wireless communication system;

Figure 8 shows a NRM "Containment/Naming and AssociationX" according to the invention;

- Figure 9 shows a NRM "Inheritance Hierarchy_X" according to the invention; and

Figure 10 illustrates mapping of services onto different DRBs.

Detailed Description of Embodiments of the Invention

The present invention relates to a method for determining QoS fulfilment of one or more data flows between a first communication node Nl and a second communication node N2 in a wireless communication system. Preferably, the first communication node Nl is base station and the second node N2 is a mobile station arranged to be in communication with the first node Nl by means of one or more data flows which are transmitted on at least one radio link between the first Nl and second N2 nodes.

Each data flow is associated with one or more QoS attributes, and the present method comprises the steps of: measuring at least one measurement value for each one of the one or more QoS attributes; comparing the at least one measurement value for each one of the one or more QoS attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and determining the QoS fulfilment of the one or more data flows based on the one or more satisfied or non- satisfied representations. A flow chart of the method is shown in Figure 5.

A data flow represents a bundle of one or more user application streams that are handled equally with respect to QoS by a RAN, and the RAN does not look inside each data flow to identify individual application streams, instead the RAN treats all packets of the data flows equally. Thus a data flow has its own QoS attributes and QoS requirements or needs. A mobile station can support multiple data flows simultaneously. In Universal Mobile Telecommunications System (UMTS) or LTE/LTE-A a data flow is a Data Radio Bearer (DRB) - the RAN is responsible for managing the QoS of each DRB admitted to the system. In General Packet Radio Service (GPRS) QoS differentiation is applied between different Packet Data Protocol (PDP) contexts of a mobile station, so a data flow in this system represents a set of services mapped to one PDP context. In Worldwide Interoperability for Microwave Access (WiMAX), "service flows" are used to represent "data flows". In a preferred embodiment of the invention, the method steps in the method according to the invention are performed by a base station if the system is a LTE or a LTE-A; or performed by a base station or a RNC if the system is a UMTS. However, if the system is a GPRS

communication system, the steps are performed by a Base Station Controller (BSC); and if the system is a WiMAX system, the steps are performed by a base station or an access service network gateway.

QoS fulfilment of a data flow means that the QoS satisfaction for each of its QoS attributes are met, which e.g. implies that a delay threshold or an error rate threshold is met for a QoS attribute. The QoS attributes can belong to the group comprising the attributes: air interface packet error rate, packet delay, packet delay percentile (e.g. xl, see below), drop rate of untransmitted packets, active bit rate, blocking or termination of a data flow due to congestion or poor radio condition, and active bit rate subject to minimum buffer occupancy. However, other relevant QoS attributes may be used together with the method according to the invention. Figure 7 schematically illustrates communication between a first node Nl and a second node N2 in a wireless communication system. The first Nl and second N2 nodes are in this example an eNB and a UE, respectively, and the communication is performed by means of one or more data flows on at least one radio link (i.e. a physical radio communication link between a cell and the mobile station) between the first Nl and second N2 nodes. It should be noted that the communication between the first Nl and second nodes N2 can be in the DL

(from the first node Nl to the second node N2) or in the UL (from the second node N2 to the first node Nl).

According to an embodiment of the invention a satisfied or non- satisfied representation corresponds to a Boolean true or false value (e.g. as a binary "0" or "1") which means that a measurement value associated with a QoS attribute has a QoS satisfaction or not. Further, the method comprises the step of: performing multiple input AND-operations using the one or more true or false values for determining the QoS fulfilment of the one or more data flows. A flow chart of this embodiment is shown in Figure 6.

According to the present invention, the step of determining the QoS fulfilment of the one or more data flows also means that the QoS fulfilment of the second node N2, to which the one or more data flows belong, can be determined by performing multiple AND-operations for the true or false representation for all data flows belonging to the second node N2.

However, it is also possible to perform multiple AND-operations for all data flows belonging to the second node N2 and to a given QoS Class Identifier (QCI) value, which gives a more narrow definition of the QoS fulfilment of the second node N2 than the former definition. Hence, different aspects of the QoS fulfilment of individual mobile stations can be determined with the present method which provides flexibility for the operator. The present invention exploits a hierarchical model of QoS. The hierarchical model is illustrated in Figures 3 and 4. The foundations of the hierarchy are definitions of the QoS satisfaction which e.g. in a LTE communication system relates to the satisfaction of radio bearers (RBs). Different QoS definitions, which relates to different QoS attributes, can be employed for Guaranteed Bit Rate (GBR) and non-GBR RBs, and for the DL and UL between a first Nl and a second node N2.

Based upon the RB satisfaction values (e.g. Boolean true or false) the overall QoS satisfaction of a UE may be determined, e.g. by using logic multiple input AND-operations on the QoS satisfaction values of each RB of the UE. Hence, definitions of QoS satisfaction of RBs for a UE can be exploited to determine the QoS satisfaction of single UEs as shown in Figure 3, which also schematically shows how the different QoS attributes and thresholds employed can be tuned by a mobile operator.

Similarly, according to an embodiment of the invention the QoS satisfaction of all RBs belonging to a specific UE with the same QCI value can be determined, which is shown in Figure 4. Since there are typically DL and UL RBs for a given QCI value, separate tuning lines are shown for these QoS attributes and thresholds in said Figure. Instead of a Boolean combination of satisfaction values as described above a combination using fuzzy logic may be used. In this embodiment, the truth of the satisfaction of a QoS attribute is given a ranking between 0 and 1, where 1 indicates "true" and 0 indicates "false". To determine the QoS fulfilment of a data flow, the AND operation is replaced by taking the minimum of these truth values (across the set of QoS attributes). To determine the satisfaction of a second node N2, the minimum of the truth values across the data flows of the second node N2 is taken (and similarly to determine the QoS fulfilment of the data flows of a particular QCI for a second node N2). The QoS fulfilment of a second node N2 measured with fuzzy logic can be converted to a Boolean value (e.g. for a performance measurement report to the NM) by a simple comparison with a threshold. Typically this threshold would be set by the NM in a similar way to other thresholds exploited in the present invention.

The inventors have also realised that the attainment of QoS by an eNB is not a black and white area. They have further realised that it is important to be able to adjust the criteria for QoS attainment according to mobile operator's needs, e.g. based upon the service mix for a particular QCI value, or based on customer feedback, since a notable aspect of the existing performance counters is that they do not exploit any tunable parameters. Therefore, the invention according to an embodiment of the invention provides a solution in which the outcome of a performance counter, e.g. percentage of satisfied UEs, is influenced by a parameter that may be set by e.g. a NM according to operator's preferences. This gives the operator control over how data flow satisfaction is determined, allowing adaptation to different types of traffic, different grades of subscribers, subscriber feedback to the operator, etc. Most of the following preferred embodiments of the invention are illustrated in a LTE or a

LTE-A communication system context. However, the present invention is not limited to such communication systems, which will be clear from the following disclosure.

In a LTE communication system an eNB manages the QoS of RBs. A RB represents a logical resource for carrying data to or from a UE, and RBs can carry either signalling (signalling RBs) or user-plane traffic (i.e. DRB). For the purposes of this disclosure RBs and DRBs are used synomonously. A RB corresponds to a data flow within the meaning of the present disclosure. An individual UE may be configured with multiple RBs, both on the DL and on the UL. Multiple RBs are configured when the services they carry require different levels of QoS. For example, if a UE is involved in a speech call and at the same time is performing a FTP download then two RBs would be configured on the DL; one carrying speech packets and the other FTP packets. The speech service has tight latency requirements but is relatively robust to packet loss, whilst the FTP service has opposite requirements.

Every RB in a LTE system is associated with one of nine Quality of service Class Identifier (QCI) values, wherein each QCI value has three different attribute values, namely: priority, Packet Delay Budget (PDB), and Packet Error Loss Rate (PELR); however CQI8 and CQI9 values have the same attribute values for PDB and PELR.

Priority indicates the importance of meeting the PDB for a specific QCI value. PELR represents the loss rate over the air interface only, i.e. discarding of packets that have not been attempted to be sent over the radio channel is excluded. The standardized QCIs (QCI1-QCI9) in LTE are given below in Table 1 taken from 3GPP TS23.203. In the above example, the speech service would be mapped to a RB using QCI1, and the FTP service to a second RB using QCI8 or QCI9.

Table 1: Standardized QCI characteristics

QCI Resource Priority PDB PELR Example services Type

1 2 100 ms 10^"2 Conversational Voice

2 4 150 ms lo-³ Conversational Video (Live Streaming)

GBR

3 3 50 ms lo-³ Real Time Gaming

4 5 300 ms 10^-6 Non-Conversational Video (Buffered

Streaming)

5 1 100 ms 10^-6 IMS Signalling

6 Video (Buffered Streaming)

6 300 ms 10^-6 TCP-based (e.g. www, e-mail, chat, ftp, p2p file sharing, progressive video, etc.)

7 Non-GBR Voice,

7 100 ms lo ³ Video (Live Streaming)

Interactive Gaming

8 Video (Buffered Streaming)

8 300 ms 10^-6 TCP-based (e.g., www, e-mail, chat, ftp, p2p file

9 9 Sharing, progressive video, etc.

Furthermore, RBs are either GBR RBs or non-GBR RBs. QCI1 to CQI4 are used for the former type, while QCI5 to QCI9 for the latter. One key difference between GBR and non- GBR RBs is that a GBR RB has an additional attribute, the GBR attribute (in bits/sec). If the services mapped to the RB offer an aggregate load which is less than or equal to this GBR then the QoS attributes in the table, PELR, etc., should be met by an eNB. If the offered load goes beyond the GBR no QoS guarantee is necessarily required to be offered by the eNB. A second difference is that delay budget values for non-GBR RBs are soft requirements and longer delays are permitted. In fact, because of the statistical nature of traffic and the variability of the radio interface even the delay requirement for GBR RBs is not expected to be met by every packet. It should be noted that there are no requirements on packet discard during congestion, only on air interface losses.

Of further interest to this invention is the PM IRP. The PM IRP allows a NM to monitor the performance of a NE. The present invention is focused on monitoring the performance of the RAN, and in a LTE system this means monitoring the performance of an eNB. In a LTE system, monitoring means taking measurements of different radio aspects at the eNB and sending these to the NM.

As discussed in the previous section QoS performance measurements have been standardized by 3GPP. However, these standardized QoS performance measurements are all cell-based measurements, not UE-based. For example, a performance counter for "Average DL PDCP SDU delay" is defined in TS32.425, and this specification references a detailed calculation in specification TS36.314. Looking at the formula in the latter specification, the packet delays for all packets of a given QCI are accumulated (added together) and then this sum is divided by the number of packets (for the given QCI). There is no attempt to distinguish between packets to individual UEs.

The cell-based QoS counters defined in TS32.425 are unable to show the QoS performance of individual UEs. For example, if the DL packet delay for a given QCI is greater than the PDB for that QCI we do not know if all users are suffering poor performance or if the delay issue is isolated to few UEs, e.g. those in poor radio conditions. Therefore, the present invention provides a solution which gives performance measurements of the QoS satisfaction of individual UEs in a wireless communication system. In LTE a GBR RB has the following QoS attributes according to the TS23.203 specification:

• QCI value which implies: a PELR which represents packet losses over the air interface; a PDB which represents the delay budget between Policy Control Enforcement Function (PCEF) and the UE (the PCEF lies within the PDN Gateway which is the core network node that interfaces to the wide area internet); and a priority which is used internally by the scheduler during congestion;

• A guaranteed bit-rate which represents a ceiling to the sending rate that the service mapped to the RB should employ if the PELR and PDB are to be met. If the sending rate exceeds the guaranteed bit-rate the QoS offered by an eNB is not specified and is open to implementation; and · Allocation Retention Priority (ARP) is used to prioritise admission and retention of the

RB. Also, during heavy congestion it is possible that (IP) packets may be discarded before they have had a chance to be sent over the air interface. A counter is defined that captures the "DL PDCP SDU drop rate" representing the fraction of incoming IP packets that are discarded; note that this is a cell-based counter. It should also be noted that this is distinct from air interface losses (see PELR), and it is important to include such congestion losses within the remit of RB satisfaction. TS23.203 states what QoS requirements a GBR RB expects:

"Services using a GBR QCI and sending at a rate smaller than or equal to GBR can in general assume that congestion related packet drops will not occur, and 98% of the packets shall not experience a delay exceeding the QCI's PDB". In addition, the packet loss rate over the air interface should not exceed the PELR for the QCI.

Hence, according to an embodiment of the invention a DL GBR RB is satisfied if:

• Air interface (IP) packet error rate < PELR_QCI AND

• xl-percentile delay measured across all packets is < (PDB_QCI- delay _correction) AND

• Congestion packet drop rate < thresholdQci AND

• The RB is not terminated because of congestion or poor radio conditions AND

• The RB is not blocked because of congestion or poor radio conditions,

where air interface (IP) packet error rate is the fraction of packets unsuccessfully delivered over the air interface; xl-percentile delay is the delay value which x % of the packets have a delay less than; and congestion packet drop rate is the fraction of incoming packets that are discarded without any attempt being made to transmit them over the air interface.

Furthermore, PELR_QCI is the PELR value defined for the QCI (see Table 1), PDB_QCI - delay ^correction is the PDB value defined for the QCI (see Table 1), and threshold_Qci is a threshold parameter configured by the operator representing the acceptable congestion packet drop rate; and where "AND" in this application denotes a logic AND-operation using true or false input representations (Booleans).

Here, the xl delay is included because the delay guarantee is not absolute. TS23.203 suggests that xl is 98 percent for satisfaction, but here it is configurable by the operator to values other than 98. Similarly, the TS23.203 specification also suggests that the thresholdQci should be zero for GBR RBs, but again it is suggested that said threshold should be configurable. Finally, if the RB is not available because it has been terminated because of congestion or poor radio conditions then no traffic can be delivered and the RB is not satisfied overall.

Similarly, if the addition of a RB is blocked by the eNB when an attempt is made to establish the RB, then this counts as a dissatisfied RB. Typically this would occur when the eNB considers that the QoS needs of the candidate RB cannot be met because of congestion in the cell (there is not enough spare capacity to satisfy the RB), or the radio conditions of the UE limit the bandwidth that it can support. Thus the RB must be existing and available to be able to offer QoS satisfaction. Furthermore, if the service is not compliant to its GBR the RB is marked as satisfied since any scheduler behaviour is legitimate in these circumstances.

For UL GBR RBs it is not possible to accurately estimate the packet delay or any packet drops at the eNB, thus according to another embodiment of the invention an UL GBR RB satisfied if:

• Air interface (IP) packet error rate < PELR_QCIANO

• The RB is not terminated because of congestion or poor radio conditions AND

• The RB is not blocked of congestion or poor radio conditions.

A non-GBR RB has the following attributes according to the TS23.203 specification:

• QCI value which implies: a PELR which represents packet losses over the air interface; a PDB which represents the delay budget between the PCEF and the UE; and a priority which is used internally by the scheduler during congestion; and

• ARP is used to prioritise admission and retention of the RB.

Again TS23.203 states what QoS requirements a non-GBR RB expects: "Services using a Non-GBR QCI should be prepared to experience congestion related packet drops, and

98 percent of the packets that have not been dropped due to congestion should not experience a delay exceeding the QCI's PDB". In addition, the PELR should be met. The QoS satisfaction criteria could mirror those of the GBR RB requirements. However, it has been postulated that for non-GBR RBs it is perhaps more important to assess the data rate that is achieved when there is data queued for the RB - a form of active data rate. This is particularly useful for applications, such as http and ftp, since it gives a feel for "speed" experienced by an end user. However, to ensure that applications with very low data rates (e.g. MSN) do not spuriously impact the QoS assessment a configurable minimum data buffered requirement is included. Note that the delay metric is retained, but this can be disabled by setting xl to zero.

Thus, according to yet another embodiment of the invention a DL non-GBR RB is satisfied if:

• Air interface (IP) packet error rate < PELR_QCI AND

• xl-percentile delay measured across all packets is < (PDB_QCI- delay ^correction) AND

• Congestion packet drop rate < threshold_Qci AND

• Active rate > target_rateQci AND maximum volume of data buffered > min_bufferQci AND

• The RB is not terminated because of congestion or poor radio conditions AND

• The RB is not blocked because of congestion or poor radio conditions.

The active data rate is the data rate measured when there is data queued for the RB. The operator may set the values target_rate_Qci and min_buffer_QCi. It should be noted that the active rate requirement can be nulled by setting the target_rateQci to zero.

On the UL for non-GBR RBs it is not possible to measure delay accurately at the eNB so a throughput metric is also useful here, thus according to another embodiment of the invention an UL non-GBR RB satisfaction occurs if:

• Air interface (IP) packet error rate < PELR_QCI AND

• The RB is not terminated because of congestion or poor radio conditions AND

• The RB is not blocked because of congestion or poor radio conditions. It should also be understood that the above criterions relating to DL GBR, UL GBR, DL non- GBR and UL non-GBR may be may be applied at the same time for a specific UE, or one or more of the criterions may be applied in different combinations depending on the intended use. Further, a UE typically holds multiple RBs with the same QCI value. Typically, there is at least one DL RB and at least one UL RB. Additional RBs for the same QCI may be employed if services require the same QCI properties but different GBR values (assuming a GBR QCI), or ARP values (or both). Thus according to another embodiment QoS fulfilment is achieved for a UE if all RBs that that UE holds for one QCI are satisfied if every RB of that QCI meets the satisfaction requirements above. This assessment is particularly useful if the operator is interested in QoS satisfaction for one particular service offering, e.g. conversational VoIP using QCI 1. Continuing with this example, the operator can assess if end users are happy with the VoIP (speech) quality, and this works even if the mobile station acts as a hub supporting multiple speech calls at the same time.

According to yet another embodiment a UE is satisfied if all of the RBs for that UE meet the satisfaction requirements irrespective of their QCI. This embodiment allows the operator to assess the satisfaction of an end user as a whole, even when the end user is running multiple applications over the radio channel at the same time.

Furthermore, there are some examples of performance measurements, in this case

performance measurement counters, according to the invention, which can be taken in a cell or in a collection of cells:

• Counter 1: percentage of satisfied UEs, where a UE is satisfied if all of the RBs (data flows) for that UE meet the satisfaction requirements irrespective of their QCI. The formula for this counter 1 is therefore: 100* N2_Fui/N2_Totai, where N2_Fui denotes the number of UEs with QoS fulfilment, and N2r₀tai denotes the total number of UEs with at least one RB;

• Counter 2: percentage of UE having RBs (data flows) with the same QCI value are all satisfied for that specific QCI value. There can be one such performance counter per

QCI value. The formula for this performance counter is: 100* N2FuiQc N2TotaiQci, where N2p_ui_Qci denotes the number of UEs with QoS fulfilment for a given QCI value, and N2_TotaiQci denotes the total number of UEs having at least one RB with that given QCI value.

An optional extension of the invention would be to exclude UEs that have RBs established but no queued data during the period over which the measurement takes place for counter 1. Similarly, for counter 2 for QCI equal to "QCF UEs which have RBs established for QCI value QCF but have no queued data during the period could be excluded.

An example is given below in Table 2 on how to determine counters 1 and 2, respectively, described above. In Table 2, "O" indicates a RB that has a QoS fulfilment, and "X" indicates a RB that has not a QoS fulfilment for a specific QCI value. There are 6 UEs in the cell in this example, and they have 1, 2 or 3 RBs each. UE1, e.g. has one RB with QCI1 (satisfied), and two RBs with QCI9 (both satisfied), so UE1 is satisfied overall (last column). UE3 is not satisfied overall because its first RB is not satisfied. Counter 1 is determined using the values in the last column; whilst counter 2 (last row) considers the satisfaction of each user for a given QCI value.

Table 2: Performance measurement counters 1 and 2 given RB satisfaction

UEs are

satisfied

Counter 2: UE 1 UE 3 is UE 1 and

UE and 2 satisfied, 6 are

satisfaction are so 100% satisfied

for QCI satisfied, for QCI 9

value UE 3 while UE

and 5 4 is not,

are not, counter is

so 50% 66%

As described above, the invention also provides means for tuning QoS attributes and corresponding threshold values. Table 3 below summarises some proposed tunable parameters according to an embodiment of the invention, and these are the same parameters as given in the satisfaction criteria conditions above. It should be clear that the DL and UL parameters may be treated independent, and the same applies to GBR and non-GBR parameters in this and the following embodiments of the invention.

Table 3: Tunable parameters I

A more advanced scheme according to another embodiment will allow the setting of xl for each QCI value individually, which is shown in Table 4 below. This allows the operator to decide upon different percentile delay values in the assessment of RBs - a useful feature if services mapped to different QCIs have different tolerances to delay variation in the packets. For example, a streaming video service mapped to QCI2 can probably tolerate a lower xl than a VoIP speech call mapped to QCI1; more packets with delay greater than the PDB value of the QCI can be accepted because a large playback buffer is typically employed at the mobile station for video streaming.

Table 4: Tunable parameters II DRB Tunable parameters/thresholds

DL GBR XIQCI, thresholdQci

DL non-GBR XIQCI, thresholdQci, DL_target_rateQci, DL_min_bufferQci

UL GBR None

UL non-GBR UL_target_rateQci, UL_min_bufferQci

To clarify how mobile operators may tune some of the above mentioned parameters, two examples are given below. Example 1 : an operator is receiving reports of poor speech quality from customers,

particularly the receive quality at the handset. The performance measurements of user satisfaction for QCI1 are good (99% satisfied), GBR_xl parameter is set to 95%, and thresholdQci parameter is set to 5%. Action: the poor quality is either due to packet loss or to delay variations (jitter). To determine the cause the operator can in turn tighten up the parameters GBR_xl and

thresholdQci, and see what happens to the user satisfaction level. This can identify the problem/problems and actions can be taken. Air interface loss is acceptable because the overall user satisfaction is high. So to begin with, the GBR_xl parameter can be increased in steps of 1% until the user satisfaction level falls. Then the GBR_xl parameter can be reset to 95% and the threshold_Qci parameter can be reduced in steps of 1%.

Example 2: an operator wants to know if he can advertise that web browsing users will achieve a rate of at least 640kb/s when they are active. Previously no such guarantees were made.

Action: HTTP is mapped to QCI9 (which is a choice of the operator). The operator sets a DL_target_bitrate of 640kb/s for QCI9 and check the satisfaction of users with this QCI value. If the capacity of the system and the service mix/traffic load are not suitable to achieve this active bit rate it will be visible in the performance measurements. To disable other aspect of QoS satisfaction the non-GBR_xl , and thresholdQci parameters can be set to 0% and 100%, respectively. Then any dissatisfaction is only due to the active bit-rate criterion. The tunable parameters above may be defined, set, configured, controlled and/or tuned using at least two different approaches, namely one approach in which the NRM management mechanism is used, and another approach using an Interface IRP management mechanism.

With an embodiment corresponding to the NRM approach, new NRM attributes for the QoS variable thresholds are defined - e.g. a new NRM attribute for xl -percentile; another NRM attribute for thresholclQci, etc. If the tunable parameters take on the same values for the whole communication network, these new NRM attributes can be added to a new IOC or to an existing IOC applicable for the whole network. Table 5 is an example of a new IOC named "QoSSet". All the parameters outlined in Table 3 (Tunable Parameters I) are captured in Table 5. Figure 8 shows an example of the NRM Containment/Naming and AssociationX for a new IOC named "QoSSet", and Figure 9 shows an example of the NRM Inheritance Hierarchy_X for a new IOC named "QoSSet".

Table 5: Attribute table for QoSSet

UL_min_buffer_qci5 M M M

UL_min_buffer_qci6 M M M

UL_min_buffer_qci7 M M M

UL_min_buffer_qci8 M M M

UL_min_buffer_qci9 M M M

Table 6 below shows all the parameters from Table 4 (Tunable Parameters II) captured as IOC.

Table 6: Attribute table for QoSSet (extended variant)

The attributes in Table 5 or 6 may be set using the current Basic/Bulk CM IRP - these are Configuration Management IRPs which may be used to configure attributes at the NE (3GPP TS 32.600).

Each attribute has a name and an associate 3-tuple supportQualifier, readQualifier, and writeQualifier where:

• The supportQualifier indicates whether the attribute is Mandatory (M), Optional (O), Conditional-Mandatory (CM), Conditional- Optional (CO), SS-Conditional (C) or Not supported (— );

• The readQualifier indicates whether the attribute shall be readable by the

"IRPManager". Allowed values are: Mandatory (M), Optional (O) and Not supported (— ); and

• The writeQualifier indicates whether the attribute shall be writeable by the

"IRPManager". Allowed values are: Mandatory (M), Optional (O) and Not supported (— ).

According to another embodiment of the invention corresponding to the Interface IRP approach QoS variable thresholds setting operations are defined by reusing some existing IRPs or by defining a new Interface IRP for QoS management.

The struct of a new operation can be a list, in which an element is struct { ThresholdName, ThresholdValue} . There is one struct for each attribute. There is at least one attribute to be set in every operation request. Here, the ThresholdName is "threshold_qcil/2/../x", corresponding to the "Attribute Name" in above tables, the ThresholdValue is the attribute value of the relative ThresholdName. For example, the input parameters of a new operation named "SetQoSSet" can be { {threshold_qcil, 95 },{ threshold_qci2, 95 },{ threshold_qci3, 95%} }.

By these new operations, the managing system, e.g. by means of a NM, can control the setting for the QoS variable thresholds. As a response to the set operation, a managed system, e.g. EM and/or NE could send responses, e.g. result, legal value: ENUM (SUCCESS, FAILURE), to the managing system, in this case the NM. The IRP operation approach can also be used for the Itf-S and/or the direct interface configuration.

According to yet another embodiment of the invention additional tunable parameters can be defined that are Boolean in type, which means that a satisfaction criterion for a QoS attribute will be either applied or not according to the true/false value of a parameter. For example, the criterion that the air interface (IP) packet error rate < PELR_QCI could be applied or not according to an attribute "Apply_pelrQCI". This provides a very simple interface to the operator to disable/enable QoS aspects in the satisfaction. Note that some QoS attributes can be effectively disabled by setting extreme values for the thresholds, an example of which was discussed above. Further, in another embodiment, if the parameter is not set by the NM the criterion is assumed to be not applicable.

Figure 10 illustrates mapping of services onto RBs for a single LTE UE. In LTE, QoS differentiation in the RAN is performed at the RB level, i.e. between RBs, and the RAN does not look at different flows mapped to the same RB. A similar picture applies for UMTS systems.

As discussed above GPRS systems offer the ability for a user to support multiple PDP contexts, each with its own QoS Profile. So the packet flow(s) mapped to one PDP context receive the same QoS - so these are similar to RBs in LTE. Therefore, the present invention can also be applied to GPRS systems in which the satisfaction of individual UEs with multiple PDP contexts is treated in the same manner as for those systems above which employ RBs.

UMTS also uses the concept of RBs in a similar way as in a LTE system. Therefore, the present invention can also be applied to UMTS communication systems, which is well understood by the skilled person. However, the details of the RB satisfaction criteria will differ since UMTS systems employ different QoS framework compared to LTE system.

WiMAX systems also offer QoS differentiation between different "service flows". Therefore, when applying the present invention to a WiMAX system the satisfaction/fulfilment of individual service flows is assessed. The invention also relates to a method in a first communication node Nl for determining QoS fulfilment of one or more data flows between the first communication node Nl and a second communication node N2 in a wireless communication system. The one or more data flows are transmitted on at least one radio link between the first Nl and second N2 communication nodes, and each data flow is associated with one or more QoS attributes. The method comprises the step of: measuring at least one measurement value for each one of the one or more QoS attributes; comparing the at least one measurement value for each one of the one or more QoS attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and determining the QoS fulfilment of the one or more data flows based on the one or more satisfied or non-satisfied representations.

The above method in a first communication node Nl can also be modified according to the different embodiments of the method in a wireless communication system described above.

In accordance with the method in a first communication node described above the invention also relates to a first communication node Nl apparatus. The apparatus is arranged for wireless communication with a second communication node N2 apparatus in a wireless communication system. The communication system employs one or more data flows for communication between the first Nl and second N2 communication node apparatuses, wherein the one or more data flows are transmitted on at least one radio link between the first Nl and second N2 communication node apparatuses, and each data flow is associated with one or more QoS attributes. The first communication node Nl apparatus is configured for: measuring at least one measurement value for each one of the one or more QoS attributes; comparing the at least one measurement value for each one of the one or more QoS attributes to a corresponding threshold value for obtaining one or more satisfied or non-satisfied representations; and determining the QoS fulfilment of the one or more data flows based on the one or more satisfied or non-satisfied representations. It should be understood that the first communication node Nl apparatus above corresponds to the method in a first communication node Nl, and therefore that the first communication node Nl apparatus can be arranged in accordance with the different embodiments of said method. Furthermore, as understood by the person skilled in the art, a method for determining the QoS fulfilment according to the present invention may be implemented in a computer program, having code means, which when run in a computer causes the computer to execute the steps of the method. The computer program is included in a computer readable medium of a computer program product. The computer readable medium may consist of essentially any memory, such as a ROM (Read-Only Memory), a PROM (Programmable Read-Only Memory), an EPROM (Erasable PROM), a Flash memory, an EEPROM (Electrically Erasable PROM), or a hard disk drive.

Finally, it should be understood that the present invention is not limited to the embodiments described above, but also relates to and incorporates all embodiments within the scope of the appended independent claims.

Claims

1. Method for determining quality of service (QoS) fulfilment of one or more data flows between a first communication node (Nl) and a second communication node (N2) in a wireless communication system, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication nodes, and each data flow is associated with one or more quality of service (QoS) attributes, said method being characterised by the steps of:

2. Method according to claim 1, wherein said one or more satisfied or non-satisfied representations correspond to one or more logic true or false values, and said method further comprises the step of:

- performing multiple input AND-operations using said one or more true or false values for determining the quality of service (QoS) fulfilment of said one or more data flows.

3. Method according to claim 1, wherein said one or more quality of service (QoS) attributes belong to the group comprising: air interface packet error rate, packet delay, packet delay percentile, drop rate of untransmitted packets, active bit rate, termination or admission blocking of a data flow due to congestion or poor radio condition, and active bit rate subject to minimum buffer occupancy.

4. Method according to claim 1, wherein said wireless communication system is a long term evolution (LTE), long term evolution advanced (LTE-A), general packet radio service (GPRS), universal mobile telecommunications system (UMTS) or a worldwide interoperability for microwave access (WiMAX) communication system; said first communication node (Nl) is a base station (BS), such as a Node B or a eNB; said second communication node (N2) is a mobile station (MS), such as a user equipment (UE); and said one or more data flows are downlink (DL) or uplink (UL) data flows between said first (Nl) and second (N2) communication nodes.

5. Method according to claim 4, wherein if said wireless communication system is a long term evolution (LTE) or a long term evolution advanced (LTE-A) communication system, said at least one measurement and/or said one or more quality of service (QoS) attributes and/or said corresponding threshold value being defined, set, configured and/or controlled by a network manager (NM) or a element manager (EM) by using an interface integration reference point (IRP) or a network resource model (NRM), wherein said network manager (NM) or element manager (EM) are connected to said base station (BS) by means of a communication interface, such as interface N or interface S.

6. Method according to claim 4, wherein if said wireless communication system is a long term evolution (LTE) or a long term evolution advanced (LTE-A) communication system said method steps are performed by said base station (BS); if said wireless

communication system is an universal mobile telecommunications system (UMTS) said method steps are performed by said base station (BS) or a radio network controller (RNC); and if said wireless communication system is a general packet radio service (GPRS) communication system said method steps are performed by a base station controller (BSC), and if said wireless communication system is a worldwide interoperability for microwave access (WiMAX) communication system said method steps are performed by a said base station or an access service network gateway.

7. Method according to claim 4, wherein said one or more data flows are one or more data radio bearers (DRBs); each data radio bearer (DRB) having an associated quality of service (QoS) class identifier (QCI) value, and are guaranteed bit rate (GBR) data radio bearers (DRBs) or non-guaranteed bit rate (non-GBR) data radio bearers (DRBs).

8. Method according to claim 7, wherein said corresponding threshold value is determined from said quality of service (QoS) class identifier (QCI) value for said one or more data radio bearers (DRBs).

9. Method according to claim 7, wherein said corresponding threshold value is determined from whether said one or more data radio bearers (DRBs) are downlink (DL) or uplink (UL) data radio bearers (DRBs).

10. Method according to claim 7, wherein said corresponding threshold value is determined from whether said data radio bearers (DRBs) are guaranteed bit rate (GBR) data radio bearers (DRBs) or non- guaranteed bit rate (non-GBR) data radio bearers (DRBs).

11. Method according to any of claims 1-10, further comprising the step of:

- determining the quality of service (QoS) fulfilment of said second node (N2), wherein said second node (N2) has a quality of service fulfilment if all data flows belonging to said second node (N2) have quality of service (QoS) fulfilment.

12. Method according to any of claims 11, further comprising the step of:

- determining a performance measurement counter defined as:

where N2p_ui denotes the number of second nodes (N2) with quality of service (QoS) fulfilment, and N2r₀tai denotes the total number of second nodes (N2) with at least one data flow.

13. Method according to any of claims 7-10, further comprising the step of:

- determining the quality of service (QoS) fulfilment for a given quality of service (QoS) class identifier (QCI) value of said second node (N2), wherein said second node (N2) has a quality of service (QoS) fulfilment for a given quality of service (QoS) class identifier (QCI) value if all radio bearers (RB) belonging to said second node (N2) having said given quality of service (QoS) class identifier (QCI) value have quality of service (QoS) fulfilment.

14. Method according to claim 13, further comprising the step of:

- determining a performance measurement counter defined as: 100* N2_Fui_Qc N2_Totai_Qci, where N2_fui_Qci denotes the number of second nodes (N2) with quality of service (QoS) fulfilment for a given quality of service (QoS) class identifier (QCI) value, and N2_TotaiQci denotes the total number of second nodes (N2) having at least one radio bearer (RB) with said given quality of service (QoS) class identifier (QCI) value.

15. Method in a first communication node (Nl) for determining quality of service

(QoS) fulfilment of one or more data flows between said first communication node (Nl) and a second communication node (N2) in a wireless communication system, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication nodes, and each data flow is associated with one or more quality of service (QoS) attributes, said method being characterised by the steps of:

16. Method according to claim 15, wherein said one or more satisfied or non- satisfied representations correspond to one or more logic true or false values, and said method further comprises the step of:

17. Method according to claim 15, wherein said one or more quality of service (QoS) attributes belong to the group comprising: air interface packet error rate, packet delay, packet delay percentile, drop rate of untransmitted packets, active bit rate, termination or admission blocking of a data flow due to congestion or poor radio condition, and active bit rate subject to minimum buffer occupancy.

18. Method according to claim 15, wherein said wireless communication system is a long term evolution (LTE), long term evolution advanced (LTE-A), general packet radio service (GPRS), universal mobile telecommunications system (UMTS) or a worldwide interoperability for microwave access (WiMAX) communication system; said first communication node (Nl) is a base station (BS), such as a Node B or a eNB; said second communication node (N2) is a mobile station (MS), such as a user equipment (UE); and said one or more data flows are downlink (DL) or uplink (UL) data flows between said first (Nl) and second (N2) communication nodes.

19. Method according to claim 18, wherein if said wireless communication system is a long term evolution (LTE) or a long term evolution advanced (LTE-A) communication system, said at least one measurement and/or said one or more quality of service (QoS) attributes and/or said corresponding threshold value being defined, set, configured and/or controlled by a network manager (NM) or a element manager (EM) by using an interface integration reference point (IRP) or a network resource model (NRM), wherein said network manager (NM) or element manager (EM) are connected to said base station (BS) by means of a communication interface, such as interface N or interface S.

20. Method according to claim 18, wherein if said wireless communication system is a long term evolution (LTE) or a long term evolution advanced (LTE-A) communication system said method steps are performed by said base station (BS); if said wireless

21. Method according to claim 18, wherein said one or more data flows are one or more data radio bearers (DRBs); each data radio bearer (DRB) having an associated quality of service (QoS) class identifier (QCI) value, and are guaranteed bit rate (GBR) data radio bearers (DRBs) or non-guaranteed bit rate (non-GBR) data radio bearers (DRBs).

22. Method according to claim 21, wherein said corresponding threshold value is determined from said quality of service (QoS) class identifier (QCI) value for said one or more data radio bearers (DRBs).

23. Method according to claim 21, wherein said corresponding threshold value is determined from whether said one or more data radio bearers (DRBs) are downlink (DL) or uplink (UL) data radio bearers (DRBs).

24. Method according to claim 21, wherein said corresponding threshold value is determined from whether said data radio bearers (DRBs) are guaranteed bit rate (GBR) data radio bearers (DRBs) or non- guaranteed bit rate (GBR) data radio bearers (DRBs).

25. Method according to any of claims 15-24, further comprising the step of:

26. Method according to any of claims 25, further comprising the step of:

- determining a performance measurement counter defined as:

27. Method according to any of claims 21-24, further comprising the step of:

28. Method according to claim 27, further comprising the step of:

29. Computer program, characterised in code means, which when run in a computer causes said computer to execute said method according to any of claims 15-28.

30. Computer program product comprising a computer readable medium and a computer program according to claim 29, wherein said computer program is included in said computer readable medium, and consist of one or more from the group: ROM (Read-Only Memory), PROM (Programmable ROM), EPROM (Erasable PROM), Flash memory, EEPROM (Electrically EPROM) and hard disk drive.

31. A first communication node (Nl) apparatus arranged for wireless

communication with a second communication node (N2) apparatus in a wireless

communication system, said wireless communication system employing one or more data flows for communication between said first (Nl) and second (N2) communication node apparatuses, wherein said one or more data flows are transmitted on at least one radio link between said first (Nl) and second (N2) communication node apparatuses, and each data flow is associated with one or more quality of service (QoS) attributes, wherein said first communication node (Nl) apparatus is characterised in that being configured for: