WO2019007486A1

WO2019007486A1 - Method and apparatus for supporting instantiation of network slice in communications network

Info

Publication number: WO2019007486A1
Application number: PCT/EP2017/066661
Authority: WO
Inventors: Patrick Maguire
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ)
Priority date: 2017-07-04
Filing date: 2017-07-04
Publication date: 2019-01-10

Abstract

A method for supporting instantiation of a network slice in a communications network in response to a service request. The communications network comprises a Radio Access Network, RAN, partitioned into a plurality of RAN resource partitions. The method performed by a domain manager for the RAN comprises: receiving (202) from a network manager information indicative of required capacity of the network slice specified in the service request, instantiating and/or reconfiguring (204) Virtual Network Functions required for fulfilling the service request. The method also comprises selecting (206) a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice, and binding (210) the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition (208). The domain manager is also disclosed.

Description

METHOD AND APPARATUS FOR SUPPORTING INSTANTIATION OF NETWORK SLICE IN COMMUNICATIONS NETWORK

Technical Field The present invention relates to method and apparatus for a communications network in general and, in particular, to a method and apparatus for supporting instantiation of a network slice in a communications network.

Background

As the communications networks evolve the number of services provided grows. From a simple voice communication at the very beginning to voice and data communication, including texting, email, web browsing, video transmission and others in 4G wireless networks. When looking at the wide range of applications and services to be provided by a 5G network, it is quite obvious that these cannot effectively be addressed with a traditional approach of having a purpose-built network for each application. This would lead to high cost for networks and devices as well as inefficient use of valuable frequency resources. An operator may have one physical network infrastructure and one pool of frequency bands, which may support many separate virtualized networks, also called network slices. Each network slice may have unique characteristics for meeting the specific requirements of the use case/s it serves. A key function of 5G Core network is to allow for flexibility in network service creation, making use of different network functions suitable for the offered service in a specific network slice, e.g. Evolved Mobile Broadband (MBB), Massive Machine Type Communication (MTC), Critical MTC, Enterprise, etc.

Network slicing is about creating logically separated partitions of the network, addressing different business purposes. These network slices are logically separated to a degree that they can be regarded and managed as networks of their own. Network slices can be associated with different slice tenants, namely entities that provide services to the end user via the network slice. Examples of slice tenants could be: armed forces providing, for example, voice communication via a dedicated network slice;

- vehicles manufacturers providing, for example, vehicle diagnostic services via a dedicated network slice.

It is also common to associate a Service Level Agreement (SLA) to each network slice. An SLA indicates the treatment to which services provided within the network slice should be subjected.

Network slicing is a new concept that applies to both LTE Evolution and new 5G RAT. The key driver for introducing network slicing is business expansion, i.e. improving the cellular operator's ability to serve other industries, e.g., by offering connectivity services with different network characteristics (performance, security, robustness, and complexity).

The current working assumption is that there will be one shared Radio Access Network (RAN) infrastructure that will connect to several Core Network (CN) instances, such as Evolved Packet Core (EPC) instances (one EPC instance per network slice) or a further EPC evolution. Although it is assumed that there will be a one to one mapping between slice and CN Instance, where an instance is a CN node, concepts, features, aspects may be equally applicable to multiple slices supported by the same CN instance. As the CN functions (e.g. EPC functions) are being virtualized, it is assumed that the operator may instantiate a new Core Network when a new slice should be supported. In another case the network slices could be implemented based on existing monolithic EPC architecture based on special purpose hardware. This architecture is shown in Figure la where Slice 0 can for example be a Mobile Broadband slice and Slice 1 can for example be a Machine Type Communication network slice. When it comes to one physical RAN used for multiple network slices, two aspects should be taken into account. The first aspect is that network slicing should make possible to support several different virtual networks on the same physical network infrastructure in order to reduce costs and energy consumption compared to deploying separate physical networks for the different use cases or business scenarios. In order to fully exploit this benefit, it is required that the slicing concept allows for efficient usage of common resources such as radio resources and infrastructure, and transport links between the slices such as fronthaul and backhaul. The second aspect is that mechanisms are required to protect common channels or resources used for UEs (user equipment) accessing system so that congestion in one slice does not have a negative impact on another slice (often called slice isolation).

To address slice isolation, it is necessary to implement resource isolation in RAN, and to achieve this resource partitioning is being proposed. Resource partitioning in RAN entails dividing up the full radio resource allocation into maximum of six partitions (the number of RAN partitions may be extended to 32 in the future). The operator will be given parameters to control whether the RAN resource partition is based on Public Land Mobile Network (PLMN) or Subscriber Profile Identity (SPID), and then control how much resources each RAN resource partition shall be given.

Figure lb summarises resource partitioning. Under low load bearers from one partition can use resources from another partition. At high load, when partitions are competing for same resources, the bearers of a partition are bound by the configured partition share size. In other words, partitioning is ringfencing of resources at high load, which means that a tenant of a RAN resource partition has guaranteed X% of resources as specified in a contract. These X% will be provided when the tenant needs them, but when the tenant needs less than X% of resources the unused portion may be used by other partitions. However, if the tenant need its X% in full then the ringfenced resources will be pre-empted so that the tenant can get the guaranteed X%. This is illustrated in Figure lb by the thick lines. When partition B needs resources and partition A does not use them then partition B may use them, but no more than is left unused. In case both partitions would need more resources than their allocated share size they are limited to what has been allocated to then (this is shown by the thick line running along the straight solid line). With the adoption of network slicing, it will be possible that two (or more) different slices contain all the services, which exist today, but there will still be the need to differentiate access as two (or more) slices represent different SLAs. In fact, scenarios may be foreseen where the same type of service is provided by different slices associated with SLAs with different slice tenants. This would lead to more slices and to a further need for slice awareness information at the RAN.

Today, operator's Network Management System (NMS) see subnetworks consisting of a management node (domain manager or subnetwork manager depending on naming convention) and the nodes (managed elements) it manages. The interface between network manager, 902, (or network management system, NMS) and underlying domain management systems, 904, 906, is called Interface-N (Itf-N) and is illustrated in Figure 9.

Summary

According to a first aspect of the present invention there is provided a method for supporting instantiation of a network slice in a communications network in response to a service request. Said communications network comprises a Radio Access Network, RAN, partitioned into a plurality of RAN resource partitions. The method performed by a domain manager for the Radio Access Network comprises: receiving from a network manager information indicative of required capacity of the network slice specified in the service request and instantiating and/or reconfiguring Virtual Network Functions required for fulfilling the service request. The method also comprises selecting a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice; and binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

According to a second aspect of the present invention there is provided a domain manager for a Radio Access Network, RAN, for supporting instantiation of a network slice in a communications network in response to a service request. Said communications network comprises said Radio Access Network partitioned into a plurality of RAN resource partitions. The domain manager comprises a processor and a memory. The memory contains instructions executable by the processor such that the domain manager is operative to receive from a network manager information indicative of required capacity of the network slice specified in the service request and to instantiate and/or reconfigure Virtual Network Functions required for fulfilling the service request. The domain manager is also operative to select a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice; and to bind the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

According to a third aspect of the present invention there is provided a domain manager for a Radio Access Network, RAN, for supporting instantiation of a network slice in a communications network in response to a service request. Said communications network comprises said Radio Access Network partitioned into a plurality of RAN resource partitions, wherein said domain manager comprises: a receiver for receiving from a network manager information indicative of required capacity of the network slice specified in the service request and a module for instantiating and/or reconfiguring Virtual Network Functions required for fulfilling the service request. The domain manager also comprises a selector for selecting a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice and a binder for binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition. Further features of the present invention are as claimed in the dependent claims.

Brief description of the drawings

The present invention will be understood and appreciated more fully from the following detailed description taken in conjunction with the drawings in which:

Fig. la is a diagram illustrating an architecture of a wireless communications network operating two network slices;

Fig. lb is a chart illustrating how resources are used by resource partitions depending on network load;

Fig. 2 - Fig. 6 are flowcharts illustrating embodiments of a method for supporting instantiation of a network slice in a communications network; Fig. 7 is a diagram illustrating a domain manager in one embodiment of the present invention;

Fig. 8 is a diagram illustrating a domain manager in an alternative embodiment of the present invention;

FIG. 9 is a diagram illustrating relation between a network manager and domain managers.

Detailed description

In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, interfaces, techniques, etc. in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary details.

Reference throughout the specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of the present invention. Thus, the appearance of the phrases "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily all referring to the same embodiment. Further, the particular features, structures or characteristics may be combined in any suitable manner in one or more embodiments.

In order to meet the evolving requirements on network management in a 5G context, there is now a need to be able to deploy and manage logical networks for a particular business purpose. Such logical networks formed of Managed Elements (ME) Group (network slices) create the need to be able to manage the associated group of network elements as a collective entity.

A network slice is composed of a subnetwork instance or multiple subnetwork instances (e.g. Radio Access Network, transport network, packet core, etc.). A subnetwork instance (SNI) is managed by a domain manager whereas a network slice (which is an end-to-end logical entity) is managed by a network manager. It must be noted, however, that management operations performed on a network slice by a network manager or management operations performed on a subnetwork instance by a domain manager affect operation of the network of which the network slice and the subnetwork instance are parts. The solution in embodiments disclosed in this document proposes a method to align network slice management with RAN resource partitioning thus ensuring optimum use of radio resources in a partitioned configuration.

As the number of network slice instances will exceed the number of RAN resource partitions, network slices will be multiplexed on to the available RAN resource partitions. From RAN resource usage perspective, it is optimal that network slices with same/similar requirements/characteristics will be multiplexed onto the same RAN resource partition, i.e. the configuration profile applied to the RAN resource partition is optimized for a particular radio behaviour, thus being able to serve many network slices with same/similar requirements/characteristics. This would also be beneficial for end user experience.

The problem with resource partitioning in this context is that there is no relationship between the configuration of the partition share size and the number of network slices and their individual associated capacity requirement which may vary over time. Without a runtime relationship between network slice life cycle management and configuration of RAN resource partitions, it is likely that over time the share size configuration of resource partitions will be far from optimal.

With reference to Figure 2 an embodiment of a method for supporting instantiation of a network slice in a communications network, is now to be described. The communications network comprises a Radio Access Network, RAN, partitioned into a plurality of RAN resource partitions. The network slice instantiation starts when a network manager receives a service request specifying that the network slice needs to be instantiated. A Network Slice Life Cycle Manager residing in the network manager starts the instantiation of the requested network slice. In a preferred embodiment, the method for supporting instantiation of a network slice is performed by a domain manager for the Radio Access Network. As illustrated in Figure 2 in a preferred embodiment the method comprises receiving, 202, from the network manager information indicative of required capacity of the network slice specified in the service request. Figure 9 illustrates example of communication interface between, the domain manager for the Radio Access Network, 904, receiving, 202, from the network manager, 902, information indicative of required capacity of the network slice specified in the service request. As further illustrated in Figure 2, in a preferred embodiment the method comprises instantiating and/or reconfiguring, 204, Virtual Network Functions (VNFs) required for fulfilling the service request. In some embodiments, in order to fulfil the service request it may be necessary instantiate new Virtual Network Functions that have not been provided by the RAN, or it may be necessary to perform scaling of existing VNFs. Sometimes one of these operations will be enough and sometimes both will need to be carried out - this depends on the characteristics of the service request and configuration of the RAN network. In the following step the method comprises selecting, 206, a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice and binding, 210, the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice. The binding operation is performed if a sum of the required capacity (i.e. capacity required for operating the network slice) and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition, 208. This claim preferred embodiment assumes that there is enough capacity in the selected partition to support the service request (i.e. instantiation of the network slice).

In the operation of binding logical associations between elements to be bound are created. A network manager as well as radio access network nodes (e.g. eNodeBs) are notified of the binding so that the network management can take this into account and traffic steering can be effected as intended for the network slice. In the example given above the logical associations are created linking the selected RAN resource partition and the instantiated Virtual Network Functions with the network slice. A consequence of binding, once the slice is instantiated, is that traffic related to end user services associated with this network slice would only use the selected partition. In a preferred embodiment the network slice life cycle manager, after the network manager received the service request, sends a trigger to the domain manager for the Radio Access Network to instantiate /reconfigure the required resources (i.e. subnetwork life cycle management is being triggered). In short, network slice life cycle management commences when network slice life cycle manager receives the service request from customer via the network manager. Network slice life cycle manager subsequently send a domain level service request to the RAN domain manager (i.e. domain manager for the Radio Access Network) to trigger RAN subnetwork life cycle management, i.e. instantiation of virtual RAN (vRAN) resources and, in the end, partition-to-slice binding. In the event that the sum of the required capacity and a current capacity supported by the selected RAN resource partition is exceeds the capacity limit of the selected RAN resource partition the RAN resource partitions need to be recalibrated, 208, 212.

Recalibration, 212, in its various embodiments is illustrated in Figures 3 - 5. The goal of the recalibration is to reallocate capacity from the other RAN resource partitions which have not been selected in step 206 to the selected RAN resource partition. If recalibration succeeds the selected RAN resource partition will have enough capacity to support the network slice requested in the service request and the binding 210 of the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice will proceed. In case there is not enough spare capacity in the other RAN resource partitions that could be reallocated to the selected RAN resource partition the recalibration fails and in consequence the network slice instantiation fails and an alarm is raised for the operator to act upon.

Now, the various embodiments of the recalibration operation will be described in more detail.

With reference to Figure 3, after it has been determined that the selected RAN resource partition does not have enough of its own capacity to support the network slice requested in the service request, step 208 in Figure 2, the method determines, 302, additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request. This additional capacity is the capacity shortfall of the selected RAN resource partition. In the following step share size of at least one remaining RAN resource partition is reducing, 304, to produce spare capacity for allocation to the selected RAN resource partition. The share size of the at least one remaining RAN resource partition is reduced to such an extent that it leaves enough capacity to support requirements of said at least one remaining RAN resource partition. Depending on circumstances and approach taken it may be that there is enough spare capacity in just one RAN resource partition and by reducing its share size it is possible to obtain capacity that after allocation to the selected RAN resource partition will be enough to support the service request.

Alternatively, the spare capacity can be obtained by adding up capacity freed up after reducing share size of more than one RAN resource partition. The method then comprises allocating, 308, at least part of the spare capacity to the selected RAN resource partition if the spare capacity is greater or equal to the additional capacity, 306. Once the selected RAN resource partition has enough capacity to support the service request (to operate the network slice) the method comprises binding, 310, the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice.

As mentioned earlier, it may be that there is not enough spare capacity in the other RAN resource partitions that could be reallocated to the selected RAN resource partition and then, 306-No, the network slice instantiation fails, 312, and an alarm is raised, 314.

Another alternative embodiment of the recalibration operations, 212, is illustrated in Figure 4. In this embodiment the recalibration starts with the same operation, 302, of determining additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request. The method then moves on to reducing share size of the remaining RAN resource partitions and it starts, 402, with the partition with the largest share size of the remaining RAN resource partition and is carried out in descending order based on share size, 402 - 408, of the remaining RAN resource partitions. Preferably the spare capacity is monitored in each iteration and once there is enough spare capacity, i.e. spare capacity is equal or above the additional capacity required fulfil the service request, the iteration stops, 306 - Yes, and the method then allocates, 308, at least part of the spare capacity to the selected RAN resource partition. Once the selected RAN resource partition has enough capacity to support the service request (to operate the network slice) the method comprises binding, 310, the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice. In this embodiment the operation of reducing starts with the largest partition because this where potentially are the largest reserves. In a preferred embodiment a magnitude of reduction of share size of an individual remaining RAN resource partition is proportional to the share size of said individual remaining RAN resource partition.

Alternatively, it may be that after the last iteration, when the last available RAN resource partition had its share size reduced the spare capacity is less than the additional capacity, 406 - yes. In this situation network slice instantiation fails, 312, and an alarm is raised, 314.

An alternative embodiment of the recalibration operation is illustrated in Figure 5. In this embodiment, the step of reducing share size of at least one remaining RAN resource partition is carried out in descending order based on unused capacity of the remaining RAN resource partitions, 502 - 506, and starts with a RAN resource partition having a largest unused capacity, 502. The advantage of this approach is that it reduces the number of iterations needed before enough spare capacity is collected because we're taking first the biggest chunks available. As in the previous embodiment, also in this one, a magnitude of reduction of share size of an individual remaining RAN resource partition is proportional to the share size of said individual remaining RAN resource partition. Also in the embodiment depicted in Figure 5 it may be that after the last iteration, when the last available RAN resource partition had its share size reduced the spare capacity is less than the additional capacity, 506 - yes. In this situation network slice instantiation fails, 312, and an alarm is raised, 314. In one embodiment, if the spare capacity is greater than the additional capacity required to fulfil the service request then capacity not smaller than the additional capacity is allocated from the spare capacity to the selected RAN resource partition and what then remains from the spare capacity is returned to the at least one remaining RAN resource partition. Preferably, if the spare capacity is greater than the additional capacity required to fulfil the service request then only capacity equal to the additional capacity is allocated from the spare capacity to the selected RAN resource partition and the remaining capacity from the spare capacity is returned to the at least one remaining RAN resource partition. In this embodiment the method takes only as little as needed for the selected partition to fulfil the service request and returns the rest of the spare capacity to the remaining RAN resource partitions. Preferably, what remains from the spare capacity after allocation of part of the spare capacity to the selected RAN resource partition, 308, is redistributed to more than one of the remaining RAN resource partitions proportionally to the remaining RAN resource partitions share size. In a preferred embodiment, illustrated in Figure 6, the operation of reducing share size of a RAN resource partition, which is executed when it is known that there is enough spare capacity in the remaining RAN resource partitions. In this embodiment the step of reducing comprises reserving, 602, a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition and executing reduction, 604, of share size of the at least one RAN resource partition if the spare capacity is greater than or equal to the additional capacity, 306. In this embodiment a number of iterations may be carried out as in the previously described embodiments until either enough spare capacity is reserved or until the last RAN resource partition is approach without securing enough spare capacity. The difference is that the actual reduction is carried out only when enough capacity has been reserved. The advantage is of this embodiment is that there is no need to return any unused capacity to the remaining RAN resource partitions because the actual reduction is executed only if there is enough spare capacity for the recalibration to be successful and takes only what is necessary (i.e. the additional capacity) and the rest of the reserved capacity is released and no longer reserved.

Alternatively, if the iterations failed to reserve enough spare capacity, i.e. if the spare capacity is smaller than the additional capacity required to fulfil the service request, 306, after reaching the last one of the remaining RAN resource partitions, 406, 506, the method comprises releasing, 606, the capacity by removing the reservation.

In a preferred embodiment the operation of reducing is carried out only on those of the remaining RAN resource partitions which are eligible for share size reduction, because there may be scenarios in which configuration does not allow some RAN resource partitions to undergo recalibration.

The benefit provided by recalibration, apart for providing resources for a network slice to be instantiated, is that it provides a runtime relationship between network slice life cycle management and configuration of RAN resource partitions which helps maintaining share size of resource partitions optimal over time.

Figure 7 illustrates a domain manager, 700, which implements the method for managing a wireless communications network operating at least one network slice described earlier. In a preferred embodiment, the domain manager, 700, is implemented as a physical apparatus operating in the wireless communications network. The domain manager, 700, is for a Radio Access Network, RAN, and supports instantiation of a network slice in a communications network in response to a service request. The communications network comprises the Radio Access Network partitioned into a plurality of RAN resource partitions. In a preferred embodiment the domain manager, 700, comprises a processor, 702, and a memory 704. The memory, 704, contains instructions executable by the processor, 702, such that the domain manager, 700, is operative to receive from a network manager information indicative of required capacity of the network slice specified in the service request. The domain manager, 700, is also operative to instantiate and/or reconfigure Virtual Network Functions required for fulfilling the service request. In some embodiments, to fulfil the service request, it may be necessary for the domain manager to instantiate new Virtual Network Functions that have not been provided by the RAN, or it may be necessary to perform scaling of existing VNFs. Sometimes one of these operations will be enough and sometimes both will need to be carried out - this depends on the characteristics of the service request and configuration of the RAN network. The domain manager, 700, is further operative to select a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice and bind the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

In other words, when there is enough capacity in the selected partition to support the service request then logical associations are created linking the selected RAN resource partition and the instantiated Virtual Network Functions with the network slice and in consequence the instantiation of the network slice may go ahead.

It may happen, however, that the sum of the required capacity and a current capacity supported by the selected RAN resource partition is above the capacity limit of the selected RAN resource partition. This is a situation in when there is not enough capacity in the selected partition to support the service request. To address this problem the domain manager, 700, is operative to determine additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request. The additional capacity is the deficit of capacity at the selected RAN resource partition. In addressing this problem the domain manager, 700, is further operative to reduce share size of at least one remaining RAN resource partition and in this way to produce spare capacity for allocation to the selected RAN resource partition. It is important, however, that the reduction of share size of the at least one remaining RAN resource partition leaves enough capacity to support requirements of said at least one remaining RAN resource partition. This operation has been described earlier and is called recalibration. If the recalibration produces spare capacity that is greater than or equal to the additional capacity the domain manager, 700, is operable to allocate at least part of the spare capacity to the selected RAN resource partition and then bind the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice.

Two alternative embodiments of the domain manager, 700, operative to carry out the recalibration will not be described. In one embodiment, the domain manager carries out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on share size of the remaining RAN resource partitions starting with a largest of the remaining partitions. Alternatively, the domain manager, 700, carries out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on unused capacity of the remaining RAN resource partitions starting with a RAN resource partition having a largest unused capacity. These operations are illustrated in Figures 4 and 5 accordingly.

Preferably, the domain manager, when carrying recalibration, stops the recalibration once it collected enough spare capacity - this is illustrated by the loop 306 - 406 and 306 - 506 in Figures 4 and 5. It may be, however, that the domain manager goes with reducing capacity of each one of the RAN resource partitions eligible for reduction and only then allocates at least part of the spare capacity to the selected RAN resource partition. If the spare capacity is greater than the additional capacity required to fulfil the service request the domain manager, 700, allocates from the spare capacity to the selected RAN resource partition capacity not smaller than the additional capacity. The remaining capacity from the spare capacity is then returned to the at least one remaining RAN resource partition. Preferably, as described earlier, the remaining capacity is redistributed to more than one of the remaining RAN resource partitions proportionally to the remaining RAN resource partitions' share size.

In a preferred embodiment if the spare capacity is greater than the additional capacity required to fulfil the service request the domain manager, 700, allocates from the spare capacity to the selected RAN resource partition only capacity equal to the additional capacity. The remaining capacity from the spare capacity is then returned to the at least one remaining RAN resource partition and, preferably, is redistributed proportionally as described earlier.

It may also happen that after reducing share size of all remaining RAN resource partitions, the spare capacity is still smaller than the additional capacity required to fulfil the service request. In this situation, the recalibration is not successful and the domain manager, 700, is operative to raise an alarm indicating that instantiation of the network slice failed. The spare capacity is returned in the same way as described in earlier embodiments. Preferably, the domain manager, 700, is operative to reduce share size of a RAN resource partition, when it is known that there is enough spare capacity in the remaining RAN resource partitions that in the end the recalibration will be successful and it will be possible to go ahead with the binding. In this embodiment the domain manager, 700, is operative to reserve a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition and execute reduction of share size of the at least one RAN resource partition if the spare capacity is greater than or equal to the additional capacity. The domain manager, 700, is further operative to release the reservation if the spare capacity is smaller than the additional capacity required to fulfil the service request. Details of how the domain manager, 700, operates in this embodiment are illustrated in Figure 6 and described earlier. The advantage is of this embodiment is that domain manager does not need to return any unused capacity to the remaining RAN resource partitions because the actual reduction is executed only if there is enough spare capacity for the recalibration to be successful and takes only what is necessary (i.e. the additional capacity) and the rest of the reserved capacity is released and no longer reserved.

In the embodiment illustrated in Figure 7 the domain manager, 700, also comprises an interface, 706, for communicating with other elements of the network. The interface, 706, processor, 702 and memory, 704, in the embodiment illustrated in Figure 7 are connected in series, however, other architectures of the apparatus, 700, are also possible and would be envisaged by those skilled in the art. For example, the elements 702 - 706 could be connected to a bus. Figure 8 illustrates another embodiment of a domain manager, 800, which implements the method for managing a wireless communications network operating at least one network slice described earlier. In a preferred embodiment, the domain manager, 800, is implemented as a physical apparatus operating in the wireless communications network. The domain manager, 800, is for a Radio Access Network, RAN, and supports instantiation of a network slice in a communications network in response to a service request. The communications network comprises said Radio Access Network partitioned into a plurality of RAN resource partitions. In a preferred embodiment the domain manager, 800, comprises a receiver, 802, for receiving from a network manager information indicative of required capacity of the network slice specified in the service request. The domain manager also comprises a module, 804, for instantiating and/or reconfiguring Virtual Network Functions required for fulfilling the service request and a selector, 806, for selecting a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice. Further, the domain manager comprises a binder, 808, for binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

In a preferred embodiment, the domain manager also comprises a recalibration module, 812. If the sum of the required capacity and a current capacity supported by the selected RAN resource partition is above the capacity limit of the selected RAN resource partition it is necessary to perform recalibration in order to fulfil the service request. The recalibration module, 812, operative to determine additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request. The recalibration module, 812, is also operative to reduce share size of at least one remaining RAN resource partition to produce spare capacity for allocation to the selected RAN resource partition, wherein said reduction of share size of the at least one remaining RAN resource partition leaves enough capacity to support requirements of said at least one remaining RAN resource partition. Moreover, the recalibration module is operative to allocate at least part of the spare capacity to the selected RAN resource partition if the spare capacity is greater or equal to the additional capacity. In a preferred embodiment the recalibration module, 812, is operative to carry out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on share size of the remaining RAN resource partitions starting with a largest of the remaining partitions and in an alternative embodiment based on unused capacity of the remaining RAN resource partitions starting with a RAN resource partition having a largest unused capacity.

Preferably, the recalibration module, 812, is operative to raise an alarm indicating that instantiation of the network slice failed if, after reducing share size of all remaining RAN resource partitions, the spare capacity is smaller than the additional capacity required to fulfil the service request.

Preferably, if the spare capacity is greater than the additional capacity required to fulfil the service request the recalibration module, 812, is operative to allocate from the spare capacity to the selected RAN resource partition capacity not smaller than the additional capacity and to return to the at least one remaining RAN resource partition the remaining capacity from the spare capacity. The remaining capacity from the spare capacity is then returned to the at least one remaining RAN resource partition. Preferably, as described earlier, the remaining capacity is redistributed to more than one of the remaining RAN resource partitions proportionally to the remaining RAN resource partitions' share size.

Alternatively, if the spare capacity is greater than the additional capacity required to fulfil the service request the recalibration module, 812, is operative to allocate from the spare capacity to the selected RAN resource partition only capacity equal to the additional capacity. What capacity then remains from the spare capacity is returned to the at least one remaining RAN resource partition. Possible details of returning the remaining capacity are already discussed above.

In a preferred embodiment of the domain manager, 800, the recalibration module, 812, is operative to reduce share size of a RAN resource partition, only when it is known that there is enough spare capacity in the remaining RAN resource partitions that in the end the recalibration will be successful and it will be possible to go ahead with the binding. In this embodiment the recalibration module, 812, is operative to reserve a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition and execute reduction of share size of the at least one RAN resource partition if the spare capacity is greater than or equal to the additional capacity. If the spare capacity is smaller than the additional capacity required to fulfil the service request the recalibration module (812) is operative to release the reservation. In the embodiment illustrated in Figure 8 the domain manager, 800, also comprises an interface, 810, for communicating with other elements of the network. The interface, 810, receiver, 802, instantiating module, 804, selector, 806, binder, 808, and recalibration module, 812, in the embodiment illustrated in Figure 8 are connected to a bus, however, other architectures of the apparatus, 800, are also possible and would be envisaged by those skilled in the art. For example, the elements 802 - 812 could be connected in series.

Claims

1. A method for supporting instantiation of a network slice in a communications network in response to a service request, whereas said communications network comprises a Radio Access Network, RAN, partitioned into a plurality of RAN resource partitions, and the method performed by a domain manager for the Radio Access Network comprises:

receiving from a network manager information indicative of required capacity of the network slice specified in the service request;

instantiating and/or reconfiguring Virtual Network Functions required for fulfilling the service request;

selecting a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice; and

binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

2. The method according to claim 1, wherein if the sum of the required capacity and a current capacity supported by the selected RAN resource partition is above the capacity limit of the selected RAN resource partition the method comprises:

determining additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request;

reducing share size of at least one remaining RAN resource partition to produce spare capacity for allocation to the selected RAN resource partition, wherein said reduction of share size of the at least one remaining RAN resource partition leaves enough capacity to support requirements of said at least one remaining RAN resource partition, allocating at least part of the spare capacity to the selected RAN resource partition if the spare capacity is greater or equal to the additional capacity; and

binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice.

3. The method according to claim 2, wherein the step of reducing is carried out in descending order based on share size of the remaining RAN resource partitions and starts with a largest of the remaining partitions.

4. The method according to claim 2, wherein the step of reducing is carried out in descending order based on unused capacity of the remaining RAN resource partitions and starts with a RAN resource partition having a largest unused capacity.

5. The method according to any one of claims 2 - 4, comprising raising an alarm indicating that instantiation of the network slice failed if, after reducing share size of all remaining RAN resource partitions, the spare capacity is smaller than the additional capacity required to fulfil the service request.

6. The method according to any one of claims 2 - 4, wherein if the spare capacity is greater than the additional capacity required to fulfil the service request then capacity not smaller than the additional capacity is allocated from the spare capacity to the selected RAN resource partition and the remaining capacity from the spare capacity is returned to the at least one remaining RAN resource partition.

7. The method according to claim 6, wherein if the remaining capacity from the spare capacity is redistributed to more than one of the remaining RAN resource partitions the redistribution is in proportion to the remaining RAN resource partitions share size.

8. The method according to any one of claims 2 - 7, wherein the operation of reducing is carried out on those of the remaining RAN resource partitions which are eligible for share size reduction.

9. The method according to any one of claims 2 - 8, wherein the step of reducing comprises:

reserving a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition;

executing reduction of share size of the at least one RAN resource partition if the spare capacity is greater than or equal to the additional capacity.

10. The method according to claim 9, when dependent on claim 5, further comprising releasing the reservation if the spare capacity is smaller than the additional capacity required to fulfil the service request.

11. A domain manager for a Radio Access Network, RAN, for supporting instantiation of a network slice in a communications network in response to a service request, whereas said communications network comprises said Radio Access Network partitioned into a plurality of RAN resource partitions; said domain manager comprising a processor and a memory, the memory containing instructions executable by the processor such that the domain manager is operative to:

receive from a network manager information indicative of required capacity of the network slice specified in the service request;

instantiate and/or reconfigure Virtual Network Functions required for fulfilling the service request;

select a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice; and

bind the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

12. The domain manager according to claim 11, wherein if the sum of the required capacity and a current capacity supported by the selected RAN resource partition is above the capacity limit of the selected RAN resource partition the domain manager is operative to:

determine additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request;

- reduce share size of at least one remaining RAN resource partition to produce spare capacity for allocation to the selected RAN resource partition, wherein said reduction of share size of the at least one remaining RAN resource partition leaves enough capacity to support requirements of said at least one remaining RAN resource partition,

- allocate at least part of the spare capacity to the selected RAN resource partition if the spare capacity is greater or equal to the additional capacity; and

bind the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice.

13. The domain manager according to claim 12, operative to carry out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on share size of the remaining RAN resource partitions starting with a largest of the remaining partitions.

14. The domain manager according to claim 12, operative to carry out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on unused capacity of the remaining RAN resource partitions starting with a RAN resource partition having a largest unused capacity.

15. The domain manager according to any one of claims 12 - 14, operative to raise an alarm indicating that instantiation of the network slice failed if, after reducing share size of all remaining RAN resource partitions, the spare capacity is smaller than the additional capacity required to fulfil the service request.

16. The domain manager according to any one of claims 12 - 14, wherein if the spare capacity is greater than the additional capacity required to fulfil the service request the domain manager is operative to allocate from the spare capacity to the selected RAN resource partition capacity not smaller than the additional capacity and to return to the at least one remaining RAN resource partition the remaining capacity from the spare capacity.

17. The domain manager according to claim 16, wherein if the remaining capacity from the spare capacity is to be redistributed to more than one of the remaining RAN resource partitions the domain manager is operative to redistribute the remaining capacity in proportion to the remaining RAN resource partitions share size.

18. The domain manager according to any one of claims 12 - 17, operative to carry out the reduction on those of the remaining RAN resource partitions which are eligible for share size reduction.

19. The domain manager according to any one of claims 12 - 18, wherein when carrying out reducing the share size of the at least one remaining RAN resource partition the domain manager is operative to:

- reserve a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition;

execute reduction of share size of the at least one RAN resource partition if the spare capacity is greater than or equal to the additional capacity.

20. The domain manager according to claim 19, when dependent on claim 15, further operative to release the reservation if the spare capacity is smaller than the additional capacity required to fulfil the service request.

21. A domain manager for a Radio Access Network, RAN, for supporting instantiation of a network slice in a communications network in response to a service request, whereas said communications network comprises said Radio Access Network partitioned into a plurality of RAN resource partitions, wherein said domain manager comprises:

- a receiver for receiving from a network manager information indicative of required capacity of the network slice specified in the service request;

a module for instantiating and/or reconfiguring Virtual Network Functions required for fulfilling the service request;

a selector for selecting a RAN resource partition from said plurality of RAN resource partitions for binding with the network slice; and

a binder for binding the selected RAN resource partition and the instantiated Virtual Network Functions to the network slice if a sum of the required capacity and a current capacity supported by the selected RAN resource partition is equal to or below a capacity limit of the selected RAN resource partition.

22. The domain manager according to claim 21, further comprising a recalibration module, wherein if the sum of the required capacity and a current capacity supported by the selected RAN resource partition is above the capacity limit of the selected RAN resource partition the recalibration module is for:

- determining additional capacity above the capacity limit of the selected RAN resource partition required to fulfil the service request;

reducing share size of at least one remaining RAN resource partition to produce spare capacity for allocation to the selected RAN resource partition, wherein said reduction of share size of the at least one remaining RAN resource partition leaves enough capacity to support requirements of said at least one remaining RAN resource partition, and

allocating at least part of the spare capacity to the selected RAN resource partition if the spare capacity is greater or equal to the additional capacity.

23. The domain manager according to claim 22, wherein the recalibration module is operative to carry out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on share size of the remaining RAN resource partitions starting with a largest of the remaining partitions.

24. The domain manager according to claim 22, wherein the recalibration module is operative to carry out the reduction of the share size of the at least one remaining RAN resource partition in a descending order based on unused capacity of the remaining RAN resource partitions starting with a RAN resource partition having a largest unused capacity.

25. The domain manager according to any one of claims 22 - 24, wherein the recalibration module is operative to raise an alarm indicating that instantiation of the network slice failed if, after reducing share size of all remaining RAN resource partitions, the spare capacity is smaller than the additional capacity required to fulfil the service request.

26. The domain manager according to any one of claims 22 - 24, wherein if the spare capacity is greater than the additional capacity required to fulfil the service request the recalibration module is operative to allocate from the spare capacity to the selected RAN resource partition capacity not smaller than the additional capacity and to return to the at least one remaining RAN resource partition the remaining capacity from the spare capacity.

27. The domain manager according to claim 26, wherein if the remaining capacity from the spare capacity is to be redistributed to more than one of the remaining RAN resource partitions the recalibration module is operative to redistribute the remaining capacity in proportion to the remaining RAN resource partitions share size.

28. The domain manager according to any one of claims 22 - 27, wherein when carrying out reducing the share size of the at least one remaining RAN resource partition the recalibration module is operative to:

reserve a fraction of capacity of the at least one remaining RAN resource partition for allocation to the selected RAN resource partition;

29. The domain manager according to claim 28, when dependent on claim 25, wherein the recalibration module is further operative to release the reservation if the spare capacity is smaller than the additional capacity required to fulfil the service request.