WO2011072735A1

WO2011072735A1 - Coordinated transport and radio interface scheduling for a mobile communication network

Info

Publication number: WO2011072735A1
Application number: PCT/EP2009/067312
Authority: WO
Inventors: Hans Kroener
Original assignee: Nokia Siemens Networks Oy
Priority date: 2009-12-16
Filing date: 2009-12-16
Publication date: 2011-06-23

Abstract

The application provides a method of operating a base station (20) of a mobile communication network (10). The method comprises the step of determining system capacity of the radio interface (13) and system capacity of the transport network (21). The method comprises the further step of changing data rates of the plurality of the UEs (12) when the transport network (21) causes a bottleneck of system capacity of the mobile communication network (10). The method also comprises the further step of changing a radio interface capacity when the radio interface (13) causes the bottleneck of system capacity of the mobile communication network (10). The balancing between radio and transport capacity will be achieved by adjusting the trade-off between capacity of the radio interface and fairness of the data rate allocation to different UEs.

Description

DESCRIPTION

Coordinated transport and radio interface scheduling for a mobile communication network

This application relates to a mobile communication network. In particular, this application relates to scheduling for the mobile communication network. The mobile communication system uses radio cells. The radio cell refers to a geographical area that is assigned to a cor^¬ responding base station of the mobile communication system. The base station is connectable to a fixed communication net^¬ work. System capacity of the base station is usually quite high since its available signal bandwidth may be reused among its various cells. The fixed communication network may in^¬ clude a mesh network that comprises numerous switches con^¬ nected together by communication links. When a mobile user of the mobile communication system wishes to place a call, the call is transmitted through a communica^¬ tion medium, often a radio channel, to the base station that is assigned to the radio cell of the mobile user. From the base station, the fixed communication network carries the call to its intended destination. When the mobile user moves from one radio cell to another, a call handoff or handover between the base stations takes place.

The application provides a mobile communication network.

The mobile communication network provides wireless communica^¬ tion services for its users.

The mobile communication network comprises a transport net^¬ work that is communicatively connectable to a base station. The communicative connection enables transmission of data packets for carrying voice or data information. The base station receives data packets from the users of the mobile communication network via radio signals and forwards them to the transport network. The transport network then sends them to other parts of the mobile communication net- work. Similarly, the transport network receives data packets from the other parts of the mobile communication network and it forwards them to the base station for transmitting to the user . The application provides a radio interface scheduler of a base station. The radio interface scheduler rules or governs the data packets transmission between the base station and user equipments (UEs) that are communicatively connectable to the base station.

The user of the mobile communication network uses the UE for voice or data communication. The UE includes a hand-held telephone or a card in a laptop computer. It is believed that network system throughput of the mobile communication network may be improved by adapting behaviour of the radio interface scheduler to different situations.

When a transport network of the mobile communication network is a bottleneck for system capacity, the radio interface scheduler may be adapted to provide a fair or equal alloca^¬ tion of data rates to different active UEs (User Equipments) that are attached to a base station of the mobile communica^¬ tion network. The system capacity refers to amount of infor- mation or signal that the mobile communication network may handle .

Operational expenditure (OPEX) is a major concern for most mobile network operators. A significant part of the OPEX is incurred by costs for the transport network. In particular, cost that is incurred for a last mile transport of the trans^¬ port network is significant. Therefore, transmission rate of the transport network is usually limited to reduce cost. Furthermore, fairness is another major concern for most mo^¬ bile network operators. This may be seen in 3GPP (Third Gen^¬ eration Partnership Project) requirements for 3G Long Term Evolution (LTE) system, as described in 3GPP specification, 3GPP TR 25.913, http://www.3gpp.org/ftp/specs /html- info/25913.htm.

On the other hand, when the radio interface of the mobile communication network is the bottleneck for the system capacity, the radio interface scheduler may be adapted to optimize or to improve a radio interface capacity.

For this purpose, network traffic over the transport network as well as the network traffic over the radio interface over all involved radio cells is monitored.

Data packet load, data packet delay, and data packet loss of the network traffic over the transport network for the base station and over the radio interface for the radio cell is then considered or evaluated. The consideration may afterward be used to estimate or to determine which of the two bottle^¬ necks is limiting the system capacity. Based on this estimation, the radio interface scheduler is then tuned or adjusted such that system signal throughput and UE signal throughput with fairness for UE data rate through^¬ out a signal transmission chain is optimized or improved. In other words, global optimization or overall improvement is achieved.

The application advantageously allows the mobile network op^¬ erators to meet its objective of system capacity and fair^¬ ness.

Put differently, the application provides a method of operat^¬ ing a base station of a mobile communication network. The mo- bile communication network comprises at least one base sta^¬ tion .

The mobile communication network comprises a transport net- work that is communicatively connectable to the base station. The mobile communication network also comprises a plurality of UEs (User Equipments) that are communicatively connectable to the base station via a radio interface. The method comprises the step of determining system capacity of the radio interface and determining system capacity of the transport network. The system capacity of the radio capacity may be determined by a first determining section of the base station whilst the system capacity of the transport network may be determined by a second determining section of the base station .

Using this determination, a degree of fairness is adjustable by the changing of the data rates of the plurality of the UEs according to a scheduling strategy e.g., a fair scheduler

(47), a maximum carrier to interference ratio (CIR) scheduler (49), a proportional fair scheduler (51) . In particular, us^¬ ing the determination, data rates of the plurality of the UEs is changed when the transport network is causing a bottleneck of system capacity of the mobile communication network. The data rate of the UE may be changed by a first changing sec^¬ tion of the base station. Likewise, a radio interface capac^¬ ity is changed when the radio interface is causing the bot^¬ tleneck of system capacity of the mobile communication net- work. The radio interface capacity may be changed by a second changing section of the base station.

The change of the data rates of the plurality of the UEs may be done in a manner that increases a degree of fairness of the data rates of the plurality of the UEs. The degree of fairness refers to degree or amount of equal data rate of transfer for the UEs. In particular, the change of the data rates of the plurality of the UEs according to a first sched- uling strategy e.g., a fair scheduler (47) may ensure a converging of the data rates of the UEs, which may increase a degree of fairness of the data rates of the plurality of the UEs .

The change of the radio interface capacity may also be done in a manner that decreases a degree of fairness of the data rates of the plurality of UEs e.g., by exploiting channel de^¬ pendent scheduling gains or by assigning a data rate of a UE according to its radio channel conditions. In particular, the change of the radio interface capacity according to a second scheduling strategy e.g., a maximum carrier to interference ratio (CIR) scheduler (49) may ensure a prioritising of the data rates of the UEs according to their radio conditions. The change of the radio interface capacity according to the second scheduler may decrease a degree of fairness of the data rates of the plurality of the UEs and may increase the total radio interface capacity. This has the effect that the change of the radio interface capacity increases the total radio interface capacity. The increase of the radio interface capacity may occur for at least one radio cell with a prede^¬ termined characteristic. The predetermined characteristic re^¬ fers to a good or better radio condition that is able to ac^¬ cept transmission of a higher data rate without significant degradation of signal quality. The good radio connection al^¬ lows transmission of higher data rate.

The determination of system capacity of the radio interface may comprise a step of monitoring data packet traffic over a radio cell or radio cells of the radio interface to determine one or more data packet characteristics of the radio inter^¬ face packet traffic. The data packet characteristic may refer to data packet load information, data packet delay informa^¬ tion, or data packet loss information.

In addition, the determination of the system capacity of the transport network may further comprise a step of monitoring data packet traffic over a base station or base stations of the transport network to determine one or more data packet characteristics of the transport network packet traffic. The data packet characteristic may also refer to data packet load information, data packet delay information, or data packet loss information.

For determining the bottleneck of a system capacity of the mobile communication network, it is convenient to use one or more determined data packet characteristics of the radio in- terface packet traffic. In addition or alternatively, one or more determined data packet characteristics of the transport network packet traffic may be used.

In addition, the application provides a method of operating a mobile communication network that comprises one of the above- mentioned methods.

The application provides a base station of a mobile communication network.

The base station comprises a transport network and a plural^¬ ity of UEs (User Equipments) . The transport network is commu^¬ nicatively connectable to the base station whilst the plural^¬ ity of UEs are communicatively connectable to the base sta- tion via a radio interface.

The base station comprises a first determining section for determining system capacity of the radio interface and a sec^¬ ond determining section for determining system capacity of the transport network.

In addition, the base station includes a first changing section for changing data rates of the plurality of the UEs when the transport network causes a bottleneck of system capacity of the mobile communication network. The first changing section may include a transport scheduler. In a special case, the first changing section does not include the transport scheduler but works with or exchanges information with the transport scheduler.

The base station also comprises a second changing section for changing a radio interface capacity when the radio interface causes the bottleneck of system capacity of the mobile commu^¬ nication network. The second changing section may comprise a radio interface scheduler. In an alternative implementation, the second changing section does not comprise the radio in- terface scheduler. The second changing section works with or exchanges information with the radio interface scheduler.

A Digital Signal Processor may conveniently provide the first determining section and the second changing section. The Digital Signal processor comprises a microprocessor architec^¬ ture that processes signals via a digital means. The signal carries audio or streams of information in a digital format.

The digital signal may be converted from an analog signal. The processing uses numerical calculations. For example, the processing may remove noises from streams of information us^¬ ing Digital Signal Processing techniques. It may be part of a system-on-chip device, which contains analog and digital cir^¬ cuitries, or, alternatively, it may be in a form of a stand- alone device.

Similarly, a Network Processor may easily include the second determining section and the first changing section. The Network Processor comprises electrical circuits, which has a feature set that is intended for network application.

The network processor comprises programmable chips, like gen^¬ eral-purpose microprocessors, which are optimized for the packet processing, as required in network devices. The net- work devices include Internet equipment, such as routers, switches, Voice over IP (VoIP) bridges, and virtual private network (VPN) gateways. The base station may use 3GPP protocol. The 3GPP protocol is widely used and may be easily implemented for the base sta^¬ tion. The base station may also comprise a monitoring device to monitor data packet traffic over a radio cell or radio cells of the radio interface to determine data packet charac^¬ teristics of the radio interface packet traffic. These char^¬ acteristics may later be used to determine or compute the system capacity of the radio interface. Similarly, the base station may also comprise a monitoring device to monitor data packet traffic over the transport net^¬ work to determine data packet characteristics of the trans^¬ port network packet traffic. Afterward, these characteristics may be used to determine or compute the system capacity of the transport network.

The application provides a mobile communication network. The mobile communication network comprises one or more base sta^¬ tions .

The application provides a computer program of a base station for executing one of above-mentioned methods.

The computer program has the advantage of flexible implemen- tation of the methods. The method may be realized in a quick manner and any error in the realization may be fixed easily.

The application provides a data medium for holding above- mentioned computer program and a computer system of a base station for executing the above-mentioned computer program.

The computer system comprises one or more processors that are connected to computer memories. The above-mentioned computer program is loaded into the memory. The processor may include a Digital Signal Processor for per^¬ forming tasks that includes running or executing a radio interface scheduler. Alternatively, the processor may also in^¬ clude a Network Processor for performing tasks that includes running or executing a transport scheduler. The Network Processor includes an electrical circuit that has a feature set that is used for network application. The application provides an apparatus of a mobile communica^¬ tion network, which is being configured to perform scheduling in a communications network. The apparatus comprises

- means configured to determine the capacity of the radio in^¬ terface ;

- means configured to determine the capacity of the transport network;

- means configured to change data rates of the plurality of UEs if transport network is causing a bottleneck of system capacity; and

- means configured to change a radio interface capacity if radio interface is causing a bottleneck of system capacity. The apparatus may be a part of a base station of the mobile communication network e.g., a computer system or a processor within the base station. An example of a processor within the base station is a Digital Signal Processor or a Network Proc^¬ essor. The apparatus advantageously allows the mobile network operators to meet its objective of system capacity and fair^¬ ness.

Fig. 1 illustrates an overview of an embodiment of a mo^¬ bile communication network,

Fig. 2 illustrates an example of traffic control mecha^¬ nisms of the mobile communication network of Fig. 1 that uses WCDMA (Wideband Code Division Multiple

Access) mechanism,

Fig. 3 illustrates an example of a trade-off between

throughput and fairness for different scheduling strategies of the mobile communication network of Fig. 1, and

Fig. 4 illustrates a high-level functional view of an im^¬ proved scheme for operating the mobile communica^¬ tion network of Fig. 1, and Fig. 5 illustrates an embodiment of an implementation of a base station of the mobile communication network of Fig. 1. Figs. 1 to 4 have similar parts. The similar parts have the same names or same reference numbers. The description of the similar parts is thus incorporated by reference.

Fig. 1 provides an overview of an embodiment of a mobile com- munication network 10.

The mobile communication network 10 comprises a Radio Access Network (RAN) 11 that is connected to multiple User- Equipments (UEs) 12 via a radio interface 13. The UEs 12 are also called Mobile Stations.

The RAN 11 is connected to a core network 15 that is con^¬ nected to a fixed line network 17. The fixed line network 17 is connected to a remote application 18.

The RAN 11 includes a base station (BS) 20, a transport net^¬ work 21, and other RAN elements 22. The BS 20 is also called a base transceiver station (BTS) , Node B, or eNode B. The BS 20 is connected to the UEs 12 via the radio interface 13 and to the transport network 21 that is connected to the other RAN elements 22.

The transport network 21 comprises a last mile access 24 that is connected to the BS 20 and a switching and/or routing net- work 26 that is connected to the last mile access 24. The last mile access 24 is also called a last mile transport.

In a generic sense, the other RAN elements 22 are an optional part and do not exist for all radio standards. The other RAN elements 22 may include a radio network controller (RNC) , as in the case of UMTS (Universal Mobile Telecommunications Sys^¬ tem) , for providing interconnection to the core network 15 of the mobile communication network 10. The last mile access 24 may include PDH ( Plesiochronous Digi^¬ tal Hierarchy) lines, DSL (Digital Subscriber Loop) lines or microwave transmission. The last mile access 24 may also in- elude SDH (Synchronous Digital Hierarchy) or Ethernet.

The UE 12 comprises a user and a terminal that is capable of communicating with the RAN 11. The terminal includes a mobile phone or a computing device.

The BS 20 receives network traffic from the UEs 12 and for^¬ wards it to the core network 15. The network traffic is also called traffic. Similarly, the BS 20 receives traffic from the core network 15 and sends it to the UEs 12. The traffic comprises signals or data packets.

The BS 20 communicates with the UEs 12 via different radio cells 28 with an appropriate multiple-access scheme, such as CDMA (Code Division Multiple Access) or OFDMA (Orthogonal Frequency Division Multiple Access) . The radio cells 28 may be separated either by different radio frequencies or by an^¬ tennas serving different geographical sectors.

The last mile access 24 acts as a communication channel be- tween the BS 20 and the routing and/or switching network 26. The last mile access 24 has a limited bandwidth. The switch^¬ ing and/or routing network 26 aggregates traffic towards the other RAN elements 22. The core network 15 utilizes the RAN 11 for radio communica^¬ tion. It provides circuit-switched transmission of voice sig^¬ nals and packet-switched transmission of data packet signals. The fixed line network 17 provides a line of communication on land. The remote application refers to a software service.

Downlink traffic towards the UE 12 may be sent from the fixed line network 17 via the core network 15 and the other op^¬ tional RAN elements 22. The downlink traffic also travels through the transport network 21 that includes the last mile transport 24 and through the BS 20. From the BS 20, the downlink traffic is sent via the corresponding radio cell 28 over the radio interface 13 to the target UE 12.

In the other direction, the UE 12 may send uplink traffic via the corresponding radio cell 28 to the BS 20. The BS 20 then sends the uplink traffic over the transport network 21 that includes the last mile transport 24 to the other optional RAN elements 22, to the core network 15, to the fixed line net^¬ work 17, and finally to the remote application 18.

Depending on the dimensioning of the network system, it is believed that two potential major bottlenecks 29 and 30 for system capacity exist in the RAN 11. The system capacity re^¬ lates to amount of data or information that the mobile commu^¬ nication network 10 may handle or manage. The bottleneck 29 may exist in the last mile transport 24 whilst the bottleneck 30 may exist in the radio interface 13.

System capacity of the last mile transport 24 is usually dy^¬ namically shared between all UEs 12 of the base station 20 whilst system capacity of the radio interface 13 of a certain radio cell 28 is dynamically shared by those UEs 12 that are connected to this radio cell 28. Furthermore, the sharing usually takes Quality of Service (QoS) requirements of the different traffic flows into account.

Fig. 2 depicts an example of a network traffic control mecha- nisms 32 of the mobile communication network 10 of Fig. 1. The mobile communication network 10 uses a WCDMA (Wideband Code Division Multiple Access) mechanism.

A traffic flow in the downlink is controlled by several con- trol elements or control loops.

Media access control protocols in the BS 20 and in the UE 12 usually handle an access to the RAN 11. More specifically, a radio interface scheduler 34 that resides in the BS 20 grants the access to radio resources. A Digital Signal Processor of the BS 20 runs or operates the radio interface scheduler 34. The Digital Signal Processor has a microprocessor architec- ture that processes signals via a digital means. The signal carries audio or streams of information in a digital format.

The radio interface scheduler 34 allocates the radio re^¬ sources to the different UEs 12. The allocation considers buffer status, quality of service requirements of the traf^¬ fic, as well as quality of radio link between the BS 20 and the UE 12.

Moreover, the base station 20 also includes a radio interface capacity determination device 35 and a data packet traffic monitoring device 39. In particular, the radio interface ca^¬ pacity determination device 35 computes system capacity of the radio interface of the base station 20 for use by the ra^¬ dio interface scheduler 34. The data packet traffic monitor^¬ ing device 39 tracks or monitors data packet traffic over a radio cell or radio cells of the radio interface to determine characteristics of the data packet. The data packet charac^¬ teristics can relate to data packet load information, to data packet delay information, or to data packet loss information.

In addition, a link adaptation unit adjusts modulation and coding formats to provide radio link quality. A higher data rate is selected for UEs with good channel quality whilst a lower data rate is selected for UEs with poor channel qual^¬ ity. The higher data rate is accomplished by a higher order modulation format or less channel coding. The link adaptation unit is not shown in the figure.

Appropriate packet transport schedulers 36 control an access to the transport network 21 according to its QoS require^¬ ments. The packet transport scheduler 36 comprises a DL transport scheduler that resides in the routing/switching network 26 and a UL transport scheduler that resides in the base station 20. A Network Processor of the BS 20 runs or operates the transport schedulers 36. In a special case, a dedicated piece of hardware runs the transport schedulers 36. The Network Processor includes electrical circuit that has a feature set that is used for network application. The Network Processor may be implemented using software programmable de^¬ vices.

In a general sense, the radio interface scheduler 34 can be operated or ran by the digital signal processor or by the network processor. Similarly, the transport schedulers 36 can also be operated or ran by the digital signal processor or by the network processor.

Further, the base station 20 also includes a transport net^¬ work capacity determination device 37 and a data packet traf^¬ fic monitoring device 41. The transport network capacity de^¬ termination device 37 computes system capacity of the trans^¬ port network 21 whilst the data packet traffic monitoring de^¬ vice 41 tracks or monitors data packet traffic over the transport network 21 to determine characteristic of the data packet. The data packet characteristics can refer to data packet load information, to data packet delay information, or to data packet loss information.

For the UMTS network, a third control loop is established be^¬ tween the BS 20 and a radio network controller (RNC) of the other RAN elements 22 via an Iub frame protocol. This proto^¬ col provides an Iub flow and congestion control 38 between the RNC of the other RAN elements 22 and the BS 20.

On top of these protocols, a radio link control (RLC) proto^¬ col provides a flow control 40 between the RAN 11 and the UE 12. For the WCDMA implementation, the RLC protocol is located in the RNC. For the 3G LTE (Long Term Evolution) implementation, the RLC protocol is located in the base station. Finally, for typical data applications like file transfer or web surfing, a Transmission Control Protocol (TCP) between the remote application 18 and the UE 12 provides an end-to- end TCP flow and congestion control 42.

The different control loops may work more or less independ^¬ ently. That means that each of the control loops may be opti^¬ mized or adjusted to the specific part of the mobile communi^¬ cation network 10 where it is applied without considering the other control loops.

More specifically, the radio interface scheduler 34 is usu^¬ ally adjusted in such a way that it focuses on radio inter^¬ face performance in terms of throughput and of fairness. The fairness refers to equal data rate of transfer for the UE 12.

The transport scheduler 36 relies on scheduling and QoS han^¬ dling mechanisms that are used in ATM (Asynchronous Transfer Mode) or in IP networks.

The Iub flow and congestion control 38 focuses on handling of the lub flow and congestion control issues.

The RLC flow control 40 is designed to hide non-ideal behav- iour of the mobile communication network 10 from the TCP congestion and flow control 42. The non-ideal behaviour relates mainly to cell loss.

The TCP congestion and flow control 42 is developed for the usual IP based networks and does not take into account the specifics of mobile communication networks.

Hence, some improvement potential exists when the different control mechanisms are considered jointly. In particular, it is believed that it makes sense to consider the two major bottlenecks 29 and 30 of the mobile communication network 10 jointly or to consider a closer coupling of the RAN 11 and the radio interface 13. The cooperation or closer coupling may avoid a situation wherein the radio interface scheduler 34 may try to optimize or to improve the radio interface capacity even where it is not required because the last mile transport 24 of the trans^¬ port network 21 is the bottleneck for system capacity. System capacity for throughput of the transport network 21 is always or mostly constant and it cannot be changed by allocating different data rates to different UEs 12.

Fig. 3 depicts an example of a graph 45 of trade-off between radio cell throughput and fairness between UE data rates for different scheduling strategies of the mobile communication system of Fig. 1, which is described or elaborated below.

The different multiple scheduling strategies can be used by the radio interface scheduler 34 and they can be selected based on location of bottleneck of system capacity. The scheduling strategies include a fair scheduler 47, a maximum CIR (carrier to interference ratio) scheduler 49, and a proportional fair scheduler 51.

The fair scheduler 47, that allocates to the UE 12 a same data rate, leads to a low cell throughput. In contrast, the maximum CIR scheduler 49 that allocates all radio resources to the UE 12 that has currently the best propagation condi^¬ tions achieves a high cell throughput. The typically used proportional fair scheduler 51 generates a cell throughput that is in between the low cell throughput and the high cell throughput. The proportional fair scheduler 51 allocates a fair amount of radio resources, which means that the data rate of the UE 12 is proportional to radio channel quality of the UE 12. This proportional fair sched^¬ uler 51 is mostly used for data services that does not re^¬ quire guaranteed throughput. A fundamental theorem of Shannon states that signal capacity of a channel depends on signal to noise ratio of the channel. The channel capacity relates to a maximum data transmission rate at which data is transferred without errors.

Therefore, modern radio standards like EDGE (Enhanced Data Rates for GSM Evolution) , WCDMA (Wideband Code Division Multiple Access) , WiMAX (Worldwide Interoperability for Micro^¬ wave Access) , or 3GPP LTE adapts modulation and coding scheme according to radio channel conditions. The radio channel con^¬ ditions is also called radio conditions.

The adaptation allows data to be transferred at a higher rate by using higher order modulation schemes with lower coding for the UE 12 when the radio conditions are favourable. Fa^¬ vourable here refers to higher signal to noise ratio. On the other hand, more robust lower order modulation schemes and higher channel coding is employed when the UE 12 has bad ra^¬ dio channel condition. The bad radio condition occurs mainly at the edge of the radio cell 28. The adaptation of the modu^¬ lation and coding scheme to the radio conditions is performed by a link adaptation.

Thus, system capacity of the radio interface 13 may be in- creased by exploiting or utilizing channel quality of the different UEs 12 using a scheduling algorithm. That means that a radio channel is preferably allocated to UEs 12 with good channel conditions that may receive a higher data rate or consume less radio interface resources for a given data rate.

A number of scheduling algorithms, like the proportional fair scheduler 47 or the maximum CIR scheduler 49, increase the system throughput by considering the radio channel conditions for the allocation of the radio channel to the different UEs 12. On the other hand, those algorithms achieve this gain at the expense of throughput variations between the different UEs 12 that have different radio channel conditions. That is, the UE 12 with good radio channel conditions gets a higher transfer data rate whilst the UE 12 with bad radio channel conditions gets a lower transfer data rate. The difference in UE throughputs is usually defined by the fairness of the ra- dio interface scheduler 34.

Thus, it is desirable that mobile communication networks en^¬ hance not only system throughput but also enhance fairness, in particular for cell edge UE . For example, the 3G LTE, as described in 3GPP TR 25.913, E-UTRA, aims to deliver signifi^¬ cantly improved spectrum efficiency and increased cell edge bit rate whilst maintaining the same site locations as de^¬ ployed today. Fig. 4 shows a high-level functional view 55 of an improved scheme of operating the base station 20 of the mobile commu^¬ nication network of Fig. 1.

The improved scheme optimizes or improves system behaviour by enabling the radio interface scheduler 34 to provide a fair or equal allocation of data rates to different UEs 12 when the transport network is the bottleneck for system capacity. Similarly, it optimizes or improves the radio interface ca^¬ pacity when the radio interface 13 is the bottleneck for the system capacity.

The improved scheme comprises a monitoring of network traffic over the transport network 21 as well as the network traffic over the radio interface 13 over all involved radio cells 28.

The monitoring determines packet loss 57, packet load 59, packet delay 61, and other parameters for the transport net^¬ work 21 as well as packet loss 57, packet load 59, packet de^¬ lay 61, and other parameters of radio cell load for the radio interface 13.

Based on the determined packet loss 57, the determined packet load 59, the determined packet delay 61, and other determined parameters, a transport load as well as a radio cell load is estimated for each QoS (Quality of Service) class, in steps 63 and 65. The uplink (UL) and downlink (DL) scheduling parameters is then adjusted using the estimated information, in a step 67. The adjustment is intended to achieve global optimum or over^¬ all improvement for system throughput and for fairness. The UL and DL scheduling parameters are later sent to the DL radio interface scheduler and the UL radio interface sched^¬ uler, in steps 69 and 71.

Then, based on this estimated information, the radio inter- face scheduler 34 adjusts the system throughput and the UE throughput with fairness between UE data rates to optimize or to improve throughout the data transmission chain.

A specific example of the improved scheme is provided below. In this example, only one QoS class for best effort traffic is considered.

Furthermore, the radio interface scheduler 34 modifies its behaviour according to a location of bottleneck for system capacity. The radio interface scheduler 34 behaves like the fair scheduler 47 when the transport network 21 is the bottleneck for system capacity and it behaves like the propor^¬ tional fair scheduler 51 when the radio interface 13 is the bottleneck for system capacity. A proposed or suggested op- eration range 75 of the improved scheme is shown in Fig. 3.

The radio interface scheduler 34 bases its decisions on a scheduling metric C for the different UEs 12 that it is serving .

For the fair throughput scheduler 47, the scheduling metric C of the UE 12 is defined by (1)

R,(t) where

i denotes an index of the UE 12 at a scheduling time in- terval t , and

R_j(t) denotes the averaged or filtered data rate of the UE i at the scheduling time interval t , R_i(t) = a-R_i(t-\) + (\-a)-R_i(t-\) , (2) where

R_t(t) denotes scheduled instantaneous data rate for the UE i in the scheduling time interval t , and a filter coefficient a defines a filtering or averaging time, wherein a has a value that lies between 0 and 1.

The UE i with a highest value of the scheduling metrics C. , which has the lowest average data rate R_t{t) is selected for data transmission. This leads to a balancing of UE throughputs that denotes a fair allocation of data rates.

On the other hand, the proportional fair scheduler 51 takes radio channel conditions of the considered UE 12 into account by estimating a peak throughput R,-(t) that could be allocated to the UE 12 in the considered scheduling time interval t . The scheduling criterion is defined as

where R_t(t) is defined by equation (2) . Considering equations (1) and (2), a general scheduling criterion may be defined by a weighted addition of both schedul^¬ ing criteria. Denoting a weighting factor by w , where a value of w lies between 0 and 1, the following generalized scheduling crite^¬ rion is defined

A tuning algorithm modifies the weighting factor w such that a global optimum or overall improvement for transport network throughput and fairness as well as radio interface throughput and fairness is achieved.

A simple algorithm that increases the weighting factor w by a certain factor Aw when a load in the transport network 21 exceeds a load in the radio cell 28 may do this. On the other hand, the weighting factor w is decreased by a factor Aw when the load in the radio cell 28 exceeds the load in the transport network 21. This is described by a following equa^¬ tion, i min(w_; ( - l) + A v,l), if transport network load > load in radio cell j, w (t) = \ '

1 max(w_y. (t - 1) - Aw, 0), if transport network load < load in radio cell j, where an index j denotes the considered radio cell 28.

In another words, a balance between a load of the transport network 21 and a load of the radio cell 28 may be achieved by adjusting a degree of fairness for the different UEs 12. In this manner, a maximum or good degree of fairness for a given system load is also obtained.

The load may be estimated, for example, from a ratio of used resources to unused resources. In the last mile transport 24, the load may be defined as number of bits per second trans- ferred over the last mile transport 24 divided by the trans^¬ mission rate of the last mile transport. For the radio inter^¬ face 13, the load definition is more complex and is defined based on its underlying radio standard. Metrics that may be used for the load estimation include downlink transmission power, uplink interference level, number of used spreading codes for a CDMA (Code Division Multiple Access) system, or number of used resource blocks for an OFDMA (Orthogonal Fre^¬ quency Division Multiple Access) system.

In a generic sense, this improved scheme may be adapted to other situations wherein other bottlenecks limit the system capacity. The other bottlenecks may include BTS hardware ca^¬ pability or software licenses for the radio cell or for the BS that limit the system throughput.

This improved scheme may also be used to take into account additional system bottlenecks, especially within the base station, to optimize or improve total system behaviour.

Fig. 5 depicts an embodiment of an implementation of the base station 20 of the mobile communication network 10 of Fig. 1.

The base station 20 has a computer system 80 and a data me- dium 82. The computer system 80 has a processor 84 and memory 86.

The data medium 82 is for holding a computer program 85 for executing a method to operate the base station 20.

The computer system 80 is able to read in the computer pro^¬ gram 85 of the data medium 82 into its memory 86 and executes or performs instructions of the computer program 85 using its processor 84. The computer system 80 uses the computer pro- gram 85 to perform functions or steps of the radio interface scheduler 34 and to perform functions or steps of the trans^¬ port scheduler 36 of Fig. 2. The memory 86 includes RAM (Random Access Memory) memory and ROM (Read Only Memory) memory. The RAM and ROM provide a storage area for the computer program 85. The computer program 85 performs steps for operating the base station 20 using software code that has instructions for per^¬ forming by the processor 84. The software code relates to a certain programming language. The processor 84 may be in the form of a RISC (Reduced In^¬ struction Set Computing) processor or a general-purpose processor for performing instructions of the computer program 85. The processor 84 may be implemented as hardware components using a certain hardware technology, such as MOS (Metal Oxide Semiconductor) , CMOS (Complementary Metal Oxide Semiconductor) , BiCMOS (Bipolar Complementary Metal Oxide Semiconduc^¬ tor) , ECL (Emitter Coupled Logic) , or TTL (Transistor- Transistor Logic) . The hardware components may comprise ASIC (Application Specific Integrated Circuit) components or DSP (Digital Signal Processing) components. The hardware compo^¬ nents may also include FPGA (Field Programmable Gate Array) . The method steps may also be implemented using software, hardware, or combination of software and hardware. The hard^¬ ware includes individual discrete components.

In a general sense, the base station 20 has one or more com^¬ puter systems 80. The processor 84 may be in the form of a digital signal processor for running the radio interface scheduler 34. The processor 84 may also be in the form of a network processor for running the transport scheduler 36. List of abbreviations

3GPP 3rd Generation Partnership Project

ATM Asynchronous Transfer Mode

BS Base Station

CDMA Code Division Multiple Access

CIR Carrier to Interference Ratio

DL Downlink

DSL Digital Subscriber Loop

EDGE Enhanced Data Rates for GSM Evolution

GSM Global System for Mobile communications

IP Internet Protocol

LTE Long Term Evolution

OFDMA Orthogonal Frequency Division Multiple Access

OPEX Operational Expenditures

PDH Plesiochronous Digital Hierarchy

QoS Quality of Service

RAN Radio Access Network

RLC Radio Link Control

RNC Radio Network Controller

TCP Transmission Control Protocol

UE User Equipment

UMTS Universal Mobile Telecommunications System

UL Uplink

WiMAX Worldwide Interoperability for Microwave Ac^¬ cess

WCDMA Wideband Code Division Multiple Access

Reference numbers

10 mobile communication network

11 Radio Access Network (RAN)

12 User-Equipments (UEs)

13 radio interface

15 core network

17 fixed line network

18 remote application

20 base station (BS)

21 transport network

22 other RAN elements

24 last mile access

26 switching and/or routing network

28 radio cell

29 bottleneck

30 bottleneck

32 network traffic control mechanisms

34 radio interface scheduler

35 radio interface capacity determination device

36 transport scheduler

37 transport network capacity determination device

38 Iub flow and congestion control

39 data packet traffic monitoring device

40 flow control

41 data packet traffic monitoring device

42 TCP flow and congestion control

45 graph

47 fair scheduler

49 maximum CIR (carrier to interference ratio) scheduler

51 proportional fair scheduler

55 high-level functional view

57 packet loss

59 packet load

61 packet delay

63 step 65 step

67 step

69 step

71 step

75 operation range

80 computer system

82 data medium

84 processor

85 program

86 memory

Claims

A base station (20) of a mobile communication system, said communication system comprises

a transport network (21) that is connectable to the base station (20) and

a plurality of UEs (User Equipments) (12) that are connectable to the base station (20) via a radio interface (13) ,

characterised in that

the base station (20) comprises

a first determining section (35) for determining system capacity of the radio interface (13),

a second determining section (37) for determining system capacity of the transport network (21),

a first changing section (36) for changing data rates of the plurality of the UEs (12) when the transport network (21) is causing a bottleneck of system capacity, and

a second changing section (34) for changing a radio interface capacity when the radio interface (13) is cau^¬ sing the bottleneck of system capacity.

A base station (20) of claim 1

characterised in that

the base station (20) uses 3GPP (Third Generation Part^¬ nership Project) protocol.

A base station (20) of claim 1 or 2

characterised in that

the base station (20) further comprises

a monitoring device (39) for monitoring packet traffic over the radio interface (13) to determine data packet characteristics of the radio interface packet traffic .

4. A base station (20) of one of claims 1 to 3

characterised in that the base station (20) further comprises

a monitoring device (41) for monitoring packet traffic over the transport network (21) to determine ta packet characteristics of the transport network pa cket traffic.

A method applied in a mobile communication network (10), wherein the mobile communication network (10) comprises a base station (20),

a transport network (21) that is connected to the base station (20) and

a plurality of UEs (User Equipments) (12) that are connected to the base station (20) via a radio in^¬ terface (13) ,

the method comprises

determining capacity of the radio interface (13), determining capacity of the transport network (21), a changing of data rates of the plurality of the UEs (12) if the transport network (21) is causing a bottleneck of system capacity of the mobile communication network (10) according to a first scheduling strategy (47), and

a changing of said radio interface capacity if the radio interface (13) is causing the bottleneck of system capacity of the mobile communication network (10) ac^¬ cording to a second scheduling strategy (49) .

A method of claim 5

characterised in that

the change of the data rates of the plurality of the UEs (12) according to a first scheduling strategy (47) increases a degree of fairness of the data rates of the plurality of the UEs (12) .

A method of claim 5 or 6

characterised in that the change of the radio interface capacity according to a second scheduling strategy (49) decreases a degree of fairness of the data rates of the plurality of UEs.

8. A method of claim 7

characterised in that

the decrease of fairness leads to an increase of the ra^¬ dio interface capacity which occurs for at least one ra^¬ dio cell (28) with a predetermined characteristic.

9. A method of claim 5

characterised in that

a degree of fairness is adjustable by the changing of the data rates of the plurality of the UEs (12) accord^¬ ing to the first and/or second scheduling strategy (47, 49) .

10. A method of claim 5

characterised in that

the changing of said data rates of the plurality of UEs

(12) according to the first scheduling strategy (47) wherein the first scheduling strategy (47) ensures a converging of data rates of the UEs (12) .

11. A method of claim 5 or 10

characterised in that

the changing of said radio interface capacity according to a second scheduling strategy (49)

wherein the second scheduling strategy (49) prioritizes the data rates of the UEs (12) according to the radio conditions of the UEs (12) .

12. A method of one of the claims 5 to 11

characterised in that

the determining system capacity of the radio interface

(13) further comprises monitoring packet traffic over the radio interface (13) to determine at least one data packet characteris^¬ tic of the radio interface packet traffic.

A method of one of claims 5 or 12

characterised in that

the determining system capacity of the transport network (21) further comprises

monitoring packet traffic over the transport net^¬ work (21) to determine at least one data packet charac^¬ teristic of the transport network packet traffic.

A method of claim 13

characterised in that

the method further comprises

determining the bottleneck of system capacity of the mobile communication network (10) using the at least one determined data packet characteristic of the radio in^¬ terface packet traffic and the at least one determined data packet characteristic of the transport network pa^¬ cket traffic.

15. An apparatus in a mobile communication network, said ap^¬ paratus being configured to perform scheduling in a communications network, comprising:

- means (35) configured to determine the capacity of the radio interface;

- means (37) configured to determine the capacity of the transport network;

- means (36) configured to change data rates of the plu^¬ rality of UEs if transport network is causing a bottle^¬ neck of system capacity; and

- means (34) configured to change a radio interface ca^¬ pacity if radio interface is causing a bottleneck of system capacity

16. A computer program (85) for executing a method according to any of claims 5 to 14.