Slice Assurance in Mobile Networks
By introducing slice assurance function (SAF) in mobile networks, monitoring and adjusting slice resources, the SLA satisfaction problem of network slices in mobile networks under time-varying packet traffic is solved, and dynamic optimization and efficient utilization of slice resources are achieved, ensuring that Stability and reliability of network performance.
Patent Information
- Application Number
- CN202080081799.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-11-27
- Filing Date
- 2020-11-27
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-11-27
AI Technical Summary
In mobile networks, prior art is difficult to effectively ensure that network slices meet service level agreements (SLAs) in the face of time-varying packet traffic, especially dynamically adjusting resources across regions and time to meet RAN specificity in tracking areas of multiple cells. There are challenges in SLA.
Introducing slice assurance functionality (SAF) ensures that each slice instance meets SLA in the mobile network by monitoring key performance metrics (KPIs) and reconfiguring slice resources if necessary. SAF redistributes resources across multiple cells, including dynamic adjustments across regions and time, and uses a new interface to send reconfiguration commands to the base station to achieve optimization of slice resources.
It effectively ensures the SLA satisfaction under time-varying packet traffic, improves the reliability and reliability of network performance, and realizes the efficient utilization of slice resources through dynamic resource adjustment, meeting the bandwidth, latency and reliability requirements of different services.
Smart Images

Figure CN114747249B_ABST
Abstract
Description
[0001] This application claims the benefit of U.S. Provisional Patent Application No. 62 / 940,887, filed on November 27, 2019, the entire contents of which are incorporated herein by reference. Technical Field
[0002] The present disclosure relates to mobile networks, and more particularly, to slicing implemented within a mobile network. Background Art
[0003] One of the main technical challenges facing mobile operators today is the ability to provide the various network performance characteristics required by future applications and services. To name a few, bandwidth, latency, packet loss, security and reliability will vary significantly between different services. Emerging applications such as robotic teleoperation, massive IoT and autonomous vehicles require connectivity but with very different characteristics. The combination of architectural flexibility, software programmability, the needs of different verticals (medical, factory, military, public safety, etc.) and the various types of applications has led to the concept of network slicing in 5G mobile networks. Network slicing provides a convenient way to completely segment the network to support specific types of services or businesses. In addition, each slice is optimized based on capacity, coverage, connectivity, security and performance characteristics (input, latency). Since slices can be isolated from each other as if they were physically separated in the control plane and user plane, the user experience of network slices will be the same as on separate networks.
[0004] Network slices within the mobile operator domain span software applications running on network nodes, core network components, transport networks, and radio access network (RAN) components. 3GPP standards build a sliceable 5G infrastructure to provide many logical network segments on a common single physical network (see 3GPP's TR28.801 document entitled "Telecommunication management; Study on management and orchestration of network slicing for next generation network"). Technologies such as software-defined networking (SDN) and network function virtualization (NFV) with separation of the control plane (CP) and user plane (UP) are key factors in segmenting traditional network structures. Customizable and virtualized network components can be connected together using only software to provide the appropriate level of connectivity.
[0005] 5G standardization work has begun to define specific slices and their service level agreements (SLAs) based on application / service types. For example, a user equipment (UE) can now directly specify its desired slice using a new field in the control message called Network Slice Selection Assistance Information (NSSAI). A subfield of the NSSAI is the Slice / Service Type (SST) used to indicate the slice type. These standards have defined the most commonly used network slice types and reserved the corresponding standardized SST values (see 3GPPTS23.501). For example, SST values 1, 2, and 3 correspond to slice types for enhanced mobile broadband (eMBB), ultra-reliable low latency communications (uRLLC), and massive IoT (MIoT), respectively. The access and mobility management function (AMF) of the core network retrieves the slices allowed by the user subscription and interacts with the network slice selection function (NSSF) of the core network to select the appropriate network slice instance for the traffic. In addition, 3GPP specifies new slice management functions, such as Network Slice Management Function (NSMF) and Network Slice Subnet Management Function (NSSMF), whose only role is to create, manage and monitor slice instances within the mobile operator's network for users / applications. The slice catalog and all slice-specific service level agreements (SLAs) are stored in the NSMF. The specific SLA corresponding to the slice segment / subnet is stored in the NSSMF.
[0006] SLAs are divided into contractual (static) SLAs and network (dynamic) SLAs. Static SLAs only define the legal and financial terms and conditions between the slice user and the operator, applicable to the activation, operation, penalties and termination of the slice. These SLAs do not change as network conditions change and therefore remain unchanged. On the other hand, network SLAs mainly define the Quality of Service (QoS) requirements. These roughly include slice availability / reliability (the percentage of service uptime to total time), slice throughput (bits per second), packet latency (average and maximum packet latency in milliseconds), packet loss (the percentage of packets lost to total packets in a defined time interval), capacity (bits per second per km2), etc. These network slice SLAs are well documented (see 3GPP TS22.261). There are key performance indicators (KPIs) collected from in-band or out-of-band network segments to measure the fulfillment of these SLAs over time. Network monitoring is the task of the Service Management and Orchestration (SMO) system and is clearly defined in 3GPP standard documents. Summary of the invention
[0007] In general, techniques for slice assurance within a mobile network are described. During the lifecycle of a network slice, from start to finish, its service level agreement (SLA) must be met in the mobile network, especially on the Radio Access Network (RAN) side over a Tracking Area (TA), where each area is formed by multiple cells that must be able to meet the SLA under time-varying packet traffic. In some examples, a new control network function called Slice Assurance Function (SAF) and its interface to a base station (e.g., gNodeB) are defined for near real-time (near-RT) and non-real-time (non-RT) coordination of cell slice resource allocation across those multiple cells, with the goal of meeting RAN-specific SLAs for slices within a region.
[0008] For example, the SAF may first determine the target RAN specific slice key performance indicators (KPIs) that must be monitored for each slice instance within the TA and their corresponding thresholds, and the SAF is able to activate the relevant performance management (PM) jobs within each gNodeB it controls. The SAF monitors these KPIs and checks them against the thresholds, and based on these checks, the SAF can repeatedly fine-tune the slice resource allocation of the gNodeB. When the mobile traffic density or characteristics within the TA changes over time, the SAF calculates the new distribution of available resource parameters for any or all of the 1st, 2nd and 3rd RAN protocol layers and sends reconfiguration commands to the gNodeB using a new interface. In this way, the SAF can facilitate the practical application of slice resource reallocation across regions and across time. The SAF can retrieve the slice attributes and the SLA for each slice from the Network Slice Subnet Management Function (NSSMF), send slice-specific notifications to the responsible party when / if the resource reconfiguration cannot meet the SLA, and send PM jobs for KPI monitoring to the gNodeB.
[0009] In some examples, a method includes: obtaining, by a slice assurance function (SAF) executed by a device, a key performance indicator (KPI) value of a first slice among multiple slices implemented by multiple base stations serving a tracking area of a mobile network; determining, by the SAF based in part on the KPI value of the first slice, that a service level agreement (SLA) of the first slice is not met; in response to the determination, reallocating, by the SAF, slice resources associated with any one of the multiple slices to calculate new slice configuration parameters for the first slice; and reconfiguring, by the SAF, at least one of the multiple base stations to implement the new slice configuration parameters for the first slice.
[0010] In some examples, a slice assurance function (SAF) for a mobile network includes: a slice performance collector, including a processing circuit and configured to obtain a key performance indicator (KPI) value of a first slice among multiple slices, the multiple slices being implemented by multiple base stations serving a tracking area of the mobile network; a slice optimizer subsystem, including a processing circuit and configured to determine that a service level agreement (SLA) of the first slice is not met based in part on the KPI value of the first slice, wherein the slice optimizer subsystem is further configured to reallocate slice resources associated with any one of the multiple slices in response to the determination to calculate new slice configuration parameters for the first slice; and a slice control action subsystem, including a processing circuit and configured to reconfigure at least one of the multiple base stations to implement the new slice configuration parameters for the first slice.
[0011] In some examples, a mobile network includes: multiple base stations, including corresponding slice assurance function clients; and a slice assurance function, including processing circuitry, and configured to: obtain a key performance indicator (KPI) value of a first slice among multiple slices, the multiple slices being implemented by multiple base stations serving a tracking area of the mobile network; determine, based in part on the KPI value of the first slice, that a service level agreement (SLA) for the first slice is not met; in response to the determination, reallocate slice resources associated with any one of the multiple slices to calculate new slice configuration parameters for the first slice; and transmit the new slice configuration parameters for the first slice to a slice assurance function client via an interface so that the slice assurance function client reconfigures the corresponding base station to implement the new slice configuration parameters for the first slice.
[0012] The details of one or more examples of the disclosure are set forth in the accompanying drawings and the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 is a block diagram illustrating an exemplary logical 5G RAN architecture.
[0014] Figure 2 is a block diagram illustrating a distributed gNodeB architecture for a 5G mobile network.
[0015] Figure 3A is a block diagram illustrating a slice resource usage reconfiguration across two slices in accordance with the techniques of this disclosure.
[0016] Figure 3B is a block diagram illustrating another slice resource usage reconfiguration across two slices in accordance with the techniques of this disclosure.
[0017] Figure 3Cis a block diagram illustrating a slice resource usage reconfiguration across two gNodeBs of the same slice in accordance with the techniques of the present disclosure.
[0018] Figure 3D is a block diagram illustrating another slice resource usage reconfiguration across two gNodeBs of the same slice in accordance with the techniques of the present disclosure.
[0019] Figure 4A To illustrate the SAF, SAF client and E2 for near RT control according to the technology of the present disclosure ++ A block diagram of an exemplary system of interfaces.
[0020] Figure 4B is a diagram showing a SAF for non-RT control, a SAF client, and O1 according to the technology of the present disclosure ++ A block diagram of an exemplary system of interfaces.
[0021] Figure 5 is a block diagram of a SAF according to the disclosed technology.
[0022] Figure 6 A flow chart of an exemplary method for initiating KPI measurement for a configuration according to the disclosed techniques.
[0023] Figure 7 is a flow chart illustrating an exemplary method of reconfiguring RAN resources using SAF according to the techniques of this disclosure.
[0024] Like reference numerals refer to like elements throughout the specification and drawings. DETAILED DESCRIPTION
[0025] In general, a mobile operator’s SLA can be defined on a per-slice basis and includes the path from the UE to the mobile network exit point. This exit point can be a destination site or a gateway function to the public Internet. However, even within a mobile network, a slice contains many segments. This SLA typically has different parts, for example,
[0026] a) RAN segment (including both fronthaul and midhaul components)
[0027] b) Transmission phase (also known as return)
[0028] c) Core network segment
[0029] Each of these segments contributes to the SLA in a different way. For example, latency is determined by adding up the packet delays within all said segments. Availability and packet loss are determined similarly, i.e., by adding up the contributions of all segments. Conversely, throughput and capacity must be the same for all segments. From the slice SLA, the NSMF must derive the RAN-specific slice SLA portion, which will be controlled by the system according to the techniques described in this article. The RAN-specific SLA can be stored in the NSMF or NSSMF and accessible from it.
[0030] The 5G radio access network has a distributed architecture. The base station, also known as gNodeB, is divided into three functional components that can be deployed in various configurations: Central Unit (CU), Distributed Unit (DU), and Radio Unit (RU). The CU performs upper layer protocol processing. The DU performs low layer protocol processing. Using the newly defined open F1 interface, different DU and CU vendors can now interoperate. The CU is also divided into control plane and user plane functions, CU-CP and CU-UP respectively. While the CU-CP controls multiple CU-UPs using the E1 interface; the CU-CP also controls all its subtended DUs using the F1-C interface. The RU, which manages the RF layer, has antenna arrays of various sizes and shapes. Depending on the operator and service requirements, the gNodeB can be deployed as a whole, that is, the RU, DU, and CU reside within the cell site, or these functions can be distributed across sites, for example, the RU and DU can reside in the cell site, while the CU resides in the edge cloud site that controls multiple of these distributed DUs. Since the DU-CU interface is more tolerant to latency, they can be separated by a certain distance, but if the interface is more sensitive to latency, the RU and DU are more likely to be collocated. The trade-offs between implementations depend on the traffic requirements and economics of these components.For simplicity, in this document, the terms gNodeB or base station are used interchangeably and do not distinguish how its distributed components are deployed within the cell site.
[0031] The 5G standard defines a registration area (RA) that includes multiple tracking areas. The concept of tracking areas comes from 4G networks, while registration areas (RAs) are newly defined in 5G. Each tracking area (TA) includes multiple cells. Each cell is identified by a cell ID. A user equipment (UE) is considered to be within a cell, TA, and RA at any point in time. The UE is also allowed to support up to 8 different network slices that the cell / TA / RA infrastructure must support.
[0032] First, the UE registers with the AMF, which provides the UE with a TAId (TAI) and thus the registration area in which it is located. Slices are described with a service level agreement (SLA) that applies to the entire tracking area. However, a tracking area includes multiple cells, and the UE radio access network conditions and density distribution on these different cells in the area vary over time. There is no mechanism defined in the standard on how to ensure that RAN-specific slice SLAs are met across cells as traffic changes. In addition, network KPIs are defined on an end-to-end basis, including RAN and core network segments. In some examples, these techniques include deriving RAN-specific slice KPI counterparts to monitor and verify these RAN-specific SLAs.
[0033] According to some examples of the present disclosure, an exemplary slice assurance function (SAF) and exemplary methods are developed so that RAN-specific SLAs are met over time and over an area (e.g., a tracking area) when there are a large number of cells in the area. The SAF collects KPIs from all gNodeBs directly or via the service management and coordination (SMO) system of the 5G network, compares them with thresholds related to the implementation of each SLA component, and determines whether all cells of each TA meet all SLAs over time. If not, the SAF uses an algorithm to determine the reallocation ratio of radio access network resources across cells by calculating the distribution of slice resources for each cell across multiple slices. In this way, the SAF implements dynamic radio resource allocation (DRRA) over the cells in the tracking area.
[0034] According to some examples of the present disclosure, the algorithm described herein for determining reallocation can (i) reallocate slice resources that are not used by other slices to those slices that need additional slice capacity to meet the SLA on the same gNodeB, (ii) reallocate slice resources that are being used by lower priority slices to slices with higher priority, even if both slices need additional capacity to meet the SLA, and / or (iii) reallocate slice resources across multiple gNodeBs by shifting slice traffic to other gNodeBs (or components) when additional capacity is needed to meet the SLA. Depending on the choice of implementation, all of these actions can be performed in near real time, i.e., less than 1 second, or in non-real time, i.e., more than 1 second. Different examples can perform any combination of (i), (ii), and (iii).
[0035] According to some examples of the present disclosure, a Slice Assurance Function (SAF) interfaces with all gNodeBs (or subcomponents thereof) to send reconfiguration commands. These commands are received by a "SAF client," which is a software subsystem implemented in each gNodeB to receive the reconfiguration commands and, accordingly, to be able to activate configuration changes in the gNodeB.
[0036] According to some examples of the present disclosure, the SAF is implemented as a subcomponent of the RAN Intelligent Controller (RIC) well known in the prior art (also known as the RAN Edge Cloud Controller). In doing so, the SAF utilizes the interfaces with the SMO and NSMF / NSSMF already defined and implemented by the RIC to collect KPIs, RAN-specific SLAs for the tracking area, and alarm notifications. The interface between the SAF and each gNodeB is denoted as E2++, which is an extended version of the E2 interface defined between the RIC and the gNodeB. In this example, the control of the gNodeB is in near real time.
[0037] According to some examples of the present disclosure, SAF is implemented as a subcomponent of SMO. In doing so, SAF utilizes the internal NSMF / NSSMF of SMO, the performance management function for KPI collection, and the alarm / fault management function for alarm notification. The interface between SAF and each gNodeB is represented as O1++, which is an extended version of the O1 interface defined between SMO and gNodeB. In this example, the control of gNodeB may be non-real time.
[0038] According to some examples of the present disclosure, a subcomponent of SAF is implemented within SMO (SAF SMO ), implement another subcomponent (SAF) within RIC RIC ) to ensure coordination of near-RT and non-RT control of gNodeB. In addition, the A1 interface between RIC and SMO is extended to A1++ to support communication between these two subcomponents of SAF. In such a more complex architecture, SAF SMO Employ algorithms / methods (e.g., those using artificial intelligence / machine learning) to determine required controls and communicate those controls and strategies to the SAF, which executes them in near real time RIC (Using A1++). SAF can implement control in near real-time and non-real-time according to specific circumstances. In this scenario, SAF SMO The interface with gNodeB is O1++, and SAF RIC The interface with gNodeB is E2++.
[0039] 5G RAN supports Orthogonal Frequency Division Multiplexing (OFDM) in the physical layer with different numbers (e.g. different subcarrier spacing and cyclic prefix length) and adaptable time and frequency frame structure, which means selectable time slot duration and dynamic allocation of DL / UL transmission direction. In addition, UEs served by the same cell can be instructed to receive or transmit using only a subset of the cell resource grid. The aforementioned radio parameters are all adjustable in the RAN protocol layer. In addition, the reconfiguration of slice parameters within a cell in the radio network is achieved simply by changing the allocation of cell resource grid shares on, for example, the time and frequency axis. At the upper layers of the protocol stack, the slice SLA is met by appropriate handling of data radio bearers (DRBs), for example, by specific scheduling rules and / or radio protocol stack configuration for the corresponding DRBs and immediate scheduling of URLLC type flows. These are exemplary gNodeB Layer 1, Layer 2 and Layer 3 parameters that can be adjusted within each cell to provide slice SLAs, some of which are defined based on performance requirements such as throughput, energy efficiency, latency and reliability. These requirements are defined to evaluate the SLA for each slice, which needs to be properly handled in the gNodeB protocol layer. In the case of throughput requirements, slice-based physical resource block (PRB) utilization, link adaptation scheme control, modulation and coding scheme (MCS) level, and transport block (TB) size are some exemplary configurable parameters. Similarly, slices with high energy efficiency requirements can be adjusted through discontinuous reception (DRX) configuration. As mentioned earlier, energy-saving UEs can be configured with low bandwidth to prevent energy-consuming wideband operations supported by 5G New Radio (NR). The switching between bandwidth parts (BWPs) is simply guided by the policy received from the SAF. Regarding latency requirements, 5GNR supports non-slot-based scheduling and preemptive scheduling. If the utilization of resource blocks in the BWP is high, the SAF manages the preemption policy, which will allow high priority delay-sensitive services to be served by interrupting resource allocation for other delay-tolerant or lower priority services (e.g., mobile broadband). In order to support the reliability requirements of the slices, an alternative MCS table with low spectral efficiency is introduced. Based on the reliability requirements of the slice, the SAF manages the dynamic MCS table signaling in downlink and uplink data transmission. Related to reliability, the Packet Data Convergence Protocol (PDCP) may be configured to generate repeated packets, which are routed to different RLC entities. Another method for increasing reliability is to configure the UE to repeatedly receive the Physical Data Shared Channel (PDSCH) in consecutive time slots. The advantage of using repetition is reduced latency since the base station does not need to wait for the HARQ acknowledgment from the UE while increasing the probability of successful reception. Each of the above parameters is a non-exhaustive list of slice parameters that parameterize and affect slice operation.Slice parameters may also be referred to herein as "slice configuration parameters" because the SAF may configure components of the mobile network to operate according to the slice parameters.
[0040] In some examples, the SAF obtains the number and type of slices supported for each tracking area and the associated RAN-specific slice SLAs by communicating with the Network Slice Subnet Management Function (NSSMF) responsible for the specific tracking area. When there is a violation, the SAF also sends slice notifications, such as warnings and alarms, to subscribed entities, which can be the NSSMF and / or Service Management and Coordination (SMO), depending on the operator deployment. The SAF further communicates with the SMO so that policies can be received, which will be used for target RAN slice KPI derivation. Depending on the operator's choice, the NSSMF and NSMF can be implemented as components of the SMO or deployed as separate functions.
[0041] The functionality of the SAF may be implemented as (i) integrated into the RAN Intelligent Controller (RIC) for near-RT (near-RT) control, (ii) integrated into the SMO for non-RT control (non-RT), (iii) integrated on the SMO and RIC (as a separate function) for near-RT and non-RT control, or (iv) as a separate virtual or physical network function (VNF / PNF) that communicates directly with the RIC and SMO and may be used for near-RT or non-RT control as the case may be. An example interface toward the gNodeB for (i) above is an extended version of the E2 interface, which is currently defined for collecting near real-time information (e.g., UE basis and cell basis) from the gNodeB and for providing value-added services. In one example, E2 is extended to send reconfiguration commands, denoted by E2++. Another example interface toward the gNodeB for (ii) above is an extended version of the O1 interface, which is currently defined for non-real-time interaction with the gNodeB. In one example, O1 is extended to send reconfiguration commands, denoted by O1++. Another exemplary interface is an extension of the A1 interface between RIC and SMO, which is used to send commands between subcomponents of SAF implemented in RIC and SMO, represented by A1++. In addition, a completely new interface can be designed between SAF and SAF client, which does not rely on existing interfaces, such as O1 and E2.
[0042] Key performance indicators (KPIs) are collected at RAN protocol layers 1, 2 and 3 aggregated using actual user packet flows or synthetically generated packet flows, using in-band and / or out-of-band data collection methods, for example from the gNodeB. These KPIs can be further broken down into each RU, DU and CU component and midhaul and fronthaul facilities to take into account any latency, packet loss or availability associated with these specific sub-components of the RAN.
[0043] An exemplary algorithm of the SAF that can calculate RAN-specific slice resource reallocation calculates the reallocation weights that accommodate traffic bursts by using the concept of resource "ranges" (e.g., 20% ± 5%). These ranges allow only minor SLA violation events that generate non-emergency situations and do not require any RAN slice resource adjustments. In doing so, frequent reconfiguration of gNodeBs can be avoided.
[0044] An electronic device (e.g., gNodeB, UPF user plane function, AMF, controller, etc.) stores and transmits (internally and / or communicates with other electronic devices over a network) code (consisting of software instructions) and data using machine-readable media, such as non-transitory machine-readable media (e.g., machine-readable storage media, such as magnetic disks; optical disks; read-only memory; flash memory devices; phase change memory) and transitory machine-readable transmission media (e.g., electrical, optical, acoustic or other forms of propagated signals, such as carrier waves, infrared signals). In addition, such an electronic device includes hardware, such as a set of one or more processors coupled to one or more other components, such as one or more non-transitory machine-readable storage media (to store code and / or data) and network connections (to transmit code and / or data using propagated signals), and in some cases, user input / output devices (e.g., keyboards, touch screens, and / or displays). The coupling of the set of processors and other components is typically through one or more interconnects (e.g., buses and possible bridges) within the electronic device. Therefore, the non-transitory machine-readable medium of a given electronic device typically stores instructions for execution on one or more processors of the electronic device. One or more portions of the exemplary examples of the techniques of this disclosure may be implemented using different combinations of software, firmware, and / or hardware.
[0045] As used herein, a network device (e.g., a base station, switch, controller, or control function) is a network component that includes hardware and software that communicates and interconnects with other devices of the network (e.g., other network devices and end systems). A switch provides network connectivity to other network devices, such as switches, gateways, and routers that exhibit multi-layer network functions (e.g., routing, layer 3 switching, bridging, VLAN (virtual LAN) switching, layer 2 switching, quality of service, and / or user management), and / or provides support for traffic from multiple application services (e.g., data, voice, and video).
[0046] A user equipment (UE) is a user device, such as a mobile phone, tablet, mobile sensor, computer, or other type of device that is wirelessly connected to a mobile network. Any physical device in the network has a type, location, ID / name, media access control (MAC) address, and Internet Protocol (IP) address. In addition, a physical device can host a set of VNFs, each of which is identified by, for example, a virtual port number and / or a virtual IP address.
[0047] Note that while the illustrated examples in the specification primarily discuss 5G networks that rely on SDN and NFV, the examples for implementing the techniques described herein may also be applicable to other types of networks (mobile and non-mobile) that can be sliced.
[0048] Figure 1 An exemplary radio access network (RAN) 10 is shown with a registration area (RA) 14 having three tracking areas 11, 12 and 13. Tracking area 11 has cells 101, 102, 103 and 104. Tracking area 12 has cells 1111, 1112, 1113 and 1114. Finally, tracking area 13 has cells 2111, 2112, 2113 and 2114. A unique tracking area Id (TAI) identifies each tracking area. RA 14 is defined as a set of tracking areas. A unique cell Id identifies each cell within a TA.
[0049] Each cell in the tracking area has an associated base station (gNodeB). The 5G base station architecture distributes the various components that make up the base station. Figure 2 As shown, the central unit (CU), distributed unit (DU) and radio unit (RU) are the basic components. The RU supports the radio interface and carries the antenna array for this purpose. The brain of the gNodeB is the CU, which mainly performs Layer 3 operations and is controlled by the CU-CP. The CU-CP controls the CU-UP and all its subtended DUs. The figure shows a single CU with three subtended DUs: DU#2 has three RUs (RU#1, #2 and #3), where DU#1 and DU#2 have integrated RUs. Depending on the cell requirements and traffic type, the gNodeB can be deployed in various configurations.
[0050] Figure 3A , Figure 3B , Figure 3C and Figure 3DA simplified diagram comparing slice resource utilization according to existing 3GPP specifications and slice resource reconfiguration according to the techniques of the present disclosure is shown in a simple TA with two cell sites (supported by gNodeB1 and gNodeB2, respectively) and two slice instances (slice 1 and slice 2). The total throughput initially configured for slice 1 and slice 2 is 3 Mbps per slice, which is configured on gNodeB 1 and gNodeB 2 that constitute the tracking area in this simple example.
[0051] Figure 3A The slice throughput achieved over time for slice 1 and slice 2 on gNodeB 1 is shown. We notice that while slice 2 has relatively low traffic, slice 1 traffic exceeds the configured throughput between time t1 and t2. By performing cross-time resource configuration according to this example, the throughput is “borrowed” from slice 2 and transferred to slice 1 on gNodeB 1 by reconfiguring only the throughput parameters of gNodeB 1.
[0052] Figure 3B Another slice throughput achieved at time 0 for slice 1 and slice 2 on gNodeB 1 is shown. Although both slice 1 and slice 2 have high traffic, slice 1 is the higher priority slice. Between time t1 and t2, the traffic of both slices exceeds the configured throughput. By performing cross-time resource configuration according to this example, the throughput is borrowed from the lower priority slice 2 and transferred to slice 1 on gNodeB 1 by reconfiguring only the throughput parameters of gNodeB 1.
[0053] Figure 3C The slice throughput achieved for slice 1 on gNodeB 1 and gNodeB 2 over time is shown. We note that while the traffic for slice 1 is relatively low on gNodeB 2, it exceeds the throughput configured in gNodeB 1. By performing cross-area resource configuration according to the techniques of the present disclosure, the throughput is borrowed from gNodeB 2 by forcing users of slice 1 on gNodeB 1 to switch to gNodeB 2. For example, SAF 400 may reconfigure gNodeB 1 and / or gNodeBs with different antenna powers, or change the load balancing parameters of the gNodeBs, to force the switch.
[0054] Figure 3DAnother example of slice throughput achieved over time for slice 1 on gNodeB 1 and gNodeB 2 is shown. Slice 1 has high traffic that exceeds the configured throughput on gNodeB 1 and gNodeB 2. The excess slice 1 traffic on gNodeB 1 is satisfied by borrowing throughput from gNodeB 2 by performing cross-region resource configuration according to this example. While the throughput SLA for slice 1 is still not met on gNodeB 2, it is met on gNodeB 1 by doing so. According to this example, these reconfigurations show how dynamic weights can be assigned to the throughput distribution on the base stations across regions and across time on the slices. Figure 3A and Figure 3B On gNodeB 1, between time t1 and t2, a higher weight of slice throughput is given to slice 1 (25% more). Figure 3D In the example above, a higher weight (25% more) is given to the gNodeB 1 throughput compared to gNodeB 2 of slice 1. Doing so improves the throughput across time and per region slice.
[0055] Let I = {1, 2...n} represent the slice group in a TA, and C = {1, 2...m} represent the cell group in a TA. For each slice i∈I, let d i Indicates the guaranteed data rate requirement of each UE indicated in the SLA. Let the set U i is the UE receiving service from slice i. In this case, R u Indicates UEu∈U i The experienced data rate.
[0056] Therefore, the following expression is an example performance indicator that shows whether the UE in slice i∈I achieves the slice-specific guaranteed data rate requirement.
[0057]
[0058] Among them, x u is UE u∈U i The normalized rate of , bounded between 0 and 1.
[0059]
[0060] The SAF uses the result of this expression to update RAN specific parameters, which directly affect the UE rate. As mentioned before, the PRB allocation per slice is a parameter for the throughput requirement. indicates the percentage of PRBs allocated to slice i∈I and cell j∈C. Therefore,
[0061]
[0062] The PRB distribution over different slices within a TA and the multiple cells forming the TA can be represented by a matrix to better understand the results of reconfiguration across regions and across time. The columns represent the different types of slices in the TA, while the rows represent the configurable cells / gNodeBs in the TA. In the exemplary scenario, it is assumed that there are four slices (I1, I2, I3, and I4) and three cells (C1, C2, and C3) in the TA. Slice 3 is a "high priority" slice (from the perspective of one or more SLAs). The resource allocation percentages at time t=t0 are shown below. Note that the total resources of each cell must be 1 across all slices:
[0063]
[0064] At time t=t1, in cell C1, the traffic of slice 3 increases significantly. SAF checks whether the guaranteed throughput of each UE in slice 3 is met. In addition, SAF checks the radio resource utilization and the load of the slice, for example, the average number of PRBs used for data traffic in the previous time interval. According to an example of the technology of the present disclosure, C1 resources are borrowed from I1 and I2 and transferred to I3, thereby providing a higher percentage of usage to slice 3.
[0065]
[0066] The SAF optimization algorithm calculates resource allocation configuration parameters for each cell and each slice during each observation period. Note that the SAF optimization algorithm can use simple heuristic algorithms, many of which are known in the prior art, or complex machine learning techniques using artificial intelligence (AI) (e.g., deep learning) that learn and improve the response of slice traffic to various resource allocation actions over time.
[0067] Figure 4A A tracking area 420 is shown with many cell sites: gNodeBs 422a, 422b, 422c, ... 422n. Distributed components of gNodeB 422a are also shown. These are CU-CP 452a, CU-UP 432a, and DUs 442a-k. CU-CP 452a controls CU-UP 432a using an E1 interface and controls DU 442a using an F1-C interface. Example components for implementing the techniques of the present disclosure may include: (i) SAF 400, which makes decisions on resource reallocation across time and area based on RAN-specific SLAs and collected KPIs and controls all of the base stations, (ii) interfaces between SAF 400 and all gNodeBs, E2 ++ interface, and (iii) SAF clients 451a, 451b, ... 451n, via E2 ++The interface receives reconfiguration commands and in some cases can execute them at near RT.
[0068] The SAF 400 obtains each slice type supported in the tracking area and the SLA associated with each slice from the Network Slice Subnet Management Function (NSSMF) 404 through the RIC 410. It is assumed that the NSSMF 404 obtains the RAN-specific slice SLA catalog and specific SLA component information from the NSMF 406.
[0069] SAF 400 uses E2 ++ The interface collects near real-time performance data related to the slice SLA from all base stations or directly via the SMO 408 using the A1 interface 458. The SMO 408 receives notifications such as warnings and alarms from the SAF 400 when the slice SLA cannot be met using the same interface.
[0070] Figure 4B Another example of SAF 400 is shown, where the SAF is a component of the SMO (rather than the RIC). In this scenario, the interface used for reconfiguration to the gNodeB is an extended O1 interface, with O1 ++ Because O1 is designed for non-RT operation, a consequence of this example may be that the reconfiguration of the gNodeB is less frequent. Although not shown in a separate figure, one can easily Figure 4A and Figure 4B superimposed to obtain yet another alternative configuration, in which some components of the SAF are in the RIC (such as Figure 4A ), while other components are in SMO (such as Figure 4B This may require extensions in the A1 interface between the RIC and SMO to support messaging within the SAF. This interface is called A1 ++ , according to the functional division of SAF on these two systems, they have different functions. This disclosure covers such changes in implementation.
[0071] In some cases, SAF 400 may apply artificial intelligence or other machine learning models trained using historical KPIs to predict possible failures of slices to meet corresponding SLAs. In response to identifying an upcoming possible failure based on recent KPIs, SAF 400 may reallocate slice resources and reconfigure the eNodeB before the predicted failure time to avoid the failure.
[0072] Figure 5is a high-level block diagram of the SAF 400 according to some examples of the SAF. The slice directory and slice SLA database 505 are populated with information collected from the NSSMF. The slice performance collector 503, which collects measurements from the base stations, populates the real-time (RT) RAN slice KPI database 506. The slice parameter NRT database 507 stores all determined slice parameters assigned to each base station in the TA over time. The slice optimizer subsystem 500 across time and region calculates weights to assign the slice SLA configuration of the TA to each base station configuration. The slice control action subsystem 501 maps the weights to appropriate base station control actions and sends these actions to the base stations 422a...n. The interface between the SAF and the SAF client supports simple exemplary messages, such as:
[0073] ·<ReconfigRequest{ID}{component ID}{time stamp}{param:value}{param:value}…{param:value}> : A request by SAF to reconfigure resources with new parameter values.
[0074] ·<FallBackToDefault{component ID}{time stamp}> : A request to revert to its original / default slice configuration.
[0075] ·<ReconfigResponseSuccess{param:value1}{param:value1}…{param:value1}{component ID}{time stamp}> : gNodeB's response to successful SAF reconfiguration
[0076] ·<ReconfigResponseFailure{param:value1}{param:value1}…{param:value1}{component ID}{time stamp}> : gNodeB response to failure to implement the requested reconfiguration
[0077] If the slice SLA is not met, the slice notification subsystem 502 prepares an alarm or warning to be sent to the alarm management subfunction of the SMO 408. Once the slice catalog and slice SLA are obtained, the subsystem 516 determines the slice KPIs specific to each TA RAN and their targets and thresholds. Subsystem 515 generates performance management jobs (tasks) to be executed by each gNodeB to collect and report the slice KPIs for consumption by the SAF 400.
[0078] Figure 6 and Figure 7 Exemplary methods or steps for implementing the disclosed technology are shown. Figure 6 In step 601 of the TRA, SAF 400 collects all slice types and associated SLAs supported by the TA from NSSMF 603. In step 611, SAF 400 derives the slice KPIs, their target values and thresholds. Thereafter, in step 671, SAF 400 generates the PM jobs that each gNodeB must perform for slice-specific KPI monitoring. Finally, in step 681, RIC 410 transmits the required KPIs to each gNodeB in the tracking area and starts collecting KPIs from each gNodeB for consumption by SAF 400.
[0079] exist Figure 7 At time T in step 602, SAF 400 collects KPIs related to each slice from all base stations. In check box 604, SAF 400 first checks to determine whether the slice SLA of the TA is met for each flow within the slice. This check and subsequent steps are performed for all slices, which can be done iteratively or in parallel. If the answer to check box 604 is yes, SAF 400 waits in step 611 until the next time interval (T+1) to repeat the process.
[0080] In some examples, the SAF 400 optionally (as indicated by the dashed line of 609) determines whether a prediction-based SLA adjustment should be performed (step 609). If not, in step 611, the SAF 400 waits until the next time interval (T+1) to repeat the process for the slice. If a prediction-based SLA adjustment is to be performed (the "yes" branch of step 609), the SAF 400 performs step 618, which in these examples includes re-optimizing the slice parameters by applying a prediction technique that is used to predict possible failures of the slice to meet the corresponding SLA, such as those described above. For example, the SAF 400 may predict that an SLA for the corresponding slice is not met, and in response, preemptively reallocate slice resources to calculate new slice configuration parameters for the slice. The SAF 400 may then reconfigure the base station to implement the new slice configuration parameters for the slice.
[0081] However, if some slice SLAs are not met (the "No" branch of 604), the SAF 400 checks to determine in checkbox 605 whether the total slice SLA for TA within time period T is met. If not, an alarm is sent to the SMO in step 621. Otherwise, in step 618, the SAF 400 reallocates slice resources on one or more base stations. This step can be performed by the cross-time and cross-region slice optimizer subsystem 500, which calculates new slice resource parameters, which will implement the slice resource parameters according to the reallocation when configured into the gNodeB using the slice configuration parameters calculated by the SAF 400. In step 622, the SAF 400 checks to determine whether there is a feasible solution for reallocation so that all slices can meet their associated SLAs. If not, in step 621, the SAF 400 sends an alarm to the alarm management in the SMO. Otherwise, in step 642, the SAF 400 sends new slice configuration parameters to at least one base station to implement the new slice resources calculated for the slice. Any of the CU-CP, CU-UP or DU of one or more base stations can receive new slice configuration parameters for reconfiguring one or more base stations.
[0082] abbreviation
[0083] 5G: The fifth generation of mobile wireless communication system
[0084] AMF: Access and Mobility Function
[0085] CP: Control Plane
[0086] CU: Central Unit
[0087] DU: Distributed Unit
[0088] gNodeB: base station
[0089] IoT: Internet of Things
[0090] IP: Internet Protocol
[0091] KPI: Key Performance Indicator
[0092] NRF: Network Knowledge Base Function
[0093] NSSAI: Network Slice Selection Assistance Information
[0094] NSMF: Network Slice Management Function
[0095] NSSMF: Network Slicing Subnet Management Function
[0096] PM: Performance Management
[0097] RA: Registration Area
[0098] RU: Radio Unit
[0099] QoS: Quality of Service
[0100] RAN: Radio Access Network
[0101] RIC: RAN Intelligent Controller
[0102] SDN: Software Defined Networking
[0103] SLA: Service Level Agreement
[0104] SMO: Service Management and Orchestration
[0105] SST: Slice Service Type
[0106] TA: Tracking Area
[0107] UE: User Equipment
[0108] UP: User Plane
[0109] VNF: Virtualized Network Function
[0110] Computer executable instructions include, for example, instructions and data that cause a general purpose computer, a special purpose computer, or a special purpose processing device to perform a specific function or group of functions. Computer executable instructions also include program modules that are executed by computers in a stand-alone or network environment. Typically, program modules include routines, programs, components, data structures, objects, functions, etc. inherent in the design of a special purpose processor that performs a specific task or implements a specific abstract data type. Computer executable instructions, related data structures, and program modules represent examples of program code means for performing the steps of the methods disclosed herein. A specific sequence of such executable instructions or related data structures represents an example of corresponding actions for implementing the functions described in these steps.
[0111] Processors suitable for executing computer programs include, for example, general and special purpose microprocessors and any one or more processors of any type of digital computer. Typically, the processor will receive instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for performing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include, or be operatively coupled to, receive data from or transfer data to, or both, one or more mass storage devices (e.g., magnetic, magneto-optical, or optical disks) for storing data.
[0112] In this specification, the term "software" is intended to include firmware residing in read-only memory or applications stored in magnetic storage or flash memory, such as a solid-state drive that can be read into memory for processing by a processor. In addition, in some implementations, multiple software technologies can be implemented as sub-parts of a larger program while maintaining different software technologies. In some implementations, multiple software technologies can also be implemented as separate programs. Finally, any combination of independent programs that jointly implement the software technologies described herein are within the scope of the subject technology. In some implementations, when the software programs are installed to run on one or more electronic systems, these software programs define one or more specific machine implementations that execute and perform the operations of the software programs.
[0113] A computer program (also referred to as a program, software, software application, script, or code) may be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, object, or other unit suitable for a computing environment. A computer program may, but need not, correspond to a file in a file system. A program may be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple cooperating files (e.g., files storing one or more modules, subroutines, or portions of code). A computer program may be deployed to execute on a single computer or on multiple computers located at a single site or distributed across multiple sites and interconnected by a communications network.
[0114] The above functions can be implemented in digital electronic circuits, computer software, firmware or hardware. These technologies can be implemented using one or more computer program products. Programmable processors and computers can be included in mobile devices or packaged as mobile devices. These processes and logic flows can be performed by one or more programmable processors and one or more programmable logic circuits or processing circuits. General and special computing devices and storage devices can be interconnected through a communication network.
[0115] Some implementations include electronic components, such as microprocessors, storage, and memory, which store computer program instructions in a machine-readable or computer-readable medium (also referred to as a computer-readable storage medium, a machine-readable medium, or a machine-readable storage medium). Some examples of such computer-readable media include RAM, ROM, compact disk-read only (CD-ROM), compact disk-recordable (CD-R), compact disk-rewritable (CD-RW), digital versatile disk-read only (e.g., DVD-ROM, dual-layer DVD-ROM), various recordable / rewritable DVDs (e.g., DVD-RAM, DVD-RW, DVD+RW, etc.), flash memory (e.g., SD card, mini SD card, micro SD card, etc.), magnetic or solid-state hard drives, read-only and recordable CD-ROM, ultra-density CD-ROM, any other optical or magnetic medium, and floppy disk. Computer readable media can store a computer program executable by at least one processing unit and include instruction sets for performing various operations. Examples of computer programs or computer code include machine code, such as produced by a compiler, and files including high-level code that is executed by a computer, electronic component, or microprocessor using an interpreter.
[0116] Although the above discussion primarily involves microprocessors or multi-core processors that execute software, some implementations are performed by one or more integrated circuits, such as application specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs). In some implementations, such integrated circuits execute instructions stored on the circuits themselves.
[0117] The terms "computer-readable medium" and "computer-readable media" as used in this specification and any claims of this application are strictly limited to tangible physical objects that store information in a computer-readable form. These terms exclude any wireless signals, wired download signals, and any other ephemeral signals.
Claims
1. A method comprising: obtaining, by a slice assurance function (SAF) executed by a device, a key performance indicator (KPI) value of a first slice of a plurality of slices, the plurality of slices being implemented by a plurality of base stations serving a tracking area of a mobile network; Determining, by the SAF based in part on the KPI value of the first slice, that a service level agreement (SLA) of the first slice is not met; In response to the determination, the SAF reallocates slice resources associated with any one of the plurality of slices to calculate new slice configuration parameters for the first slice, wherein the calculation comprises determining a reallocation ratio of radio access network resources across a plurality of base stations by calculating a distribution of slice resources for each base station across the plurality of slices; and reconfiguring, by the SAF, at least one of the plurality of base stations to implement the new slice configuration parameters of the first slice, wherein the at least one of the plurality of base stations comprises a SAF client, and Among them, reconfiguring at least one of the multiple base stations to implement the new slice configuration parameters of the first slice includes: the SAF outputting a configuration command to the SAF client so that the SAF client reconfigures at least one of the multiple base stations.
2. The method according to claim 1, wherein: Determining that the SLA for the first slice is not met includes determining that the SLA for the first slice is not met for at least one flow associated with the first slice.
3. The method according to claim 1, further comprising: An alarm is outputted by the SAF in response to determining that a total slice SLA for the tracking area will not be met within a period of time, the alarm indicating that resources of the tracking area are insufficient to meet the SLA for the first slice.
4. The method according to claim 1, wherein: Reallocating the slice resource associated with any one of the plurality of slices comprises: Slice resources not used by a second slice among the plurality of slices are reallocated from the second slice to the first slice to calculate the new slice configuration parameter of the first slice.
5. The method according to claim 1, wherein: Reallocating the slice resources associated with the any one of the plurality of slices to calculate the new slice configuration parameter of the first slice includes: Slice resources used by a second slice of the plurality of slices having a lower priority than the first slice are reallocated from the second slice to the first slice to calculate the new slice configuration parameter of the first slice.
6. The method according to claim 1, wherein: Reallocating slice resources associated with the any one of the plurality of slices to calculate the new slice configuration parameter of the first slice includes: The user equipment UE using the first slice is forced to switch from the first base station among the multiple base stations to the second base station among the multiple base stations that supports the first slice.
7. The method according to claim 6, wherein: The new slice configuration parameters of the first slice include an identifier of the second base station to serve the UE.
8. The method according to claim 1, further comprising: Obtaining, by the SAF, a KPI value of a second slice among the multiple slices; Predicting, by the SAF based in part on the KPI value of the second slice, that an SLA for the second slice will not be met; In response to the determination, reallocating, by the SAF, slice resources associated with any one of the plurality of slices to calculate new slice configuration parameters for the second slice; and At least one of the multiple base stations is reconfigured by the SAF to implement the new slice configuration parameters of the second slice.
9. The method according to claim 1, wherein: The SAF and the SAF client communicate via an interface, the interface comprising one of an O1 interface, an E2 interface, or a 3GPP-based performance management interface, wherein the interface is extended to enable collection of the KPI value and reporting of the KPI value from the SAF client to the SAF.
10. The method according to claim 1, wherein: The device comprises a Radio Access Network RAN Intelligent Controller RIC.
11. The method according to claim 10, wherein: An E2 interface between the RIC and the at least one of the plurality of base stations is extended to support communication between a SAF and the SAF client.
12. The method according to claim 1, wherein: The device includes a service management and coordination SMO system.
13. The method according to claim 12, wherein: An O1 interface or a 3GPP-based management interface between the SMO system and the at least one of the plurality of base stations is extended to support communication between the SAF and the SAF client.
14. The method according to claim 1, wherein: The SAF includes a SAF component executed by a radio access network RAN intelligent controller RIC and a SAF component executed by a service management and coordination SMO system.
15. The method according to claim 14, wherein: The A1 interface is extended to support communication between the SAF component executed by the RIC and the SAF component executed by the SMO system.
16. The method according to claim 1, further comprising: The SAF collects the SLAs of all the slices; determining, by the SAF based on the SLA, a slice KPI to be monitored within the tracking area for each of the plurality of slices; and The SAF generates a performance management job and outputs the performance management job to the multiple base stations so that the multiple base stations monitor and report the slice KPI.
17. The method of claim 1 , further comprising, for each additional slice in the plurality of slices, iteratively: Obtaining, by the SAF, a KPI value of the additional slice; determining, by the SAF based in part on the KPI value of the additional slice, that the SLA for the additional slice is not met; In response to the determination, reallocating, by the SAF, slice resources associated with any one of the plurality of slices to calculate new slice configuration parameters for the additional slice; and At least one of the plurality of base stations is reconfigured by the SAF to implement the new slice configuration parameters for the additional slice.
18. A computing device comprising: The slice guarantee function is executed on the processing circuit and is configured to: Obtain a key performance indicator KPI value of a first slice among multiple slices, where the multiple slices are implemented by multiple base stations serving a tracking area of a mobile network, each base station including a corresponding slice assurance function client; Determining, based in part on the KPI value of the first slice, that a service level agreement (SLA) for the first slice is not met; In response to the determination, reallocating slice resources associated with any one of the plurality of slices to calculate new slice configuration parameters for the first slice, wherein the calculating comprises determining a reallocation ratio of radio access network resources across a plurality of base stations by calculating a distribution of slice resources for each base station across the plurality of slices; and The new slice configuration parameters of the first slice are transmitted to a slice guarantee function client via an interface, so that the slice guarantee function client reconfigures the corresponding base station to implement the new slice configuration parameters of the first slice.
19. A mobile network comprising: a plurality of base stations, including respective corresponding slice assurance function clients, the plurality of base stations serving a tracking area of the mobile network; as well as The slice guarantee function is executed on the processing circuit and is configured to: Obtaining a key performance indicator KPI value of a first slice among multiple slices, where the multiple slices are implemented by the multiple base stations; Determining, based in part on the KPI value of the first slice, that a service level agreement (SLA) for the first slice is not met; In response to the determination, reallocating slice resources associated with any one of the plurality of slices to calculate new slice configuration parameters for the first slice, wherein the calculating comprises determining a reallocation ratio of radio access network resources across a plurality of base stations by calculating a distribution of slice resources for each base station across the plurality of slices; and The new slice configuration parameters of the first slice are transmitted to a slice guarantee function client via an interface, so that the slice guarantee function client reconfigures the corresponding base station to implement the new slice configuration parameters of the first slice.
20. A computer readable storage medium encoded with instructions for causing one or more programmable processors to perform the method of any one of claims 1 to 17.
