DE102021124688A1 - ALLOW AUTOMATIC SCALE OF COMPUTING POWER BETWEEN DEVICE AND NETWORK WITH DYNAMIC RESOURCE AVAILABILITY - Google Patents

Abstract

A new radio (NR) user equipment (UE) apparatus that includes a radio frequency interface coupled to processors. The processors include a Compute Client Service Function (Comp CSF) configured to register local compute resources with an application and register local and remote compute resources with an operating system/middleware (OS/middleware). . The Comp CSF is configured to perform registration for remote computing resources with a Computing Control Function (Comp CF). The Comp CSF is configured to determine the local and remote computing resources based on registration with the application, the OS/middleware and the Comp CF. The Comp CSF is configured to receive a dynamic swap message and a dynamic swap request message for a computation task from the application. The Comp CSF is configured to dynamically offload the compute task based on the determined local and remote compute resources and the dynamic swap request.

Description

CROSS REFERENCE TO RELATED APPLICATION

This patent application claims the benefit of and priority to U.S. Provisional Application No. 63/083,725, filed September 25, 2020.

TECHNICAL AREA

Various aspects may relate to the field of wireless communications in general.

BACKGROUND

Scaling data processing can be addressed in two scenarios. The first scenario can deal with scaling computing at the node, rack, facility, and computer room level. The second scenario can address scaling computing at the edge and in the cloud (data centers).

character list

• 1 illustrates the architecture that can enable scaling of computing between the UE and the network;
• 2 illustrates the interaction between the elements of a Compute Client Service Function (Comp CSF);
• 3 illustrates an example message sequence diagram for notification/notification of computing resources and events/status between the application and the Comp CSF, the Comp CSF to the operating system/middleware, or the Comp CSF and a Computing Control Function (Comp CF);
• 4 illustrates an example message sequence diagram for dynamic workload offload registration;
• 5 illustrates an example resource request/response message sequence diagram between the application and the Comp CSF;
• 6 Figure 12 shows an example message sequence diagram for requesting local resource information from the operating system/middleware;
• 7 Figure 12 shows an example message sequence diagram for requesting local resources from Comp CSF to BS/Middleware;
• 8th Figure 12 shows an example message order diagram for requesting remote computer control or data;
• 9 12 illustrates an example message sequence diagram for registering the Comp CSF with the Comp CF for dynamic data processing offloading;
• 10 Figure 12 illustrates an example message sequence diagram for the Comp CSF to request the Comp CF to discover network computing resources;
• 11 illustrates an example message order diagram for Comp CSF to Comp CF requesting network resources;
• 12 illustrates the interaction between the elements for the Comp RF;
• 13 Figure 12 illustrates an example message sequence diagram for a resource discovery request to the Comp RF-C within the UE modem for offloading of computational work;
• 14 Figure 12 illustrates an example message sequence diagram for a request for computational resources to the Comp RF-C in the modem for offloading computational work;
• 15 Figure 12 illustrates an example message sequence diagram of the operating system/middleware sending computational data to Comp RF-C for delivery to Comp RF-S for offloading of computational work;
• 16 illustrates the Comp RF in the modem for offloading computational tasks;
• 17 illustrates deployment options for computing devices connected to the CU/DU;
• 18 illustrates an example message sequence diagram of a query for computing power;
• 19 illustrates an example message order diagram of a request for a remote computing resource;
• 20 illustrates a network;
• 21 schematically illustrates a wireless network; and
• 22 Figure 12 is a block diagram illustrating components capable of reading instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium), according to some example aspects. to read and perform one or more of the methods discussed herein, all in accordance with at least one aspect described in the present disclosure/invention.

DETAILED DESCRIPTION

The following detailed description refers to the accompanying drawings. The same reference numbers may be used in different drawings to identify the same or similar items. In the following description, for purposes of explanation and not limitation, specific details such as particular structures, architectures, interfaces, techniques, etc. are set forth in order to provide a thorough understanding of the various aspects. However, it will be apparent to those skilled in the art, familiar with the present disclosure, that the various aspects may be implemented in other examples that depart from these specific details. In certain instances, descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of various aspects with unnecessary detail. For purposes of this document, the terms “A or B” and “A/B” mean (A), (B) or (A and B).

Scaling data processing can be addressed in two scenarios. The first scenario can deal with scaling computing at the node, rack, facility, and computer room level. The second scenario can address scaling computing at the edge and in the cloud (data centers).

• - Rechenressourcen auf einem Host zusammengelegt oder mit Verbindungen mit hohem Durchsatz verbunden werden können;
• - eine Rechenaufgabe entweder auf Aufgabenebene oder auf Datenebene parallelisiert werden kann;
• - die Parallelisierung von Rechenaufgaben durch Laufzeit/Betriebssystem (BS)/Orchestrator mit voller Ressourcentransparenz erfolgen kann; und
• - Datenverarbeitung und Kommunikation zur Leistungsoptimierung gemeinsam betrachtet werden können.

At the node/rack/facility, the computing infrastructure can be designed with greater consideration for latency and efficiency, where

• - Computational resources can be pooled on a host or connected to high-throughput links;
• - a computational task can be parallelized either at the task level or at the data level;
• - the parallelization of computing tasks can be done by runtime/operating system (OS)/orchestrator with full resource transparency; and
• - Data processing and communication for performance optimization can be viewed together.

• - die Rechenressourcen auf den Geräten begrenzt sein können, so dass eine Anwendung in einem Client-Server-Modell mit einem allgemein leichten Client arbeiten kann;
• - die Rechenaufgabe eine Parallelisierung auf Dienstebene beinhalten kann und von der Anwendung gehandhabt werden kann (z.B. kann die Aufteilung der Arbeitslast zwischen lokal und remote im Programm vorcodiert sein); und
• - Datenverarbeitung und Kommunikation im Allgemeinen getrennt betrachtet werden können (z.B. kann die Kommunikation als Datenleitung genutzt werden, und Edge/Cloud-Endpunkte können Dienste auf Anwendungs- oder Mikrodienstebene bereitstellen).

At the edge/cloud location, edge/cloud servers can be designed to provide interfaces to services and achieve scalability of services where:

• - the computing resources on the devices may be limited, allowing an application to work in a client-server model with a generally lightweight client;
• - the computational task may involve service level parallelization and be handled by the application (eg the division of workload between local and remote may be pre-coded in the program); and
• - Computing and communications in general can be considered separately (e.g. communications can be used as a data pipeline, and edge/cloud endpoints can provide application or microservices level services).

• - keine Vorkenntnisse über Rechenressourcen bei der Entwicklung der Anwendung;
• - die Geräte-Laufzeit bzw. das Betriebssystem enthält möglicherweise keine vollständige Übersicht über die verfügbaren Rechenressourcen;
• - ein Netzwerk-Orchestrator ist möglicherweise nicht vollständig über die Rechenressourcen und die Dynamik des Geräts informiert; und
• - eine gemeinsame Optimierung kann mit der Kommunikation erfolgen.

In the post fifth generation (5G) (B5G)/sixth generation (6G) era, a cellular network can offer orders of magnitude performance improvements, such as higher throughput on wireless links. User Equipment (UE) is typically resource constrained and may not be able to handle the ever-increasing demands of compute-intensive workloads. Meanwhile, the cellular network, with its large scale and long range, can provide not only connectivity but also a variety of services including data processing. Therefore, the issue of scaling device and network computing can be viewed with the offloading of computing between the device and the cellular network with “dynamic” resource availability due to changing network resources and the status of the connections. Dynamic scaling of device and network calculations can have the following characteristics

• - no prior knowledge of computing resources when developing the application;
• - the device runtime or the operating system may not contain a complete overview of the available computing resources;
• - a network orchestrator may not be fully aware of the computing resources and dynamics of the device; and
• - a joint optimization can take place with the communication.

• - eine Comp CSF, die oberhalb des Betriebssystems oder als Teil des Betriebssystems/der Middleware positioniert ist, um eine Schnittstelle zwischen der Netzwerk-Berechnen-Steuerung und den Funktionen der Benutzerebene herzustellen; und
• - einen Rechenressourcen-Funktions-Client (Comp RF-C) auf der Seite des UE und einen Rechenressourcen-Funktions-Server (Comp RF-S) auf der Netzwerkseite unterhalb des Betriebssystems, um ein direktes Auslagern auf Rechengeräte zu ermöglichen.

The aspects described in the present disclosure can dynamic computational scaling between the device and the zel ular network with the aforementioned characteristics and may include:

• - a Comp CSF positioned on top of the operating system or as part of the operating system/middleware to interface between network computing control and user plane functions; and
• - a compute resource function client (Comp RF-C) on the UE side and a compute resource function server (Comp RF-S) on the network side below the operating system to enable direct offloading to computing devices.

Some scaling technologies may not be able to automatically scale devices and networks with dynamic resource availability.

Aspects described in the present disclosure may enable scaling of computing between the UE and the cellular network with dynamic resource availability. The Comp CSF can interface with the application and dynamically virtualize/convert local/remote resources for offloading of computing. The Comp RF-C can interface with the Comp RF-S of the network to discover and offload computing resources. Interfaces and message exchange can be identified and described in detail. These aspects can coexist in the UE and be used for different computing scenarios.

Several benefits can accrue from deploying computing either as a new service or as a network capability based on different scenarios. First, the computing tasks can be performed at the network edge to optimize latency. The latency can include the communication latency as well as the start and execution latency of the computing tasks. Second, the terminal can extend the data processing task by providing settings related to the data processing environment and the data processing task. Third, resource efficiency and latency can be optimized using paradigms such as serverless computing to handle a more dynamic workload. The aspects presented here enable these computing scenarios and can enable computing and storage capacities on a large scale.

1 FIG. 11 illustrates an architecture that may enable scaling of computing power between the UE and the network in accordance with at least one aspect described in the present disclosure. 1 shows an architecture that enables dynamic offloading of the workload. The network computation enabling functions may include the Comp CSF on the UE side, a Radio Access Network (RAN) Comp CF and a RAN Computing Service Function (Comp SF) on the network side. The Comp RF-C on the UE side and the Comp RF-S on the network side can perform the dynamic workload shifting. The Comp RF-S can reside in a centralized unit (CU), a distributed unit (DU), or in both the CU and the DU.

1

eine Schnittstelle zwischen dem Comp CSF und einem RAN Comp CF;

2

eine Schnittstelle zwischen dem UE und einer RAN DU;

3

eine Schnittstelle zwischen dem UE und dem RAN DU;

4

eine Schnittstelle zwischen dem RAN Comp CF und dem RAN Comp SF;

5

eine Schnittstelle zwischen der RAN Comp SF und einer Datenebene;

6

eine Schnittstelle zwischen dem RAN Comp CF und einer RAN CU-Steuerebene (CP) oder einer RAN CU-Nutzerebene (UP);

7

eine Schnittstelle zwischen der RAN Comp CF und den NFs des Kernnetzwerkes (CN) (z.B. eine Network Exposure Function (NEF), eine Policy Control Function (PCF), eine Access and Mobility Management Function (AMF));

8

eine Schnittstelle zwischen dem RAN Comp CF und einer Betriebsverwaltungs- und Wartungsfunktion (OAM);

9

eine Schnittstellenverbindung zwischen dem RAN Comp CF und der Datenebene;

10

eine Schnittstelle zwischen dem RAN Comp SF und dem CN Comp SF;

11

eine Schnittstelle zwischen dem RAN Comp CF und dem CN Comp CF;

12

eine Schnittstelle zwischen dem RAN Comp CF und dem RAN CF (z.B. dem NEF, dem PCF oder dem AMF);

13

eine Schnittstelle zwischen dem RAN Comp SF und dem RAN CU-CP oder dem CU-UP;

14

eine Schnittstelle zwischen dem Comp CSF und dem RAN Comp SF;

15

eine Schnittstelle zwischen dem RAN DU und dem RAN CU-CP;

16

eine Schnittstelle zwischen dem RAN DU und dem RAN CU-UP;

17

eine Schnittstelle zwischen der RAN DU und dem RAN Comp CF; und

18

eine Schnittstelle zwischen der RAN DU und der RAN Comp SF.

Compute resources can include platforms, accelerators, and so on. The digits may refer to:

1

an interface between the Comp CSF and a RAN Comp CF;

2

an interface between the UE and a RAN DU;

3

an interface between the UE and the RAN DU;

4

an interface between the RAN Comp CF and the RAN Comp SF;

5

an interface between the RAN Comp SF and a data plane;

6

an interface between the RAN Comp CF and a RAN CU control plane (CP) or a RAN CU user plane (UP);

7

an interface between the RAN Comp CF and the NFs of the core network (CN) (e.g. a Network Exposure Function (NEF), a Policy Control Function (PCF), an Access and Mobility Management Function (AMF));

8th

an interface between the RAN Comp CF and an operations management and maintenance function (OAM);

9

an interface connection between the RAN Comp CF and the data plane;

10

an interface between the RAN Comp SF and the CN Comp SF;

11

an interface between the RAN Comp CF and the CN Comp CF;

12

an interface between the RAN Comp CF and the RAN CF (e.g. the NEF, the PCF or the AMF);

13

an interface between the RAN Comp SF and the RAN CU-CP or the CU-UP;

14

an interface between the Comp CSF and the RAN Comp SF;

15

an interface between the RAN DU and the RAN CU-CP;

16

an interface between the RAN DU and the RAN CU-UP;

17

an interface between the RAN DU and the RAN Comp CF; and

18

an interface between the RAN DU and the RAN Comp SF.

Note: The digits 1 and 14 can correspond to a logic and can be mapped to several reference points.
the COMP CSF is based on a dynamic scaling of the device network calculation.

2 illustrates the interaction between the elements of the Comp CSF according to at least one aspect described in the present disclosure. The interaction between the application, the Comp CSF, the operating system/middleware and the Comp CF/SF is in 2 shown. The application, capable of dynamically balancing the workload, can call the Comp CSF to obtain resources using application programming interfaces (APIs) through the Comp1 interface instead of calling the operating system functions.

Note: The reference numbers 1/14 and 2/3 in 2 correspond to the reference numbers in 1 .

The decision whether to perform the computation on a local or remote resource can be made by the Comp CSF or the CF on the network side. the Comp CSF can decide whether to run the workload on a local resource or a remote resource based on the information from the network via interface 1 and the information about the availability of local resources via interface Comp2 or the Comp CF on the network side should.

The decision can be made on the network side, and the Comp CSF can provide information about the task and the status of the local resource and be notified of the decision from the network via interface 1, and can perform the workload offload by triggering appropriate functions.

the Comp CSF can call the operating system/middleware functions to cause a computation to run locally. Otherwise, the Comp CSF can invoke remote services through the Comp CF using protocols such as Remote Procedures Call (RPC) on the network. The operating system/middleware on the device can enable the Comp CSF to manage local resources or cause the communication unit to transport compute tasks to use remote resources via the Comp2 interface.

• - Durchführen von CP/UP-Funktionen auf der Seite des UE, um:
  • -- ein lokales/entferntes Ressourcen-/Mikrodienst-Management für die Anwendung über die Schnittstelle Comp1 bereitzustellen;
  • -- Anwendungs-Arbeitslasten in lokale Arbeitslasten (die lokal im Gerät ausgeführt werden können) und/oder Remote-Arbeitslasten (die remote im Netzwerk ausgeführt werden können) zu übersetzen; und
  • -- mit dem Betriebssystem/der Middleware über die Schnittstelle Comp2 zur Abfrage und Überwachung von Ressourcen/Geräten, zur Ressourcenvirtualisierung und Orchestrierung zu interagieren.
• - Zusammenwirken mit dem Netzwerk Comp CF/SF für die Fernverwaltung von Ressourcen/Diensten, die Verwaltung des Rechentransports und die Signalisierung/Datenübertragung für CPIUP-Verkehr über die Schnittstellen 1 oder 14.

Comp CSF functions may include:

• - performing CP/UP functions on the UE side to:
  • -- provide local/remote resource/microservices management for the application via the Comp1 interface;
  • -- translate application workloads into local workloads (which can run locally on the device) and/or remote workloads (which can run remotely on the network); and
  • -- Interact with the operating system/middleware via the Comp2 interface for resource/device querying and monitoring, resource virtualization and orchestration.
• - Interaction with Comp CF/SF network for remote management of resources/services, management of computing transport and signaling/data transmission for CPIUP traffic via interfaces 1 or 14.

5.1.2.1 Registration/notification for resources via interface Compl, interface Comp2 and interface 1.

3 FIG. 3 shows an example message sequence diagram 300 for a computing resource logon/notification and event/status between the application and the Comp CSF, the Comp CSF to the operating system/middleware, or the Comp CSF and the Comp CF in accordance with at least one of the present disclosure described aspect. The requester (on the left) can subscribe to events or request status from the responder (on the right), as in 3 shown.

• - über die Schnittstelle Comp1 kann das Ereignis/der Status für die Anwendung, die den Comp CSF abonnieren soll, die Verfügbarkeit der lokalen/entfernten Ressource, den Status der Rechenaufgabe usw. umfassen
• - über die Schnittstelle Comp2 kann das Ereignis/der Status für das Comp CSF, das das BS/Middleware abonnieren soll, die Verfügbarkeit der lokalen Ressource, die Fähigkeiten der Kommunikationseinheit und den Status wie die Verfügbarkeit von drahtgebundenen/drahtlosen Verbindungen usw. umfassen
• - über die Schnittstelle 1 kann das Ereignis/der Status Informationen über die Rechenressource enthalten, z.B. Anzahl/Typ der Grafikverarbeitungseinheiten (GPUs), Zentraleinheiten (CPUs), Speicher usw.; der Status kann Verfügbarkeit, Nutzungsprozentsatz, Anzahl, Kapazität, Fähigkeit (z.B. Unterstützung von Open Computing Language, CUDA Computing Language usw.) umfassen.

The application or the COMP CSF can send a request to inform the Comp CSF or the BS+Middleware or the Comp CF about special events and the status of the local/remote computing resources and the network or the Sta to subscribe to the arithmetic task. The request can include the event/status subscribed to, the notification trigger conditions, a time interval, and so on. Across different interfaces, the event/status information may vary and include:

• - via the Comp1 interface, the event/status for the application to subscribe to the Comp CSF can include the availability of the local/remote resource, the status of the computational task, etc
• - via the Comp2 interface, the event/status for the Comp CSF to which the BS/middleware is to subscribe may include local resource availability, communication unit capabilities and status such as wired/wireless connection availability, etc
• - via the interface 1, the event/status can contain information about the computing resource, eg number/type of graphics processing units (GPUs), central processing units (CPUs), memory, etc.; status may include availability, usage percentage, count, capacity, capability (e.g. Open Computing Language support, CUDA Computing Language support, etc.).

The Comp CSF IBSIMiddleware/Comp CF can respond to the subscription request and indicate whether the subscription was successful or not. If successful, the application can receive a notification when the corresponding event/status occurs. The notification can be periodic or event driven based on the subscription request.

5.1.2.1 Procedures via the Comp1 interface

5.1.2.1.1 Registration

4 FIG. 4 shows an exemplary message sequence diagram 400 for dynamic workload offloading registration, in accordance with at least one aspect described in the present disclosure. The application can contact the Comp CSF via the in 4 Comp1 interface shown for dynamic workload offloading.

The application can register with the Comp CSF for dynamic workload offloading by calling the provided APIs. The registration message may include, among other things, an application manifest, e.g., an application identifier (ID), an application type, valid conditions for registration, e.g., the availability of certain connection types, the validity period of the registration, applied credentials and identifiers for the service, e.g., data network name (DNN), if used for cellular access, port number, fully qualified domain name (FQDN), unified resource identifier (URI), default settings for computing tasks in terms of latency and throughput, preferences for workload shifting, e.g., preference for using network resources; Offloading protocols like RPC, gRPC, OpenCL, MPI, Vulkan, etc.

The Comp CSF can accept or reject the registration request based on the local and network side information. A rejection reason can be related to application types, availability of network resources, and status of wireless connections.

5.1.2.1.1 Request for Resources

• - eine virtualisierte Rechenressource mit Angaben zur Ressourcenmenge wie CPU-Anzahl, Speichergröße, Rechenfähigkeiten wie Beschleunigung;
• - Mikrodienste wie künstliche Intelligenz (AI)/ maschinelles Lernen (ML), Datenbank, Authentifizierung, Autorisierung und Abrechnung (AAA); und
• - spezielle Geräte wie GPU oder Beschleuniger.

5 FIG. 5 shows an example message sequence diagram 500 for a resource request/response between the application and the Comp CSF, in accordance with at least one aspect described in the present disclosure. The application can request resources for computing tasks from the Comp CSF via the Comp1 interface with different granularity, as in 5 shown. The requirement may include:

• - a virtualized computing resource with resource amount details such as CPU count, memory size, computing capabilities such as acceleration;
• - Microservices such as artificial intelligence (AI)/machine learning (ML), database, authentication, authorization and accounting (AAA); and
• - special devices such as GPU or accelerators.

The application can send a resource request to the Comp CSF by calling the APIs. This request may include, but is not limited to, an application type (eg, AI; computational task type) that terminates and frees the resource, or a deployment that may be online until requested to suspend; a source for the math problem (eg, a Uniform Resource Locator (URL) pointing to the image of the program); a resource setting such as min/max number of CPUs, memory; Placement settings such as special requirements for GPU or FPGA (Field Programmable Gate Array) acceleration; Resource preference rules like local resource preferences, performance settings like a time budget for the Completing execution and returning results.

the Comp CSF can make a relief decision locally based on the request, or send the appropriate information to the network and receive a decision from the network's Comp CF. The Comp CSF can send a response to the application stating whether the request was accepted or denied, the reason for the denial, and other available resource options.

5.1.2.1 Procedures via the Comp2 interface

5.1.2.1.1 Local Resource Discovery

6 FIG. 6 illustrates an example message sequence diagram 600 for requesting OS/middleware local resource information, in accordance with at least one aspect described in the present disclosure. The Comp CSF can request the information about the local computing resources from the operating system/middleware, as in 6 shown.

The Comp CSF can request the information about the local computer resources or specific types of computer resources. The request message may include information about the computing resources, such as a list of GPUs, CPUs, memory, etc. and their status, such as availability, utilization level, number, capacity, capability (e.g., OpenCL, CUDA support), etc. The operating system/ the middleware can send the Comp CSF a response to the requested information.

5.1.2.1.2 Request for a local resource

7 FIG. 7 illustrates an example message sequence diagram 700 for requesting local resources from the Comp CSF to BS/middleware, in accordance with at least one aspect described in the present disclosure. The Comp CSF can request local resources to perform a locally executed computational task, as in 7 shown.

The Comp CSF can request a computational task to be performed on local computational resources. The request can include, among other things, the resources for the computing task (e.g. the number of CPUs, memory, GPUs, etc.) as well as information about the computing task (e.g. the location where the program is located, the performance settings (e.g. the time budget for the execution) and the return results. Upon receiving the request, the operating system/middleware may perform resource isolation, e.g., by creating a container/process for the computational task with dedicated CPU/GPU and/or other accelerator resources. The operating system/middleware may send a response to the Comp CSF about the results of the local resource request, e.g. whether it was successful or not.

5.1.2.1.1 Request for control reports or data for remote data processing

8th 8 shows an example message sequence diagram 800 for requesting remote control messages or data, in accordance with at least one aspect described in the present disclosure. The Comp CSF can request a remote control message or data to the operating system/middleware to initiate appropriate procedures in the modem via the Comp2 interface.

The Comp CSF can generate a request for a remote control message or data. The request for remote control messages or data can be processed by the operating system/middleware in the form of various protocols or directly by the communication entity such as a network interface card (NIC) bypassing the operating system/kernel. For example, the Comp CSF can generate an RPC message itself or invoke the middleware to generate the RPC messages using the APIs. This message may indicate the type of request (e.g. computational control or data) and may contain any protocol data unit (PDU) to be sent to the Comp CF/SF.

The operating system/middleware can call the functions in the communication unit to send the message to the network side. When the operating system is bypassed, the Comp CSF can directly call existing APIs such as the Open Fabric Enterprise Distribution (OFED) defined Verbs API for direct remote memory access (RDMA) and then call the functions in the communication unit.

Once the response comes back from the network side, the response can follow the symmetric path to get to the Comp CSF.

5.1.2.1 Processes via interface 1

5.1.2.1.1 Registration

The Comp CSF can register with the Comp CF for compute offloading, which differs from service registration with the AMF as with some technologies. This procedure can be optional if the computing identities and the context of the UE are the same as those for the communication.

9 FIG. 9 illustrates an example message sequence diagram 900 for registering the Comp CSF with the Comp CF for dynamic compute offloading, in accordance with at least one aspect described in the present disclosure.

The Comp CSF can send a registration request to the Comp CF to register the compute offload service. The registration request may include the identity to be used for the compute offload, the requested compute offload type such as microservice or specific device, the application IDs, specific settings such as runtime, and so on.

The Comp CF may send a response to the registration request to indicate the status of the registration and other information such as the protocols supported for compute offload and specific instructions for using the service.

5.1.2.1.1 Discovery of computing resources/services

The Comp CSF can send a resource/service discovery request to the Comp CF to get the status of the network side computing resources. 10 10 illustrates an example message sequence diagram 1000 for the Comp CSF to request the Comp CF to determine network-side computing resources, in accordance with at least one aspect described in the present disclosure.

The Comp CSF can request a resource discovery from the Comp CF to obtain information about network computing resources. The request can contain detection filters, such as the criteria or the type of resource to be detected. For example, the Comp CSF can request information about the total available resources, the maximum allowed capacity for the UE, the available special functions, etc. The Comp CSF can also request information about the microservices on the network side, including information/identifiers about specific microservice instances like application type, application name, version or the criteria for microservices like database, AI/ML, video transcoding etc. The Comp CSF can alternatively ask the local resources - and send application status to Comp CF to exchange capability.

The Comp CF can send a response to the Comp CSF about the network resource information as specified in the request as a detection filter. The response can include the available resource and also information on how to access the resource. For example, the identifier of a Comp SF can be returned in the message to indicate the access point of a microservice. The information can be used to make decisions about outsourcing a computing task. The Comp CF can use the capabilities (e.g. the location of the UE) in selecting the resource in the discovery response. The discovery response may also include parameters for the resource's service level, such as service area, service entitlement level (e.g., free, freemium, trial, etc.), or other service level characteristics.

Alternatively, the detection of computing resources can also take place during registration (static), since the availability of the resources is specified during registration.

5.1.2.1.1 Request for computing resources

The Comp CSF can request the network computing resource to perform a computing task on the network computing resource. This method can use existing technologies such as RPC. The Comp CSF may support corresponding software APIs.

11 FIG. 1 shows an example message sequence diagram 1 100 for the Comp CSF to the Comp CF requesting network resources, in accordance with at least one aspect described in the present disclosure.

After either the Comp CSF or the Comp CF has made a decision as to whether a computational task should be performed using remote resources, the Comp CSF may request remote resources. The request may include, but is not limited to: the computational manifest such as number/type of CPU, GPU, accelerators, memory, occupancy time, etc.; the description of the calculation such as metadata or specification of the calculation including always online (deployment) or job (execution for a specified time until completion), image URL, whether compilation is required, runtime environment of the calculation; performance level settings such as latency and throughput; Resource usage agreement such as terms of cancellation, restart, postponement, etc.

Based on the settings of the math problem, the Comp CF can send a response indicating whether it accepts the math problem, the conditions for the problem, and information about how the problem can be accessed. The Comp CF can decide to reject the computational task and instruct the Comp CSF to execute the computational task locally or specify the conditions of when/how the computational resources should be requested again.

5.1.2 The Computing Resource Function (Comp RF) is based on dynamic scaling of computing power between devices and networks.

12 illustrates the interaction between the elements of the Comp RF. 12 illustrates the interaction between the elements of the Comp RF according to at least one aspect described in the present disclosure. In 12 the interaction of different elements of the Comp RF is shown. The Comp-RF can comprise a client part Comp-RF-C and a server part Comp-RF-S. The Comp-RF may have no impact on the application (e.g. the application can continue to use the existing interfaces to trigger data processing offloading). However, via the communication unit, the Comp RF can retrieve information about remote devices and present the remote resource as if the Comp RF were local to the operating system/middleware. The operating system/middleware can schedule computational tasks using both remote and local resources. The Comp RF-C on the UE side must not interface with the Comp CF/SF on the network side.

Note: The Comp RF-S is in 12 shown as part of the communication unit, but it can also be part of the operating system/middleware or stand alone. When the UE is connected to computing devices, the cellular network can offload computing tasks to a remote device by implementing the Comp RF-S on the UE side and the Comp RF-C on the cellular network side.

• - Die Funktionen der Compute-CP des Comp RF können unter anderem die Ressourcenermittlung, die Verwaltung des Rechentransports und die Verwaltung der Compute-Ressourcen (z.B. Zeitplanung) umfassen. Der Comp RF-C kann entfernte Rechenressourcen dynamisch über die Luftschnittstelle ermitteln und entfernte Ressourcen als lokale Ressourcen für Betriebssysteme/Middleware virtualisieren, um die Auslagerung von Arbeit über die Schnittstelle Comp3 zu unterstützen;
• - die Funktionalitäten des Compute UP des Comp RF können die Lieferung von Comp UP-Daten und -Kontext sowie die Planung von Echtzeit-Rechenressourcen umfassen; und
• - die Entscheidung über die Verlagerung der Rechenleistung kann von der BS/Middleware oder der Comp RF-C/Middleware für den BS/Kernel-Bypass auf der Seite des UE oder der Comp RF-S auf der Netzwerkseite getroffen werden.

The Comp RF can be viewed as an embedded function of the air interface, and the air interface can expose both network communication and computing capabilities. The Comp RF-C/S can contain both Compute CP and UP functions and serve as a Compute CP/UP endpoint on the client/server side, where the following applies

• - The functions of the compute CP of the Comp RF can include, among other things, the determination of resources, the administration of the computing transport and the administration of the compute resources (e.g. scheduling). The Comp RF-C can dynamically discover remote computing resources over the air interface and virtualize remote resources as local OS/middleware resources to support offloading of work over the Comp3 interface;
• - Compute UP functionalities of the Comp RF may include delivery of Comp UP data and context, and scheduling of real-time computational resources; and
• - the decision on the shifting of the computing power can be made by the BS/Middleware or the Comp RF-C/Middleware for the BS/Kernel bypass on the UE side or the Comp RF-S on the network side.

5.1.3.1 Procedure via the Comp3 interface

5.1.2.1.1 Request/response to determine computing resources

The operating system/middleware can request the Comp RF-C to determine the compute resources on the Comp RF-S side. The request may cause the modem to send a request for Layer 2 (L2)/Layer 3 (L3) computing capacity over interfaces 2'/3'.

13 13 shows an exemplary message sequence diagram 1300 for a resource discovery request to the Comp RF-C within the UE modem to offload computational work, in accordance with at least one aspect described in the present disclosure.

The operating system or middleware can send a request to the Comp RF-C for remote discovery of computing resources. The request can specify the settings for the computing device, such as the FPGA or GPU, and possible supported libraries. The process can be done through the Peripheral Component Interconnect Express (PCIe) enumeration process when the Comp RF-C or modem is attached to a PCIe bus.

If the Comp RF-C has the information (e.g. because it has previously discovered it), it can send it to the operating system/middleware. Otherwise the Comp RF-C can trigger corresponding procedures/functions in the air interface.

The Comp RF-C can display the results of the discovered devices with the associated information information such as numbers, type, supported libraries, etc.

5.1.2.1.1 Request for remote access to resources/devices

The OS/middleware can send a resource/device access request to the Comp RF-C to access the remote network devices to offload data processing. The request may cause the modem to send the L2/L3 request for computing resources over the 2'/3' interface.

14 14 illustrates an example message sequence diagram 1400 for a request for computational resources to Comp RF-C in the modem for offloading computational work, in accordance with at least one aspect described in the present disclosure/invention.

The operating system/middleware can request computing resources to the Comp RF-C to offload a computing task. The request may contain the metadata describing the settings of the computing device, such as number, type, supported libraries, etc., the metadata describing the computing task, such as AI/ML, etc., and the associated protocols for computing.

The Comp RF-C can trigger appropriate functions for requesting computing resources in the air interface. After a task has been accepted by the network, the Comp RF-C can generate status information for the task, e.g. assign a task ID. The Comp RF-C can reject the request based on local information about the computing resources or the connection status.

The Comp RF-C can respond after an offloading decision has been made either locally or remotely on the request. When a request is accepted, the response can include information about the task, such as the task identifier, access to computing resources, and rules for using the device. If the request is rejected, a reason can be given.

5.1.2.1.1 Calculation data

The operating system/middleware can send computing data to Comp RF-C to access the network devices on the remote side and offload computing power.

15 15 illustrates an example message sequence diagram 1500 of the operating system/middleware sending computational data to Comp RF-C for delivery to Comp RF-S for computational offloading, in accordance with at least one aspect described in the present disclosure.

The operating system/middleware may request to send computational data to Comp RF-C with the information about the data and may include a task ID to indicate the computational task to which the data is associated and metadata about it, how the calculation data is to be handled (e.g. specifying the file type).

The Comp RF-C can indicate in the response whether the calculation data was sent successfully or not.

5.1.2.1 The Comp RF may include a built-in modem for infrastructure-level data processing offloading.

16 12 illustrates the Comp RF with built-in modem for offloading of computing work in accordance with at least one aspect described in the present disclosure. On the UE side, the UE modem can include the Comp RF-C to interface with the Comp RF-S in the network modem to get the information about remote computing devices, as in 16 shown. The Comp RF-S on the network side can obtain the device information through methods such as PCIe enumeration. On the network side, the modem can be connected directly to the data processing devices via interfaces such as PCIe, Quad Small Form-factor Pluggable (QSFP) transceiver, GPUDirect, etc. The customization layer may include the operating system/middleware to support a uniform interface to the application.

The deployment options may include connecting the computing device to the CU or connecting the computing device to the DU.

17 illustrates the provisioning options for computing devices connected to the CU/DU according to at least one aspect described in the present disclosure.

5.1.3.3 Procedure via interfaces 2'/3'.

The detection of computing devices and the request for computing resources can be integrated into the functions of the air interface.

5.1.2.1.1 Recognition of the computing device.

The UE can request the identification of the network processing units connected to the CU/DU.

18 18 illustrates an example message sequence diagram 1800 of a computational capability query, in accordance with at least one aspect described in the present disclosure.

• - eine RRC-Nachricht (Funkressourcensteuerung - Radio Resource Control) an die CU, um die Geräte zu ermitteln, die sowohl mit der CU als auch mit dem DU oder nur mit dem DU verbunden sind. Zu diesem Zweck kann eine neue RRC-Nachricht definiert werden; und
• - eine vordefinierte Mediumzugriffssteuerung(MAC) CE-artige, um die mit der DU verbundenen Geräte zu ermitteln (oder die Ermittlung zu aktivieren).

The UE may send the computing capacity exchange request to discover the devices connected to the CU/DU in the cellular network. This request may differ from the determination via interfaces 2/3 described above and may be specified in the message. For example, the discovery request may specify service-level discovery and the discovery message via interfaces 2'/3' direct infrastructure-level offload. The request may include the computing manifest, such as computing device settings, number, type, and libraries, and does not necessarily result in the establishment of compute-specific data resource blocks (DRBs). This message can contain the following options:

• - an RRC (Radio Resource Control) message to the CU to identify the devices connected to both the CU and the DU or just the DU. A new RRC message can be defined for this purpose; and
• - a predefined Medium Access Control (MAC) CE-like to discover (or enable discovery) the devices connected to the DU.

• - eine maximale Anfragerate;
• - eine maximale Antwortzeit;
• - eine prozentuale Verfügbarkeit der Geräte, die zur Bearbeitung der Anfragen zur Verfügung stehen;
• - eine maximal verfügbare Rechenressource;
• - eine maximale grafische Rechenleistung; und
• - verfügbaren Arbeitsspeicher, Speicherplatz und Verbindungsbandbreite.

The CU/entity may respond based on the request with information about the computing devices that support direct data processing offload from the CU/entity. The data processing device information may include the following information in the response:

• - a maximum request rate;
• - a maximum response time;
• - a percentage of availability of the devices available to process the requests;
• - a maximum available computing resource;
• - a maximum graphic processing power; and
• - available memory, disk space and connection bandwidth.

Alternatively, the CU/DU can transmit the information about the availability/capability of the computing resources via radio. This transmitted information can include the type, name, etc. of the computing resource.

5.1.2.1.1 Request for a remote computing resource

The UE can request access to a remote resource on the network side. This may, but does not have to, result in a compute data transport being set up.

19 19 shows an example message sequence diagram 1900 of a request for a remote computing resource, in accordance with at least one aspect described in the present disclosure.

• - eine neue RRC-Nachricht an die ZB, um den Zugriff auf die mit der ZB/Einheit verbundenen Geräte oder auf beide anzufordern; und
• - eine definierte MAC-CE-artige Anfrage für den Zugriff auf die mit dem DU verbundenen Geräte.

The UE may send a remote resource request to the CU/DU to request offloading of computing power to network devices. The request may contain the reference to the outsourcing of computing power at the infrastructure level, the computing manifest and the metadata of the computing task. If the CU/DU accepts the request, a radio bearer can be established for the transmission of the subsequent calculation data. The radio transmission path for calculations can also be established during the first request to set up the communication link based on an indication of the UE (to support calculations). The CU/entity can also reject the request. Options for this message may include:

• - a new RRC message to the AP to request access to the devices connected to the AP/entity or to both; and
• - a defined MAC-CE-style request for access to the devices connected to the DU.

If the UE is in the RRC_CONNECTED state, a new RRC message can be defined to send/return the request for computing resources. If the computing resource is available at the DU, the DU can decode the RRC message containing the request, or the AP can decode it (which is simpler) and interact with the DU to provide the response and appropriate access; the computing resource is available at the CB, the CB can decode the message as before. The distinction of where the resource is available may need to be announced in advance.

If the UE is in the RRC_INACTIVE state, to shorten the transition time for the request for the arithmetic res sources the computation request message to be piggybacked onto the two-stage RACH-MSG A. The Release (Rel)-17 optimization of the "Small Data Transmission" can still be used to send the compute resource request message (as part of msg3); and the CU/DU can respond to a remote resource request. If accepted, it may contain information on how to use the assigned device identifiers or how the device can be accessed. If the request is denied, a reason can be included in the response and sent to UE.

The 20-21 illustrate various systems, devices, and components that may implement aspects of the disclosed aspects.

20 Figure 2 illustrates a network 2000 according to various aspects. The network 2000 can operate in a manner consistent with 5G/New Radio (NR) systems. However, the example aspects are not limited in this regard, and the aspects described may also apply to other networks that benefit from the principles described herein, such as future Third Generation Partnership Project (3GPP) systems or the like.

The network 2000 may include a UE 2002, which may include any mobile or non-mobile computing device configured to communicate with a RAN 2004 over an over-the-air connection. The UE 2002 may be, but is not limited to, a smartphone, tablet computer, portable computing device, desktop computer, laptop computer, in-vehicle infotainment or entertainment device, instrument cluster, head-up display, Onboard diagnostic device, handle a mobile dashtop device or mobile terminal, electronic engine management system, electronic/engine control unit, electronic/engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked device, machine-type communication device, machine-to-machine ( M2M) or device-to-device (D2D) device, Internet of Things (IoT) device, etc.

The network 2000 may include a plurality of UEs directly connected to each other via a sidelink interface. The UEs may be M2M/D2D devices communicating via physical sidelink channels such as Physical Sidelink Broadcast Channel (PSBCH), Physical Sidelink Downlink Channel (PSDCH), Physical Sidelink - Physical Sidelink Shared Channel (PSSCH), Physical Sidelink Control Channel (PSCCH), Physical Sidelink Feedback Channel (PSFCH), etc.

The UE 2002 can additionally communicate with an application protocol (AP) 2006 over a wireless connection. The AP 2006 can manage a WLAN (Wireless Local Area Network) connection, which can be used to offload some/all network traffic from the RAN 2004. The connection between the UE 2002 and the AP 2006 can be with any IEEE 802.11 (Institute of Electrical and Electronics Engineering) protocol, where the AP 2006 can be a Wi-Fi® (Wireless Fidelity) router. The UE 2002, the RAN 2004 and the AP 2006 can use cell-WLAN aggregation. In cellular WLAN aggregation, the UE 2002 can be configured by the RAN 2004 to use both cellular radio resources and WLAN resources.

The RAN 2004 may contain one or more access nodes, for example the access network (AN) 2008. AN 2008 may terminate air interface protocols for the UE 2002 by implementing access layer protocols including RRC, Packet Data Convergence Protocol (PDCP), Radio Link Control (RLC), MAC and Layer 1 ( L1) protocols. In this way, AN 2008 can enable data/voice connectivity between CN 2020 and UE 2002 . The AN 2008 may be implemented in a separate device or as one or more software units running on server computers as part of a virtual network that may be referred to as a Cloud Radio Access Network (CRAN) or virtual baseband unit pool. The AN 2008 can be referred to as a base station (BS), gNodeB (gNB), RAN node, ng-evolved NodeB (eNB), NodeB, Road Side Unit (RSU), Transceiver Point (TRxP), Transceiver Point (TRP), etc. The AN 2008 may be a macrocell base station or a low power base station providing femtocells, picocells or similar cells with smaller coverage areas, lower user capacity or higher bandwidth compared to macrocells.

The RAN 2004 may include a plurality of ANs that may be coupled together via an Xn interface (if the RAN 2004 is a 5G RAN). Via the Xn interfaces, which can be divided into control/user plane interfaces, the ANs can exchange information related to handover, data/context transfer, mobility, load management, interference coordination, etc.

The ANs of the RAN 2004 may each manage one or more cells, cell groups, component carriers, etc. to provide the UE 2002 with an air interface for network access. The UE 2002 may be simultaneously connected to a plurality of cells provided by the same or different RAN 2004 ANs. For example, the UE 2002 and the RAN 2004 may use carrier aggregation to allow the UE 2002 to connect to a plurality of component carriers, each corresponding to a primary cell (Pcell) or secondary cell (Scell). In dual-connectivity scenarios, a first AN may be a master node providing a master cell group (MCG) and a second AN may be a secondary node providing a secondary cell group (SCG). The first/second AN can be any combination of gNB, ng-eNB, etc.

The RAN 2004 can provide the air interface over licensed spectrum or unlicensed spectrum. For operation in the unlicensed spectrum, the nodes may use License-Aided Access (LAA), Enhanced LAA (eLAA) and/or feLAA mechanisms based on Carrier Aggregation (CA) technology with PC cells/cells. Prior to accessing the unlicensed spectrum, the nodes may perform media/carrier acquisition operations, e.g., based on an LBT (listen-to-speak) protocol.

In vehicle-to-everything (V2X) scenarios, the UE 2002 or the AN 2008 may be or act as an RSU, which may refer to any traffic infrastructure entity used for V2X communication. An RSU can be implemented in or by a suitable AN or a stationary (or relatively stationary) UE. An RSU implemented in or by a UE may be referred to as a "UE-type RSU"; a gNB can be referred to as "gNBtype RSU"; and the same. In one example, an RSU is a computing unit coupled to radio frequency circuitry located at the side of the road that provides connectivity to the UEs of passing vehicles. The RSU may also contain internal data storage circuitry to store intersection geometry, traffic statistics, media, and applications/software to capture and control ongoing vehicle and pedestrian traffic. The RSU can provide very low latency communication settings for high-speed events such as crash avoidance, traffic alerts, and the like. Additionally or alternatively, the RSU may provide other cellular/WLAN communication services. The components of the RSU may be housed in a weatherproof enclosure suitable for outdoor installation and may include a network interface controller to provide a wired (e.g. Ethernet) connection to a traffic signal controller or backhaul network.

The RAN 2004 may be a Next Generation (NG) RAN 2014 with gNBs, e.g. gNB 2016, or ng-eNBs, e.g. ng-eNB 2018. The gNB 2016 can connect to 5G-enabled UEs via a 5G NR interface. The gNB 2016 may be connected to a 5G core via an NG interface, which may include an N2 interface or an N3 interface. The ng-eNB 2018 can also be connected to the 5G core via an NG interface. The gNB 2016 and the ng-eNB 2018 can be connected to each other via an Xn interface.

The NG interface can be divided into two parts: an NG-U interface (NG-U) that carries traffic data between the nodes of the NG-RAN 2014 and a UPF 2048 (e.g. N3 interface), and an NG-C interface (NG-C), which is a signaling interface between the nodes of the NG-RAN 2014 and an AMF 2044 (e.g. N2 interface).

The NG-RAN 2014 can provide a 5G NR air interface with the following features: variable subcarrier spacing (SCS); CP-Orthogonal Frequency Division Multiplexing (OFDM) for downlink (DL), CP-OFDM and DFT-s-OFDM for uplink (UL); Polar, repeat, simplex and Reed-Muller codes for control and LDPC for data. The 5G NR air interface can rely on the CSI reference signal (CSI-RS), downlink physical shared channel (PDSCH)/downlink physical control channel (PDCCH), and demodulation reference signal (DMRS). The 5G NR air interface shall not use a Cell Specification Reference Signal (CRS) but may use the Physical Broadcast Channel (PBCH) DMRS for PBCH demodulation, Phase Tracking Reference Signal (PTRS) for phase tracking for PDSCH and tracking -Use reference signal for time tracking. The 5G NR air-interface can operate in frequency ranges (FR) 1 covering bands below 6 gigahertz (GHz) or in FR2 bands covering bands from 24.25 GHz to 52.6 GHz. The 5G NR air interface may include a sync signal block (SSB), which is a portion of a downlink resource grid that includes a primary sync signal (PSS)/secondary sync signal (SSS)/PBCH.

The 5G NR air interface can use bandwidth shares (BWP) for different purposes. For example, BWP can be used for dynamic adjustment of the SCS. For example, the UE 2002 can be configured with multiple BWPs, with each BWP configuration having a different SCS. When a BWP change is indicated to the UE 2002, the SCS of the transmission is also changed. Another application example for BWP relates to energy saving. In particular, multiple BWPs can be configured for the UE 2002 with a different number of frequency resources (eg, Physical Resource Blocks (PRBs)) to support data transmission under different traffic load scenarios. A BWP containing a smaller number of PRBs can be used for data transmission with low traffic load while enabling power savings at the UE 2002 and in some cases the gNB 2016. A BWP with a larger number of PRBs can be used for higher traffic scenarios become.

The RAN 2004 is communicatively coupled to the CN 2020, which includes network elements for providing various functions in support of data and telecommunications services to customers/subscribers (e.g., users of the UE 2002). The components of the CN 2020 can be implemented in one physical node or in separate physical nodes. Network Function Virtualization (NFV) can be leveraged to virtualize some or all of the functions provided by the CN 2020's network elements onto physical compute/storage resources in servers, switches, etc. A logical instantiation of CN 2020 may be referred to as a network slice, and a logical instantiation of a portion of CN 2020 may be referred to as a network sub-slice.

The CN 2020 may be a 5G core network (5GC) 2040. The 5GC 2040 can include an authentication server function (AUSF) 2042, an AMF 2044, a session management function (SMF) 2046, a user plane function (UPF) 2048, a network slice selection function (NSSF) 2050, a NEF 2052, a network function (NF), a repository function (NRF) 2054, a PCF 2056, a unified data management (UDM) 2058, and an application function (AF) 2060, coupled together via interfaces (or "reference points") as shown. The functions of the elements of the 5GC 2040 can be briefly introduced as follows.

The AUSF 2042 may store data for authentication of the UE 2002 and perform authentication related functions. The AUSF 2042 can enable a common authentication framework for different types of access. In addition to communicating with other elements of the 5GC 2040 via reference points as shown, the AUSF 2042 may have a Nausf service based interface.

The AMF 2044 can enable other functions of the 5GC 2040 to communicate with the UE 2002 and the RAN 2004 and subscribe to notifications of mobility events related to the UE 2002. The AMF 2044 may be responsible for registration management (e.g. for UE 2002 registration), connection management, reachability management, mobility management, lawful interception of AMF-related events, and access authentication and authorization. The AMF 2044 can provide the transport of session management (SM) messages between the UE 2002 and the SMF 2046 and act as a transparent proxy for the forwarding of SM messages. AMF 2044 can also provide transport of SMS (Short Message Service) messages between UE 2002 and an SMS function (SMSF). AMF 2044 can interact with AUSF 2042 and UE 2002 to perform various security anchor and context management functions. In addition, the AMF 2044 may be an endpoint of a RAN-CP interface, which may include or be an N2 reference point between the RAN 2004 and the AMF 2044; and the AMF 2044 may be a non-access layer (NAS) (N1) signaling endpoint and perform NAS encryption and integrity protection. The AMF 2044 can also support NAS signaling with the UE 2002 via an N3 Interworking Function (IWF) interface.

The SMF 2046 may be responsible for SM (e.g. session establishment, tunnel management between UPF 2048 and AN 2008); Allocation and management of IP addresses for the UE (including optional authorization); selection and control of the UP function; Configure traffic routing at the UPF 2048 to route traffic to the correct destination; termination of interfaces for policy control functions; Controlling parts of policy enforcement, charging and QoS; lawful interception (for SM events and the interface to the Layer Indicator (LI) system); termination of SM parts of NAS messages; downlink data notification; initiation of AN specific SM information sent via AMF 2044 to AN 2008 via N2; and determining the session and service continuity (SSC) mode of a session. SM can refer to the management of a PDU session, and a PDU session or "session" can refer to a PDU Kon obtain a connectivity service that provides or enables the exchange of PDUs between the UE 2002 and the data network 2036.

The UPF 2048 can serve as an anchor point for intra-radio access technology (RAT) and inter-RAT mobility, as an external PDU session connection point to the data network 2036, and as a branch point to support multi-homed PDU sessions. The UPF 2048 can also perform packet routing and forwarding, perform packet inspection, enforce the user plane part of policy rules, legitimately intercept packets (UP collection), generate traffic usage reports, perform QoS treatment for a user plane (e.g. packet filtering, gating, UL/ DL rate enforcement), perform uplink traffic inspection (e.g. SDF to QoS flow mapping), perform transport level packet marking in uplink and downlink, and perform downlink packet buffering and downlink data notification triggering. The UPF 2048 may include an uplink classifier to help route traffic streams to a data network.

The NSSF 2050 can select a set of network slice entities that serve the UE 2002. The NSSF 2050 can also determine the allowed NSSAI and mapping to the subscribed information to support Single Network Leg Selection (S-NSSAI) if needed. The NSSF 2050 can also determine the AMF set to be used to serve the UE 2002 or a list of AMF candidates based on an appropriate configuration and possibly by querying the NRF 2054. The selection of a set of network slices Instances for the UE 2002 can be triggered by the AMF 2044 to which the UE 2002 is registered through interaction with the NSSF 2050, which can result in a change of AMF. The NSSF 2050 can interact with the AMF 2044 via an N22 reference point and can communicate with another NSSF in a visited network via an N31 reference point (not shown). In addition, the NSSF 2050 may have an Nnssf service-based interface.

The NEF 2052 can securely share services and capabilities provided by 3GPP network functions with third parties, internal exposure/re-exposure, AFs (e.g. AF 2060), edge computing or fog computing systems, etc. The NEF 2052 can authenticate, authorize, or throttle the AFs. The NEF 2052 can also translate information exchanged with the AF 2060 and information exchanged with internal network functions. For example, the NEF 2052 can translate between an AF service identifier and internal 5GC information. The NEF 2052 may also receive information from other Network Functions (NFs) based on the disclosed capabilities of other NFs. This information can be stored in the NEF 2052 as structured data or in a data storage NF using standardized interfaces. The stored information can then be shared by the NEF 2052 with other NFs and AFs or used for other purposes such as analytics. In addition, the NEF 2052 may have an Nnef service based interface.

The NRF 2054 may support service discovery functions, receiving NF discovery requests from NF entities, and forwarding the discovered NF entities' information to the NF entities. The NRF 2054 also maintains information about available NF entities and their supported services. As used herein, the terms "instantiate," "instantiation," and the like may refer to the creation of an instance, and an "instance" may refer to a tangible occurrence of an object that may occur, for example, during the execution of program code. In addition, the NRF 2054 may have the Nrf service-based interface.

The PCF 2056 can provide rules for the control plane functions to enforce, and it can also support a unified rule set to control network behavior. The PCF 2056 may also implement a front end to access subscription information relevant to policy decisions in a UDM 2058 Unified Data Repository (UDR). In addition to communicating with functions via reference points as shown, the PCF 2056 has an Npcf service based interface.

The UDM 2058 may process subscription-related information to support the handling of communication sessions by the network entities and may store UE 2002 subscription data. For example, the subscription data can be communicated between the UDM 2058 and the AMF 2044 via an N8 reference point. The UDM 2058 can comprise two parts, an application front end and a UDR. The UDR can store subscription data and policy data for the UDM 2058 and the PCF 2056 and/or structured data for exposure and application data (including packet flow descriptions (PFDs) for application recognition, application request information for multiple UEs 2002) for the NEF 2052. The interface based on the Nudr service can be issued by the UDR 221 to the Allow the UDM 2058, the PCF 2056 and the NEF 2052 to access a specific set of the stored data and to read, update (e.g. add, change), delete and subscribe to notifications of relevant data changes in the UDR. The UDM may contain a UDM Frontend (FE) that is responsible for credential processing, location management, subscription management, and so on. Several different frontends can serve the same user in different transactions. The UDM-FE accesses the subscription information stored in the UDR and performs authentication credential processing, user identification handling, access authorization, registration/mobility management, and subscription management. In addition to communicating with other NFs via reference points as shown, the UDM 2058 may have the Nudm service-based interface.

The AF 2060 can enable the application's influence on traffic routing, provide access to the NEF, and interact with the policy framework for policy control.

The 5GC 2040 can enable edge computing by selecting operator/third party services to be geographically close to a point where the UE 2002 is connected to the network. This can reduce latency and load on the network. To provide edge computing implementations, the 5GC 2040 may select a UPF 2048 in the vicinity of the UE 2002 and route traffic from the UPF 2048 to the data network 2036 over the N6 interface. This can be done based on the UE subscription data, the UE location and the information provided by the AF 2060. In this way, the AF 2060 can affect UPF (re)selection and traffic routing. If AF 2060 is considered a trusted entity, the network operator can allow AF 2060 to interact directly with the relevant NFs. In addition, the AF 2060 may have a Naf service based interface.

Data network 2036 may represent various network operator services, Internet access, or third-party services that may be provided by one or more servers, such as an application/content server 2038.

21 FIG. 2 schematically shows a wireless network 2100 according to various aspects. The wireless network 2100 may include a UE 2102 in wireless communication with an AN 2104 . The UE 2102 and AN 2104 may be similar and substantially interchangeable with like-named components described elsewhere herein.

The UE 2102 may be communicatively coupled to the AN 2104 via a link 2106 . Link 2106 is shown as an air interface to enable communicative coupling and may be compliant with cellular communication protocols such as a 5G NR protocol operating at millimeter wave (mmWave) or sub-6GHz frequencies.

The UE 2102 may include a host platform 2108 coupled to a modem platform 2110 . Host platform 2108 may include processing circuitry 2112 that may be coupled to protocol processing circuitry 2114 of modem platform 2110 . Processing circuitry 2112 may execute various applications for UE 2102 that generate/process application data. Processing circuitry 2112 may further implement one or more layer operations to transmit/receive application data to/from a data network. These layer operations may include transport (e.g., User Datagram Protocol (UDP)) and internet (e.g., IP) operations

Protocol processing circuitry 2114 may implement one or more of the layer operations to facilitate the transmission or receipt of data over link 2106 . For example, the layer operations implemented by the protocol processing circuitry 2114 may include MAC, RLC, PDCP, RRC, and NAS operations.

The modem platform 2110 may further include a digital baseband circuitry 2116 that may implement one or more layer operations that are “below” the layer operations performed by the processing circuitry 2114 in a network protocol stack. These operations may include, for example, physical layer (PHY) operations including one or more hybrid automatic repeat request (HARQ) and acknowledgment (ACK) functions, scrambling/descrambling, encoding/decoding, layer mapping/de-mapping, modulation symbol mapping , determination of received symbols/bit metric, precoding/decoding of multi-antenna ports, which may include one or more of the following functions: space-time, space-frequency or spatial coding, generation/detection of reference signals, generation and/or decoding of preamble sequences, generation/detection of synchronization sequences, blind decoding of control channel signals and other related functions.

Modem platform 2110 may further include transmit circuitry 2118, receive circuitry 2120, radio frequency (RF) circuitry 2122, and RF front end (RFFE) 2124, which may include or be connected to one or more antenna arrays 2126. Briefly, transmit circuitry 2118 may include a digital-to-analog converter, a mixer, intermediate frequency (IF) components, and so on. The receiving circuitry 2120 may include an analog-to-digital converter, mixer, IF components, etc.; RF circuitry 2122 may include a low noise amplifier, a power amplifier, power tracking components, etc.; the RFFE 2124 may include filters (e.g., surface/bulk acoustic wave filters), switches, antenna tuners, beamforming components (e.g., phased array antenna components), etc. The selection and arrangement of the components of the transmit circuitry 2118, receive circuitry 2120, RF circuitry 2122, RFFE 2124, and antenna panels 2126 (commonly referred to as "transmit/receive components") may depend on the details of a particular implementation, such as whether communication takes place in time division multiplex (TDM) or frequency division multiplex (FDM), in mmWave or sub-6 GHz frequencies, etc. The transmit/receive components can be arranged in several parallel transmit/receive chains, they can be on the same one or on different chips/modules, etc.

Processing circuitry 2114 may include one or more instances of control circuitry (not shown) to provide control functions for the transmit/receive components.

UE reception may be established through and through antenna arrays 2126, RFFE 2124, RF circuitry 2122, receiving circuitry 2120, digital baseband circuitry 2116, and protocol processing circuitry 2114. The antenna arrays 2126 may receive a transmission from the AN 2104 by receive beamforming signals received from a plurality of antennas/antenna elements of the one or more antenna arrays 2126 .

A UE transmission can be established from and via the protocol processing circuitry 2114, the digital baseband circuitry 2116, the transmission circuitry 2118, the RF circuitry 2122, the RFFE 2124 and the antenna arrays 2126. The transmit components of the UE 2104 may apply a spatial filter to the data to be transmitted to form a transmit beam emitted by the antenna elements of the antenna arrays 2126.

Similar to the UE 2102, the AN 2104 may include a host platform 2128 coupled to a modem platform 2130. Host platform 2128 may include processing circuitry 2132 coupled to protocol processing circuitry 2134 of modem platform 2130 . The modem platform may further include digital baseband circuitry 2136, transmit circuitry 2138, receive circuitry 2140, RF circuitry 2142, RFFE circuitry 2144, and antenna arrays 2146. The components of the AN 2104 may be similar to and substantially interchangeable with the like-named components of the UE 2102. In addition to performing data transmission/reception as described above, the AN 2108 components may perform various logical functions including, for example, radio network controller (RNC) functions such as management of radio bearers, dynamic management of uplink radio resources - and downlink and the scheduling of data packets.

22 12 is a block diagram depicting, in accordance with some example aspects, components capable of reading instructions from a machine-readable or computer-readable medium (eg, a non-transitory machine-readable storage medium) and performing one or more of the methods discussed herein. 22 In particular, Figure 12 shows a schematic representation of hardware resources 2200 including one or more processors (or processor cores) 2210, one or more storage devices 2220, and one or more communication resources 2230, each of which may be communicatively coupled via a bus 2240 or other interface circuit. In aspects where node virtualization (eg, NFV) is used, a hypervisor 2202 may execute to provide an execution environment for one or more network slices/sub-slices to utilize the hardware 2200 resources.

The processors 2210 may include a processor 2212 and a processor 2214, for example. The processors 2210 can be, for example, a CPU, a Reduced Instruction Set Computing (RISC) processor, a Complex Instruction Set Computing (CISC) processor, a GPU, a digital signal processor (DSP) such as a baseband processor, an application specific integrated circuit (ASIC), an FPGA, a radio frequency integrated circuit (RFIC = Radio Frequency Integrated Circuit), another processor (including those described herein), or any suitable combination thereof.

Storage/storage devices 2220 may include main memory, disk storage, or any combination thereof. The memory/storage devices 2220 may include any type of volatile, non-volatile, or semi-volatile memory, such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory , solid state memory, etc., but are not limited to these.

The communication resources 2230 may include link or network interface controllers, components, or other suitable devices to communicate with one or more peripheral devices 2204 or one or more databases 2206 or other network elements over a network 2208. Communication resources 2230 may include, for example, wired communication components (e.g., for interfacing via a universal serial bus (USB), Ethernet, etc.), cellular communication components, near field communication (NFC) components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components and other communication components.

Instructions 2250 may be software, a program, an application, an applet, an app, or other executable code for causing at least one of the processors 2210 to perform one or more of the methods described herein. Instructions 2250 may reside in whole or in part in at least one of processors 2210 (e.g., in the processor's cache memory), in memory devices 2220, or in any suitable combination thereof. Additionally, any portion of instructions 2250 may be transmitted to hardware resources 2200 from any combination of peripherals 2204 or databases 2206 . Accordingly, memory of processors 2210, storage devices 2220, peripheral devices 2204, and databases 2206 are examples of computer-readable and machine-readable media.

At least one of the components illustrated in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and/or methods as set forth in the example section below. For example, the baseband circuitry as described above in connection with one or more of the previous figures may be configured to operate according to one or more of the examples below. As a further example, the circuitry associated with a UE, base station, network element, etc., as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples listed below in the Examples section is working.

• - Anmelde-ZBenachrichtigungsprozeduren über die Schnittstelle Comp1, die Schnittstelle Comp2 und die Schnittstelle 1 können das Ereignis/den Status zur Anmeldung und die Regeln zum Erhalt einer Benachrichtigung enthalten
• - Prozeduren über die Schnittstelle Comp1
  • ◯ Registrierung, die Folgendes umfasst
    • ▪ Die anwendungsbezogenen Informationen
    • ▪die dienstbezogenen Informationen
    • ▪Identitäten für das Auslagern der Datenverarbeitung
  • ◯ Anfrage einer Ressource, die Folgendes umfasst
    • ▪Das Datenverarbeitungsmanifest
    • ▪Die ressourcenbezogenen Informationen mit unterschiedlicher Granularität
    • ▪Die Metadaten für die Berechnungsaufgabe
• - Prozeduren über die Schnittstelle Comp2
  • ◯ Lokale Ressourcenermittlung, die Folgendes umfasst
    • ▪ Computing-Manifest
    • ▪ Bedingungen für lokale Ressourcen
  • ◯ Anfrage einer lokalen Ressource, die Folgendes enthält
    • ▪ Bedingungen für die angeforderte Ressource
    • ▪ Informationen über die Datenverarbeitungsaufgabe
  • ◯ Anforderung von Steuermeldungen oder Daten für entfernte Datenverarbeitung, die aufgenommen werden sollen
    • ▪ Angabe der Kontrollnachricht oder -daten
    • ▪ Angabe des Nachrichtenprotokolls, z.B. RPC

Example 1 may include a Comp CSF for calculating swapping with dynamic resource availability that includes one or more of the following:

• - Subscription/Notification procedures via interface Comp1, interface Comp2 and interface 1 can contain the event/status to subscribe and the rules to receive notification
• - Procedures via the Comp1 interface
  • ◯ Registration, which includes the following
    • ▪ The application-related information
    • ▪the service-related information
    • ▪Identities for data processing outsourcing
  • ◯ Request for a resource that includes the following
    • ▪The data processing manifest
    • ▪The resource-related information with different granularity
    • ▪The metadata for the calculation task
• - Procedures via the Comp2 interface
  • ◯ Local resource discovery, which includes the following
    • ▪ Computing Manifesto
    • ▪ Local resource conditions
  • ◯ Request for a local resource containing the following
    • ▪ Conditions for the requested resource
    • ▪ Information about the data processing task
  • ◯ Request for tax returns or data for remote processing to be included
    • ▪ Specification of the control message or data
    • ▪ Specification of the message protocol, eg RPC

• - Prozeduren über Comp3
  • ◯ Erkennung von Rechenressourcen, einschließlich
    • ▪ Die Einstellungen für das Rechengerät
  • ◯ Anforderung einer entfernten Ressource, die Folgendes enthält
    • ▪ Die Einstellungen für das Rechengerät
    • ▪ Die Metadaten für die Berechnungsaufgabe
  • ◯ Anforderung von Rechendaten, die Folgendes enthält
    • ▪ Die Angabe für Berechnungsdaten mit einer Aufgaben-ID
    • ▪ Die Metadaten für die Bearbeitung der Aufgabe
• - Verfahren über 2'/3'
  • ◯ Rechengerät-Ermittlung, die Folgendes umfasst
    • ▪ Die Erkennungsanfrage/Antwort als RRC-Nachricht oder MAC CE
    • ▪ Die Bedingungen, welches Gerät gefunden werden soll
    • ▪ Die Angabe für das direkte Auslagern von Geräten
  • ◯ Anfrage nach einer entfernten Rechenressource, die Folgendes umfasst
    • ▪ Die Anfrage/Antwort als RRC-Nachricht oder Teil einer anfänglichen RACH-Nachricht oder eines MAC CE
    • ▪ Die Angabe zum direkten Geräte-Auslagern
    • ▪ Die Metadaten zur Rechenaufgabe
    • ▪ Das Rechenmanifest auf dem angeforderten Rechengerät

Example 2 may include a Comp RF for offloading computations with dynamic resource availability that includes one or more of the following methods:

• - Procedures via Comp3
  • ◯ Discovery of computing resources, including
    • ▪ The settings for the computing device
  • ◯ Request for a remote resource that contains the following
    • ▪ The settings for the computing device
    • ▪ The metadata for the calculation task
  • ◯ Request for computational data containing the following
    • ▪ The specification for calculation data with a task ID
    • ▪ The metadata for processing the task
• - Procedure over 2'/3'
  • ◯ Calculator discovery, which includes the following
    • ▪ The discovery request/response as an RRC message or MAC CE
    • ▪ The conditions for which device should be found
    • ▪ The specification for the direct paging of devices
  • ◯ Request for a remote computing resource that includes the following
    • ▪ The request/response as an RRC message or part of an initial RACH message or MAC CE
    • ▪ The specification for direct device swapping
    • ▪ The metadata for the arithmetic task
    • ▪ The computing manifest on the requested computing device

Example 3 may include a device of an NR UE, the device including an RF interface and one or more processors coupled to the RF interface, including a Comp CSF configured to: register for local computing resources to perform with an application of the NR UE; perform registration for local and remote computing resources with a BS/middleware of the NR UE; perform registration for remote computing resources with a Comp CF of a cellular network; determine the local and remote computing resources based on registration with the Application, the OS/Middleware and the Comp CF; receive a dynamic workload registration message from the application; receive a dynamic workload request message for a computational task from the application; and perform a dynamic workload for the computing task based on the determined local and remote computing resources and the dynamic workload requirement.

Example 4 may include the apparatus of Example 3, wherein the Comp CSF is configured to perform registration with the application over a Comp1 interface, and registering with the application includes receiving a subscription request message from the application that includes a subscribed event /indicates a subscribed status, a notification trigger condition, and a time interval of the local and remote computing resources.

Example 5 may include the device of Example 4, wherein the subscribed event/status indicates availability of local and remote computing resources and status of current computing tasks.

Example 6 may include the apparatus of Example 4, with the Comp CSF configured to register with the operating system/middleware via a Comp 2 interface and the subscribed event/status an availability of local computing resources, a or multiple capabilities of a communication unit and a wired and a wireless connection availability.

Example 7 may include the apparatus of Example 3, wherein the Comp CSF is configured to register with the Comp CF via at least one of interfaces 1 and 14, and the subscription and notification includes providing a remote subscription request message containing information about includes the remote computing resource, an availability status of the remote computing resource, and a usage status of the remote status.

Example 8 may include the device of any of Examples 3-7, wherein the dynamic workload offload registration message includes an application manifest with an application ID, an application type, a registration validity condition, a registration validity time, an applied credential and ID, a port number, an FQDN , a URI, a latency status, a throughput status, a workload offload setting and an offload log.

Example 9 may include the device of any of Examples 3-8, wherein the dynamic offload workload request message specifies a virtualized compute resource, a microservice, and a dedicated device.

Example 10 may include the apparatus of any of Examples 3-9, wherein performing dynamic workload offloading includes: providing the operating system/middleware with a local resource discovery message; receiving a local resource discovery response message from the operating system/middleware; and determining that the computational task is performed using the local resource based on the local resource discovery response message.

Example 11 may include the apparatus of any of Examples 3-9, wherein performing dynamic workload offloading comprises: providing the operating system/middleware with a remote resource detection message; receiving a remote resource discovery response message from the operating system/middleware; and determining that the computational task should be performed using the remote resource based on the remote resource discovery response message.

Example 12 may include the apparatus of any of Examples 3-9, wherein performing dynamic workload offloading includes: providing a remote resource discovery message to the Comp CF; receiving a remote resource discovery response message from the Comp CF; and determining that the computational task should be performed using the remote resource based on the remote resource discovery response message.

Example 13 may include the apparatus of any of Examples 3-12, with the Comp CSF positioned in a layer above the operating system/middleware.

Example 14 may include a device of an NR UE, the device including an RF interface and one or more processors coupled to the RF interface and including a Comp RF-C configured to: a first detection message of receive a BS/middleware of the NR UE; provide a second discovery message to a Comp RF-S of a cellular network; receive a first response message from the Comp RF-S; determine remote computing resources based on the first response message; provide a second response message to the BS/middleware based on the determined remote computing resources; receive a first compute task request message from the application; and perform dynamic workload offloading for the compute task with the Comp RF-S based on the determined remote compute resources and the first request message.

Example 15 may include the apparatus of Example 14, wherein performing the dynamic workload offloading comprises: providing a second request message for the compute task to the Comp RF-S based on the first request message; and receiving a computational resource message from the Comp RF-S, the computational resource message indicating that the Comp RF-S should perform the computational task.

Example 16 may include the apparatus of Example 15, wherein providing a second response message to the Comp RF-S includes triggering the modem to send an L2/L3 request for computing resources over at least one of an interface 2' and an interface 3' .

Example 17 may include the apparatus of any of Examples 14-16, wherein the Comp RF-C includes a compute CP function and a UP function and operates as an endpoint to perform resource discovery, compute transport management, and compute resource management, UP - Deliver data, provide UP context and plan computing resources.

Example 18 may include the apparatus of any of Examples 14-175, wherein providing the second discovery message comprises triggering a modem to send a second L2/L3 discovery message over at least one of interfaces 2' and 3'.

Example 19 may include the device of any of Examples 14-18, wherein the first discovery message specifies settings for a computing device and is transmitted according to a PCIe enumeration process.

Example 20 may include the apparatus of any of Examples 14-19 with the Comp RF-C and Comp RF-S located in a layer below the operating system/middleware.

Example 21 may include the device of any of Examples 14-20, wherein the Comp RF-C further configured to send a computing capability message to the Comp RF-S requesting the status of devices connected to a central unit and a distributed unit of the cellular network and containing a computing manifest specifying the settings of the computing devices, specifies a number of computing devices, a type of computing devices, and a library.

Example 22 may include the apparatus of example 21, wherein the computing capability message includes an RRC message identifying the computing devices connected to both the CU and the DU and a predefined MAC CE.

Example 23 may include an apparatus containing means for performing one or more elements of a method described in or associated with any of Examples 1-22 or any other method or process described herein.

Example 24 may include one or more non-transitory computer-readable media containing instructions for causing an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of a method described in any of Examples 1 -22 or related thereto, or any other method or process described herein.

Example 25 may include an apparatus having logic, modules, or circuitry for performing one or more elements of a method described in or related to any of Examples 1-22 or any other method or process described herein.

Example 26 may include a method, technique, or process as described in or associated with any of Examples 1-22, or portions thereof.

Example 27 may include an apparatus, including: one or more processors; and one or more computer-readable media containing instructions that, when executed by the one or more processors, cause the one or more processors to perform the method to perform the technique or process as described in, or associated with, any of Examples 1-22, or portions thereof.

Example 28 may include a signal as described in or associated with any of Examples 1-22, or portions or portions thereof.

Example 29 may include a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or relating to any of Examples 1-22 or parts thereof, or as otherwise described in of the present disclosure/invention.

Example 30 may include a signal encoded with data as described in or related to Examples 1-22, or portions thereof, or as otherwise described in the present disclosure/invention.

Example 31 may include a signal encoded with a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or associated with Examples 1-22, or portions or portions thereof, or otherwise described in the present disclosure/invention.

Example 32 may include an electromagnetic signal bearing computer-readable instructions, execution of the computer-readable instructions by one or more processors to cause the one or more processors to implement the method, technique, or process as in any of the examples 1-22 or parts thereof, or associated therewith.

Example 33 may include a computer program with instructions, execution of the program by a processing element to cause the processing element to perform the method, technique, or process described in or associated with any of Examples 1-22, or portions thereof.

Example 34 may include a signal on a wireless network as shown and described herein.

Example 35 may include a method of communicating over a wireless network as shown and described herein.

Example 36 may include a system for providing wireless communications as shown and described herein.

Example 37 may include a device for providing wireless communications as shown and described herein.

Any of the examples described above may be combined with any other example (or any combination of examples) unless expressly stated otherwise.

The foregoing description of one or more implementations has been presented for purposes of illustration and description, but is not intended to be exhaustive or to limit the scope of aspects to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of the various aspects.

Unless otherwise used herein, terms, definitions and abbreviations may be the same as terms, definitions and abbreviations defined in 3GPP TR 21.905 v16.0.0 (2019-06). For purposes of this document, the following abbreviations may apply to the examples and aspects discussed herein.

For purposes of this document, the following terms and definitions apply to the examples and aspects discussed herein.

The term "circuit" as used herein refers to hardware components such as electronic circuitry, logic circuitry, processor (common, dedicated or bank) and/or memory (common, dedicated or bank), application specific integrated circuit ( ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-performance PLD (HCPLD), a structured ASIC, or a programmable SoC ), digital signal processors (DSPs), etc. configured to provide the functionality described. The circuit may execute one or more software or firmware programs to provide at least some of the functions described. The term "circuit" may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) with the program code used to perform the functionality of that program code. The combination of hardware elements and program code can be referred to as a specific type of circuit.

As used herein, "processor circuitry" refers to, or is part of, circuitry capable of sequentially and automatically performing a series of arithmetic or logical operations or of recording, storing and/or transmitting digital data. The processing circuitry may include one or more processor cores for executing instructions and one or more memory structures for storing program and data information. The term "processing circuitry" can refer to one or more application processors, one or more baseband processors, a physical CPU, a single-core processor, a dual-core processor, a triple-core processor, a quad-core processor, and /or any other device capable of executing or otherwise operating computer-executable instructions such as program code, software modules and/or functional processes. The processing circuitry may include other hardware accelerators, which may be microprocessors, programmable processing devices, or the like. The one or more hardware accelerators may include, for example, computer vision (CV) and/or deep learning (DL) accelerators. The terms "application circuitry" and/or "baseband circuitry" may be considered synonymous with "processor circuitry" and referred to as such.

As used herein, “interface circuitry” refers to, or is part of or includes, circuitry that enables the exchange of information between two or more components or devices. The term "interface circuit" may refer to one or more hardware interfaces, e.g., buses, I/O interfaces, peripheral component interfaces, network interface cards, and/or the like.

The term "user equipment" or "UE" as used herein refers to a device with radio communication capabilities and may describe a remote user of network resources on a communication network. The term "user equipment" or "UE" can be used as a synonym for client, mobile device, mobile device, mobile terminal, user equipment, mobile unit, mobile station, mobile user, subscriber, user, remote station, access agent, user agent, receiver, radio, reconfigurable radio , reconfigurable mobile device, etc. are considered and referred to as such. Additionally, the term "user equipment" or "UE" can include any type of wireless/wired device or any computing device with an interface for wireless communication.

As used herein, the term "network element" refers to physical or virtualized devices and/or infrastructure used to provide network services for wired or wireless communications. The term "network element" can be used as a synonym for a networked computer, network hardware, network equipment, network node, router, switch, hub, bridge, wireless network controller, RAN device, RAN node, gateway, server, virtualized VNF, NFVI, and/or the like.

The term “computer system” as used herein refers to any type of interconnected electronic device, computing device, or component thereof. Additionally, the term "computer system" and/or "system" may refer to various components of a computer that are communicatively coupled to one another. Additionally, the term "computer system" and/or "system" may refer to multiple computing devices and/or multiple computing systems that are communicatively coupled and configured to share computing and/or network resources.

As used herein, the term “device,” “computing device,” or similar, refers to a computing device or computer system containing program code (e.g., software or firmware) specifically designed to provide a specific computing resource. A "Virtual Device" is an image of a virtual machine implemented by a hypervisor-equipped device that virtualizes or emulates a computing device or is otherwise dedicated to providing a specific computing resource.

The term "resource" as used herein refers to a physical or virtual device, a physical or virtual component within a computing environment and/or a physical or virtual component within a specific device, such as computing devices, mechanical devices, storage space, processor/CPU -Time, processor/CPU usage, processor and accelerator loads, hardware time or usage, electrical power, input/output operations, ports or network sockets, channel/link allocation, throughput, memory usage, network, database and applications , workload units and/or the like. A “hardware resource” may refer to compute, storage, and/or network resources provided by one or more physical hardware items. A "virtualized resource" may refer to compute, storage, and/or network resources provided by a virtualization infrastructure to an application, device, system, etc. The term "network resource" or "communications resource" may refer to resources that computing devices/systems can access over a communications network. The term "system resources" may refer to any type of shared service delivery device and may include computer and/or network resources. System resources can be viewed as a set of coherent functions, network data objects, or services accessible through a server, residing on a single host or multiple hosts and uniquely identifiable.

As used herein, the term "channel" refers to a tangible or intangible transmission medium used to transmit data or a data stream. The term "channel" may be synonymous and/or equivalent to "communication channel", "data communication channel", "transmission channel", "data transmission channel", "access channel", "data access channel", "link", "data link", "carrier", " radio frequency carrier” and/or any other similar term denoting a path or medium over which data is transmitted. Additionally, as used herein, the term "connection" refers to a connection between two devices via a RAT for the purpose of transmitting and receiving information.

As used herein, the terms instantiation, instantiation, and the like refer to the creation of an instance. An "instance" also refers to a concrete occurrence of an object that can occur, for example, during the execution of program code.

The terms "coupled", "communicatively coupled" and their derivations are used here. The term "coupled" may mean that two or more elements are in direct physical or electrical contact with each other, that two or more elements are in indirect contact with each other but still cooperate or interact with each other, and/or that one or more other elements are in between coupled or connected to the elements that are considered to be coupled to each other. The term "directly coupled" can mean that two or more elements are in direct contact with each other. The term "communicatively coupled" may mean that two or more elements may be in contact with one another via a means of communication, e.g., via a cable or other connection, via a wireless communication channel or link, and/or the like.

The term "information element" refers to a structural element that contains one or more fields. The term "field" refers to individual content of an information element or to a data element that contains content.

The term "SMTC" refers to an SSB-based measurement timing configuration created by SSB MeasurementTimingConfiguration is configured.

The term "SSB" refers to an SS/PBCH block.

The term “Primary Cell” refers to the MCG cell that operates on the primary frequency and in which the UE either performs the initial call setup process or initiates the reconnection process.

The term "Primary SCG cell" refers to the SCG cell in which the UE performs random access when performing the reconfiguration procedure with synchronization for DC operation.

The term "Secondary Cell" refers to a cell that provides additional radio resources on top of a dedicated cell for a CA configured UE.

The term "Secondary Cell Group" refers to the subset of serving cells comprising the PSCell and zero or more secondary cells for a DC configured UE.

The term “serving cell” refers to the primary cell for a UE in RRC_CONNECTED that is not configured with CA/DC as there is only one serving cell including the primary cell.

The term "Serving Cell" or "Serving Cells" refers to the set of cells including the Special Cell(s) and all secondary cells for a UE in RRC _CONNECTED with CA/DC.

The term "special cell" refers to the PCell of the MCG or the PSCell of the SCG for DC operation; otherwise, the term "special cell" refers to the Pcell.

Claims (10)

A new radio (NR) user equipment (UE) apparatus, the apparatus including a radio frequency (RF) interface and one or more processors coupled to the RF interface and a computing client service function (Comp CSF) that is configured to: perform registration for local computing resources with an application of the NR UE; perform registration for local and remote computing resources with an operating system/middleware (OS/middleware) of the NR UE; perform registration for remote computing resources with a cellular network computing control function (Comp CF); determine the local and remote computing resources based on registration with the Application, the OS/Middleware and the Comp CF; receive a dynamic workload offload registration message from the application; receive a dynamic workload offload request message for a compute task from the application; and perform dynamic workload offloading for the computing task based on the determined local and remote computing resources and the dynamic workload offloading request. The device after claim 1 , wherein the COMP CSF is configured to perform registration with the application via a Compl interface, and the registration with the application includes receiving a subscription request message from the application that includes a subscribed event/status, a notification trigger condition, and a time interval of local and remote computing resources; optionally wherein the subscribed event/status indicates availability of local and remote computing resources and status of current computing tasks; and/or optionally wherein the Comp CSF is configured to register with the operating system/middleware via a Comp 2 interface, and the subscribed event/status is an availability of local computing resources, one or more capabilities of a communication unit and displays wired and wireless connection availability. The device according to one of Claims 1 or 2 , wherein the Comp CSF is configured to register with the Comp CF via at least one of interfaces 1 and 14, and the registration and notification comprises providing a remote registration request message containing information about the remote computing resource, an availability status of the remote computing resource and a usage status of the remote status; and/or wherein the dynamic workload offloading registration message comprises an application manifest that includes an application identifier (ID), an application type, a registration validity condition, a registration validity time, an applied credential and ID, a port number, a fully qualified domain name (FQDN), a unique resource identifier (URI), a latency status, a throughput status, a workload offload setting, and an offload log. The device according to one of Claims 1 until 3 wherein the dynamic offload workload request message specifies a virtualized compute resource, a microservice, and a dedicated device; and/or wherein performing the dynamic workload offloading comprises: providing a local resource discovery message to the operating system/middleware; receiving a local resource discovery response message from the BS/middleware; and determining that the computational task is performed using the local resource based on the local resource discovery response message. The device according to one of Claims 1 until 4 wherein executing the dynamic workload comprises: providing a remote resource discovery message to the operating system/middleware; receiving a remote resource discovery response message from the BS/middleware; and determining that the computational task should be performed using the remote resource based on the remote resource discovery response message; or wherein executing the dynamic workload comprises: providing a remote resource discovery message to the Comp CF; receiving a remote resource discovery response message from the Comp CF; and determining that the computational task is performed using the remote resource based on the remote resource determination response message. The device according to one of Claims 1 until 5 , where the Comp CSF resides in a layer above the operating system/middleware. Apparatus of a new radio (NR) user equipment (UE), the apparatus including a radio frequency (RF) interface and one or more processors coupled to the RF interface, having a compute resource function on the client side (Comp RF-C) include that is configured to: receive a first discovery message from an operating system/middleware (OS/middleware) of the NR UE; provide a second discovery message to a computing resource function server side (Comp RF-S) of a cellular network; receive a first response message from the Comp RF-S; determine remote computing resources based on the first response message; provide a second response message to the operating system/middleware based on the determined remote computing resources; receive a first request message for a calculation task from the application; and perform dynamic workload offloading for the computing task with the Comp RF-S based on the determined remote computing resources and the first request message. The device after claim 7 wherein performing the dynamic workload offloading comprises: providing a second request message for the computational task to the Comp RF-S based on the first request message; and receiving a computational resource message from the Comp RF-S, the computational resource message indicating that the Comp RF-S is to perform the computational task; optionally wherein providing a second response message to the Comp RF-S triggering the modem to send a Layer 2 (L2)/Layer 3 (L3) request for computing resources over at least one of an interface 2' and an interface 3' includes. The device according to one of Claims 7 or 8th , where the Comp RF-C includes a Compute Control Plane (CP) function and a User Plane (UP) function and works as an endpoint to perform resource discovery, computing transport management and compute resource management, provide UP data, UP -Provide context and plan computing resources; and/or wherein providing the second discovery message comprises triggering a modem to send a second L2/L3 discovery message over at least one of interfaces 2' and 3'; and/or wherein the first discovery message specifies settings for a computing device and is transmitted according to a Peripheral Component Interconnect Express (PCIe) enumeration process. The device according to one of Claims 7 until 9 wherein the Comp RF-C and the Comp RF-S are located in a layer below the operating system/middleware; and/or wherein the Comp RF-C is further configured to that he provides the Comp RF-S with a computing capability message requesting the status of devices connected to a central unit and a distributed unit of the cellular network and a computing manifest containing the settings of the computing devices, a number of computing devices , a type of computing devices and a library; optionally wherein the computing capability message comprises a radio resource control (RRC) message to identify the computing devices connected to both the CU and the DU, and a predefined media access control (MAC) CE.

Images