WO2007111538A1

WO2007111538A1 - Controlling distribution of communication traffic load between multiple access networks employing different access technologies

Info

Publication number: WO2007111538A1
Application number: PCT/SE2006/000370
Authority: WO
Inventors: Per Magnusson; Mikael Prytz; Joachim Sachs; Teemu Rinta-Aho
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ)
Priority date: 2006-03-24
Filing date: 2006-03-24
Publication date: 2007-10-04
Also published as: TW200803354A

Abstract

The present invention relates to advanced mobile telecommunications in a multi network scenario with user terminals capable of handling a plurality of different access technologies. The object of the present invention is to enable control of traffic load between the different access networks that are based on different access technologies. The solution is a Gateway node (13) that produces measure on the traffic load and capacity in respective the network. The measure is produced from networks specific parameters and transformed into a common generic format, and provided to MRM node (12). The MRM (12) compares the generic measures from two or more networks and selects one of them for providing service to a particular user.

Description

CONTROLLING DISTRIBUTION OF COMMUNICATION TRAFFIC LOAD BETWEEN MULTIPLE ACCESS NETWORKS EMPLOYING DIFFERENT ACCESS TECHNOLOGIES

TECHNICAL FIELD OF THE INVENTION

The present invention relates to mobile telecommunications and more particular to methods and to nodes arranged to perform the methods and that enables distribution of traffic load between access networks that uses different types of access technologies. The invention is based on there are terminals that handle multiple access technologies.

DESCRIPTION OF RELATED ART

Today wireless communications can be achieved using a variety of different heterogeneous ANs (Access Network) . The ANs typically employs different Radio Access Technologies (RAT) . Therefore, the ANs differ in many respects, e.g., supported data rates, coverage, mobility, quality of service, business models, etc. Future wireless network systems are likely to include multi-access functionality to cooperatively (and efficiently) use many ANs employing different RATs ensuring that users are always best connected. With multi-access functionality operators can deploy ANs flexibly to match service and coverage requirements and still provide and manage a seamless and resource-usage efficient "single" network .

Current terminals already have multiple radio interfaces available, but in general there is limited support in networks for radio resource efficient multi-access, at least in between families of RATs . Certain RATs belonging to the same family, for example GSM and WCDMA, can be seen as more integrated from a radio resource efficiency perspective. Multi-access is typically enabled at higher layers without general mechanisms for reflecting the current radio resource state of the individual accesses in access selection decisions. This is partly due to the wide difference in how resources are used, shared, and managed, if at all, through Radio Resource Management (RRM) mechanisms in different ANs, as well as on other differences, for example supported services, quality, security, policies, and historical/industrial reasons.

Note, that this invention is described with respect to radio access and radio access technologies. The invention can he applied as well for only fixed access types, or a combination of fixed and wireless access. Radio resource management should then be generalized to access resource management, and multi-radio resource management to multi-access resource management.

SUMMARY OF THE INVENTION

The object of the present invention is to provide a solution for controlling the distribution of traffic in a multi-access network consisting of different access networks employing different access technologies.

The solution relates to a first method wherein Generic Resource Metrics are received from two or more access networks, and wherein said Generic Resource Metrics relates to a specific service for a certain user. By comparison of said Generic Resource Metrics, one of said access networks is selected for providing the specific service. The selected access network is then provoked to provide said specific service .

The solution also relates to a second method in interaction with the first method. The second method relates to identifying a Resource Management Area relevant for at least potentially providing the specific service, to retrieving access network specific parameters indicating the extent of occupation of resources in the Resource Management Area, to transforming said parameters into a Generic Resource Metric, and to report on said Generic Resource Metric .

The solution also involves a node adapted for performing the first method and to a node for performing the second method.

The advantage of the invention is better usage of the overall traffic capacity in the multi-access network. The benefit is primarily to operators of the multi-access network and to the total group of users that get a better chance of experiencing high quality sessions. Also a user that is transferred to another network is likely to experience higher quality; however, this is not the primary reason for the transfer.

DESCRIPTION OF THE DRAWINGS

Figure 1 is a block diagram of nodes in a multiple access network that are important for the present invention.

Figure Ib show three examples for relations between Resource Management Areas and service areas in different access networks .

Figure 2 is a flowchart of the steps of an inventive method as regarded from a Multi-access Resource Management function.

Figure 3 is a flowchart of the steps of an inventive method as regarded from a Generic Resource Translation Unit.

Figure 4 is staple diagram indicating the distribution of resources within a Resource Management Area.

Figure 5a, 5b, 5c & 5D discloses an example indicating the geographical presence of different users in different Resource Management Areas and indicating the Access Networks serving the respective users, at different time instances. .

DESCRIPTION OF PREFERRED EMBODIMENTS

The invention relates to a method by which a MRM (Multiaccess Resource Management) function in a multi-access network, controls admission of new users into multiple ANs (Access Networks) and controls distribution of users and services between the ANs. For this purpose the MRM decisions must be based on comparable, relevant measures for the current radio resource state, including load, and radio resource usage efficiency across all kinds of different, heterogeneous ANs. Due to that the multiple ANs are based on different types of access technologies, the capacities and capabilities cannot be immediately compared. Therefore the invention also involves an AN-specific translation entity called GRTU (Generic Resource Translation Unit) that, on- demand, self-triggered or by interval, computes values for a certain, precise set of Generic Resource Measures. The computation is based on information that is retrieved from AN-specific entities such as Radio Network Controllers' or Access Points / Access Controllers. The values are then communicated to an MRM control entity that makes multi-access decisions, such as admissions, load management, handovers etc .

Figure Ia is a block diagram of an example MAN (Multiple Access Network), 11, where the invention is to take place. The MAN 11 block diagram is simplified to involve only the nodes essential for the invention. They are the MRM-node, 12, with connections to at least two GRTUs 13. The GRTUs are connected to a respective AN, 14a, 14b.

Within the ANs, 14a, 14b, the resources are divided into RMAs

(Resource Management Areas) 15a, 15b, which are different for different ANs. Typically an RMA corresponds to a service area in an access network, for example a cell area, but it could also be a smaller or larger unit. Figure Ib shows three examples of relations between MRM, 15a, 15b, and service areas in different access networks. For one AN in the MAN the resources in each service area could be visible to MRM

(leftmost part of Figure Ib) . For another AN, MRM may only see the aggregated resources over multiple service areas

(middle part of Figure Ib) ; in this case the aggregate of service areas is seen as one RMA. It may also be that MRM can see the resources in individual service areas (rightmost part of Figure Ib) even if they are managed inside the AN by access-specific Resource Management (RM) , for example a Radio Network Controller.

Multi-access decisions are mainly interesting in the regions where RMAs of different ANs overlap. In a first example MRM, 12, is assumed to be aware of all RMAs for all ANs, 14a, 14b, in the multi-access system 11. MRM also needs to know which RMAs a certain user and specific service is connected to.

The background for the need of GRTU, 13, is in more detail the fact that different ANs have widely different mechanisms for using and sharing its available resources among users, for example division into time/frequency slots or codes or shared statistically using contention-based schemes. The notions of the amount of total resource occupied (the load level) , and the amount of resource that a particular user session occupies are therefore quite different for different ANs. The latter also depends on the quality requirements of the session/application. Different ANs may further be deployed with full or only spotty/hotspot coverage.

For MRM, 12, to exploit resource information from heterogeneous ANs in its operation, it is necessary that the GRTU, 13, derive comparable, relevant measures. The invention proposes a method with associated measures to solve this problem. The deployment of the GRTU depends on the AN; preferably it is placed where the (radio) link layer is controlled in the network. A single GRTU entity could therefore be responsible for one ore more RMAs . The GRTU computations are based on resource information retrieved from the AN-specific entities, for example Radio Network Controllers or Access Points / Access Controllers.

The method as performed in the MRM, 12, is disclosed in figure 2. Initially, S21, a request is sent from MRM, 12, to the two or more GRTUs, 13, for a Generic Resource Metric on the capacity occupied in the AN, 14a, 14b, with respect to a first service to a first user. Two different situations may apply; the AN, 14a, 14b, already supports the first service to the first user or alternatively the AN, 14a, 14b is a potential candidate for providing the first service. In next step, S12, the Generic Resource Metrics is received from the two or more GRTUs, 13, representing the capacity in the respective AN, 14a, 14b. The MRM selects one of the ANs, 14a, 14b based on the Generic Resource Metrics, in third step S13. Finally, the MRM, 12, provokes the first service to be provided by the selected one of the ANs, 14a, 14b.

The various steps of this basic method can be performed in different ways, which will be exampled further down in the description.

Figure 3 is the steps of the method performed in any of the

GRTUs, 14a, 14b, in interaction with the method disclosed in figure 2. In a first step, S31, the GRTU receives a request for a Generic Resource Metric with respect to a first service to a first user. In second step, S32, the GRTU, 13, identifies which Resource Management Area 15a, 15b, serves or would serve the first user. The identification is based on information in the received request that may specifically identify the Resource Management Area 15a, 15b, that may contain information on the geographical location of the first _η

user or may contain any other identification of the user that enables the GRTU to identify the relevant RMA, 15a, 15b via databases or measurements .

In the following step, S33, the GRTU, 13, retrieves parameters on the extent to which the resources in the RMA, 14a, 14b, are occupied. These parameters are transformed into a Generic Resource Metric. The transformation is further described into more detail and variations further down in the description.

Finally, see S35, the Generic Resource Metric is reported to the MRM, 12.

The method disclosed in figure 2 and 3 is initiated by the MRM, 12, for example when it becomes aware that the first user requests the first service to be set up. Then an AN, 14a, 14b, need be assigned for the first time. It may further be initiated by the MRM, 12, because the resources in an RMA are used too heavily. In any of these cases the AN, 14a, 14b, may be alarmed via its interface to the GRTUs, 13, or possibly another interface to the ANs 14a, 14b.

Alternatively, the methods of figure 3 and 4 are initiated by the GRTUs that reports of the Generic Resource Metric without being requested by the MRM 12. The GRTU, 13, may report Generic Resource Metric on a regular basis, or when ANs- specific events occur. For example an AN, 14a, 14b may actively try to recruit the first user if it identifies in could be served by RMA with low traffic load. Also an AN, 14a, 14b, serving the first user may actively try to get rid of it, if it is served by a high loaded RMA, 15a, 15b. In the alternative, the first step, S21, S31, of the methods is superfluous .

According to a first embodiment the MRM, 12, know, which RMA, 15a, 15b, the first user is connected to. The first user terminal regularly sends performance reports on link quality _g

to one of the ANs, 14a, 14b. The report informs of the RMA, 15a, 15b, in which the first user is present. This is made both in the situation of the AN 14a, 14b, is supporting a service if the first user terminal is in idle mode. The information on in which RMA the first user is present is forwarded to the MRM, 12, via the GRTU, 13.

When the MRM, 12, requests, S21, Generic Resource Measures it informs the requested GRTUs, 13, of the relevant RMA, 15a, 15b. Identifying, S32, in the GRTU, 13, then only includes detecting the RMA, 15a, 15b information in the request.

In the situation of the preferred embodiment, when the MRM, 12, knows the RMA, 15a, 15b, of the users, the selection of an AN, 14a, 14b, to provide service is also a selection of RMA, 15a, 15b.

The request transmitted by the MRM, 12, in step S21 and received in step S31 by the GRTU, 13, in more detail relates to, admission and load management for a user/session i requiring quality Qi in an area covered by at least one RMA. Qi can be defined in different ways, but typically it contains requirements on minimum bit rate and maximum delay as well as how much additional quality, typically extra bit rate, if possible to provide, is beneficial. A typical example of a service that can benefit from additional quality is file transfers. For each relevant and specified RMA, MRM, 12, transmits Qi to the associated GRTU, 13, and requests, S21, it to do a AN-specific mapping of Qi and compute the Generic Resource Metric. The GRTU retrieves, S33, AN, 14a, 14b, specific parameters on the usage of the RMA, 14a, 14b, resources. Based on these parameters the Generic Resource Metric is computed.

The AN specific resource parameters are transformed into the Generic Resource Metrics according to the following: Initially the AN, 14a, 14b, specific parameters on resources in each RMA are quantified into a relative resource level, see figure 3. A resource level of one corresponds to all available resources in the RMA and a level of zero corresponds to none of the available resources. Depending on what is the limiting resource (s) in the AN, 14a, 14b, it can be computed as, for example, the relative number of time slots or codes, the relative amount of downlink power, the relative occupied bandwidth, the average collision ratio, or combinations thereof such as power and slot or chunk. Three different resource levels are identified:

rmin = the current minimum required amount of resources for all active users/sessions in the RMA

rocc = the current occupied amount of resources for all active users/sessions in the RMA. This can be larger than rmin whenever extra, "elastic" resources are provided for users/sessions that can benefit from it. It is assumed that all of the extra resources (rocc - rmin) can (either instantaneously or after some delay) be reclaimed and used for other users.

rmax = the current maximum amount that can be used in the RMA. The "headroom" 1-rmax is the margin required to cope with changing resource usage of active users in the RMA due to, for example, user mobility. If the time to free up extra, "elastic" resources is very small or zero, rocc resources

(but not rmin) could be allowed to grow beyond the rmax limit as these can be freed up to give room for increasing resource usage of active users .

Based on the resource levels rmin, rocc, and rmax, further Generic Resource Metrics may be computed according to the following:

δmin = rmax - rmin = relative amount of currently available resources, including resources that can be (instantaneously or after some delay) freed up if they are currently used to provide extra quality for some (or all) users.

δocc = rmax - rocc = relative amount of currently free resources .

qi,min = relative AN-specific (instantaneous) resource usage if user/session i would be "served" in the RMA with minimum quality requirements. For example, if there are 10 slots in the RMA and one slot is required then qi,min =0,1. If it is not possible to meet the minimum quality requirements in the RMA then qi,min can be set to qi,min = 1+ε, for example qi , min = 1,1.

qi, extra = as above, but where user/session i is given as much extra quality as it benefits from or can be currently given. So qi, extra >= qi,min as it contains additional spare resources.

σi,min = qi,min / δmin = relative resource usage efficiency/impact of serving user/session i in the RMA with minimum quality requirements. For example, if δmin = 0.4 and qi,min =0,1 then σi,min = 0,25. Note that if σi,min > 1 then the amount of resources requested exceeds the available resources and such request is typically rejected.

σi, extra = qi, extra / δocc = as above, but where user/session i is given as much extra quality it benefits from or can be currently given.

Once computed the Generic Resource Metrics are signaled from the GRTU to MRM, which then uses the values for its multiaccess control decision. (Note, that absolute values can also be calculated from relative resource measures based on Qi, which is given in absolute terms.)

Selecting, S23, AN, 14a, 14b, based on the Generic Resurce

Metric in more detail: _{l χ}

In admission control a first step is to check whether σi,min <= 1 for the RMA; if not then the user/session i cannot be admitted there.

Initial access selection or load management can be based on many different algorithms. Examples can be:

Choosing the AN / RMA with maximum amount of available resources δmin:

RMA selected for user/session i = argminj { δmin (RMAj) }

Choosing the AN / RMA with maximum amount of free resources δocc :

RMA selected for user/session i = argminj { δocc (RMAj) }

Choosing the AN / RMA with minimum resource usage efficiency σi,min, that is:

RMA selected for us.er/session i = argminj { σi,min (RMAj) }

Choosing the AN / RMA with minimum resource usage efficiency σi, extra, that is:

RMA selected for user/session i = argminj { σi, extra (RMAj) }

Note that if the current load is high in all RMAs, i.e. σi, extra > 1 for all ANs, then alternatively the AN / RMA could be selected minimizing σi,min instead.

Many other MRM decision algorithms can be considered based on the MRM resource measures above. The key purpose of the measures is to provide sufficient, comparable radio information on current radio resource state and resource usage efficiency for various heterogeneous ANs for effective MRM operation.

Conversion between absolute and relative measures:

All measures apart from Qi in the above description are given as relative measures, as shown in Fig. 3. The conversion between absolute and relative measures is done in GRTU. The

(absolute) service request Qi is received at MRM from a higher level. MRM forwards Qi to one or more GRTUs, which convert it into access-specific relative resource requirements qi . Further the other relative measures are calculated and reported back to MRM.

Additional interactive negotiation between GRTU (s) and MRM can be performed, when GRTUs provide additional information about the best service performance that can be provided

(without guarantees) . For this a translation from relative measures back to an absolute measure Q is required. This is explained with the following example:

MRM gets a request for (absolute) resources Qreq,i,. E.g. Qreq, i is a request for 150 kb/s service.

MRM passes the request Qreq, i on to GRTUs, where it is translated to relative resources qmin,i.

GRTUs reply to MRM relative load values for load balancing, to determine the relative resource costs of Qreq, i (i.e. qmin, i ) .

If spare access resources are available, it would maybe be beneficial to know what absolute service level Qoffered, i could be provided to the service by the access (i.e. what maximum Q can be provided by qi, extra > qi,min) . This _{χ 3}

requires that GRTU would not only make a translation Qreq<=>qmin but also qextra<=>Qoffered.

Then MRM would get the information from GRTU, "your request Qreq can be handled, it costs the relative resources qmin. But the access could even support the service request at level Qoffered, but this would then cost the resources qextra .

Further parameters for the selection (S32) of AN, 14a, 14b/KMΑ, 15a, 15b;

Also other parameters than the Generic Resource Metric may be weighted in the selection, S23, of AN 14a, 14b/ RMA 15a, 15b. These parameters may for example be:

• Operator or terminal priorities of RAT/RMA usage

• Some ANs / RMAs may be provided by other cooperating operators. In this case additional roaming/cooperation charges may exist for the usage of those ANs / RMAs.

• Some ANs / RMAs may have less efficient operation, for example, they require more signaling for handover or AAA signaling.

• Some ANs / RMAs may provide less security.

Example on distribution of traffic generated by a plurality of users:

The following, idealistic, example shows how the MRM resource measures can be used in admission control and load management. Consider the deployment in figure 5a where the different RMAs show coverage areas for two different ANs, RMAl uses a first AN, 14a, while RMA2 and RMA3 (hot spot areas) use a second AN, 14b. Initially there are no users. Assume further for simplicity that there is no distinction between minimum required quality and extra quality and that no resource margins are used. Hence, initially δ = 1

(subscript dropped) for all RMAs. The small circles in figure 5a represent users and the numbers are the sequence in which they first appear. During the first five time events MRM is faced with an admission control / access selection decision

(ANl in RMA 1 or AN 2 in RMA 2/3), but in the sixth time event only ANl in RMAl can be selected. Finally, assume that qi = 0,1 for all i in RMA2 and RMA3 while qi = 0,11 for all i in RMAl, and that these are fixed throughout.

The following table shows the access selection decisions during the first five time events :

Figure 5b shows the access selection state after the fifth time event according to result of the table.

Figure 5c shows the state after the sixth time event when three users have appeared and been selected in RMAl . Now RMAl is becoming fairly loaded (δ = 0,45) . In a load management process MRM may reevaluate the relative resource usage efficiencies σ for the users that are in overlapping RMA regions and connected to RMAl, that is, to see the effect of changing the access selection. For user 2, σ = 0,196 in the present RMAl (virtually take away and "add back") and σ = 0,111 in RMA2 , so the user is handed over to RMA2. Next, for user 5, σ^'= 0,164 in the present RMAl (after user 2 was handed over) and σ = 0,125 in RMA3, so this user is handed over to RMA3. The net result after this "load management" operation is shown in Figure 5d.

Example: User allocation with different radio capacity.

In the following we assume that we have an overlay of two ANs with exactly the same cell structure (RMAl equals RMA2) but different cell capacities. For example, the two ANs could use the same RATs in close frequency bands but with different frequency allocations, e.g. ANl with a 20 MHz carrier and AN2 with a 5 MHz carrier. As a consequence, for the same Qi different values of qi are used. The following values are assumed:

As a start no users are in the system and then users are successively added. The following table shows the allocation of users to the system, assuming that the allocation policy is to always select the RMA with the lowest value of σ.

Example: User allocation with different radio efficiency.

In the following we assume that we have an overlay of two ANs with exactly the same cell structure (RMAl equals RMA2) but different cell capacities. We assume that the difference in capacity is caused by the RATs having different spectral efficiency. As a consequence, for the same carrier bandwidth a different rmax is achieved, and the required relative resources qi for a service request Qi is different. The following values are assumed:

Related work:

There is ongoing work within the IEEE 802.21 working group on Media Independent Handover Services [1] that have recognized the need for an abstraction model of load, but to the best of our knowledge there is no proposal for a solution, or any method for comparing different types of resources or for considering resource occupancies for certain users and services or resources usage efficiencies.

Between GSM & WCDMA there is a possibility (in release 5) to exchange load information [2] [3] [4] using generic measures in the following common measurement IE:

1) Cell Capacity - a value between 1 and 100 with a step size of 1. This value is set by OAM by the operator and is used to differentiate between different cells on a relative scale. The value is linear.

2) Cell Load (as defined on Iur in rel4, but extended to be a value between 0 and 100 with a step size of 1) . The measure should be clarified to only include "planned" capacity.

3) Estimated percentage of current load that is conversational or streaming traffic. A value between 0 and 100 with a step size of 1.

4) NRT load - having possible values: Low (meaning no problem to add another NRT user) , Medium (meaning normal operation) , High (meaning a situation where another NRT user would likely cause an overload situation) and Overload (meaning a situation where there is an overload of NRT users in the cell) . NRT = Interactive or background.

There are also 3GPP reports [5] [6] that describe Common Radio Resource Management (CRRM) for GSM and WCDMA. These compare the absolute amount of free capacity (e.g. in Kbps) in cells [5] but also the generic resource measures above [6].

The setup is still access network specific for GSM and WCDMA. Furthermore, there are no measures for resource occupancy or resource usage efficiency for a certain user and specific service.

2

REFERENCES

[1] IEEE 802.21, Media Independent Handover, http: //www.ieee802.org/21/ .

[2] 3GPP TS 48.008, "Mobile Switching Centre - Base Station System (MSC-BSS) interface; Layer 3 specification".

[3] 3GPP TS 25.423, "UTRAN Iur interface KNSAP signalling".

[4] 3GPP TS 25.413, "UTRAN Iu interface RANAP signalling".

[5] 3GPP TR 25.881, "Improvement of RRM across RNS and RNS/BSS" . [6] 3GPP TS 25.891, "Improvement of RRM across RNS and RNS/BSS (Post Rel-5); (Release 6)".

Claims

1. A method for controlling traffic load distribution between two or more access networks that uses different access technologies and provide service in geographical service areas that are at least partly overlapping, and wherein the respective service areas are divided into Resource Management Areas (15a, 15b) each assigned a specific set of resources for- handling communication, comprising the steps of:

- receiving (S22) the information from said access networks on their respective ability to provide a first service to a first user and a respective Generic Resource Metric indicating the occupation of said specific set of resources relevant for the first service;

- selecting (S23) one of said access networks based on comparison of their reported Generic Resource Metric; and,

- provoking (S24) the first service to be provided by the selected access network.

2. The method of claim 1 wherein the receiving step is preceded by the further step of:

-requesting (S21)at least the one or more of said access networks that is not supporting the first service to the first user, to provide said Generic Resource Metric.

3. The method of claim 1 or 2 wherein a further condition for said selection is the ability of said access network to provide the first service to the first user.

4. The method of claim 1 or 3 wherein said ability to provide the first service to the first user includes the admission of the first user into the access network.

5. The method of claim 1 or claim 2 wherein said received Generic Resource Metric is specific for the Resource Management Area .

6. The method of claim 2 or claim 5 wherein the in the requesting step (S21) the access networks are given information for identifying the Resource Management Area relevant for the first user.

7. A method (in a GRTU) related to management of traffic load between different access networks, comprising the steps of;

-identifying (S32) a first Resource Management Area (15) within an first access network that is supporting, or alternatively, would support, a first service to a first user;

- retrieving (S33) access network specific parameters indicating to what extent communication resources in the Resource Management Area is occupied in relation to the first service;

- transforming (S34) said parameters into a Generic Resource Metric; and

-reporting (S35) said Generic Resource Metric.

8. A method according to claim 7 wherein the identifying step is preceded by the further step of:

- receiving (S31) a request to provide said Generic Resource Metrics for the first service to the first user.

9. The method of claim 8 wherein the request further includes a query on the ability of the first access network to provide the first service to the first user.

10. The method of claim 7 or 9 wherein said reporting step also includes information on the ability of the first network to provide the first service to the first user.

11. The method of claim 10 wherein said ability includes the admission of the first user into the first access network.

12. The method of claim 10 wherein said ability includes the capacity to provide the first service to the first user.

13. The method of claim 8 wherein said request further includes information for identifying the Resource Management Area relevant for the first user.

14. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative amount of currently available resources, including resources that can be (instantaneously or after some delay) freed up if they are currently used to provide extra quality for some (or all) users.

15. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative amount of currently free resources .

16. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative AN-specific (instantaneous) resource usage if the said first service for the said first user is served with minimum quality requirements.

17. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative AN-specific (instantaneous) resource usage if the said first service for the said first user is served with as much extra quality as it benefits from or can be currently given.

18. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative resource usage efficiency if the said first service, for the said first user is served with minimum quality requirements.

19. The method of claim 1 or 7 wherein the Generic Resource Metric indicates the relative resource usage efficiency if he said first service for the said first user is served with as much extra quality as it benefits from or can be currently given.

20. A multi access management node comprising an interface for connection to at least to access networks and arranged for performing the method of claim 1, or any method claimed depending on 1.

21. A generic resource translation unit with a first interface for connection to a multi access management node and with an second interface for connection to a first access networks, and arranged for performing the method of claim 7 or on any of the claims dependent on claim 7 , and wherein said Generic Resource Metric is reported over the first interface and wherein said first access network parameters is communicated over the second interface.