Resource scheduling method, apparatus, device, communication system, and medium

Abstract

A resource scheduling method, device, equipment, communication system and medium are disclosed, which belong to the technical field of communication. The method comprises: acquiring first characteristic data of a mobile network and second characteristic data of a fixed access network, wherein the first characteristic data is used to represent the load of the mobile network, the second characteristic data is used to represent the load of the fixed access network, and the mobile network and the fixed access network share network resources; and according to the first characteristic data and the second characteristic data, scheduling first network resources from the network resources shared by the mobile network and the fixed access network to provide services for the mobile network, and scheduling second network resources from the network resources shared by the mobile network and the fixed access network to provide services for the fixed access network. By sharing network resources between the mobile network and the fixed access network, the cost can be reduced and the resource utilization rate can be improved.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a resource scheduling method, apparatus, device, communication system, and medium. Background Technology

[0002] In different environments, terminal devices will obtain network services through different types of networks. For example, when a user is at home or in the office, they will obtain network services through a fixed access network (such as a fiber optic network), while when a user is outdoors, they will obtain network services through a mobile network.

[0003] To improve network performance and better provide network services to users, both fixed access networks and mobile networks need to be configured with sufficient computing resources.

[0004] Deploying computing resources separately for different types of networks is costly and results in low resource utilization. Summary of the Invention

[0005] This application provides a resource scheduling method, apparatus, device, communication system, and medium that enables dynamic scheduling of computing network resources between mobile networks and communication networks, thereby allowing mobile networks and communication networks to share computing network resources, thereby reducing costs and improving resource utilization.

[0006] Firstly, this application provides a resource scheduling method. This method can be executed by a cooperative network controller, which can be a network controller deployed independently of the network controllers of the mobile network and the fixed access network, or it can be integrated with the network controllers of the mobile network or the fixed access network. The method includes: acquiring first characteristic data of the mobile network and second characteristic data of the fixed access network, wherein the first characteristic data characterizes the load of the mobile network, and the second characteristic data characterizes the load of the fixed access network, and the mobile network and the fixed access network share computing network resources; and scheduling first computing network resources to provide services to the mobile network and second computing network resources to provide services to the fixed access network based on the first characteristic data and the second characteristic data, wherein the first computing network resources and the second computing network resources are different computing network resources among the computing network resources.

[0007] In this embodiment, by enabling the mobile network and fixed access network to share computing network resources, the total amount of computing network resources only needs to meet the maximum sum of the resource demands of the mobile network and fixed access network at any given time. Then, based on the load of the mobile network and the load of the fixed access network, corresponding computing network resources are scheduled to provide services to the mobile network and fixed access network. In this way, while meeting the needs of the mobile network and fixed access network, the total amount of computing network resources required can be effectively reduced, costs can be lowered, and the utilization rate of computing network resources can be improved.

[0008] Optionally, network resources include computing resources and communication resources. Computing resources include at least computing power resources, and may also include storage resources. Communication resources include at least one of wavelength, transmission path, and bandwidth in the optical transport network.

[0009] Optionally, when computing network resources include computing power resources, scheduling computing power resources can be done in one or more of the following ways:

[0010] The first method involves determining a first trend of load change in the mobile network within a target time period based on the first feature data; when the first trend is an increase in load, increasing the computing power resources in the first computing network resources, or when the first trend is a decrease in load, decreasing the computing power resources in the first computing network resources.

[0011] The second method involves determining a second trend of load change in the fixed access network within a target time period based on the second characteristic data; when the second trend is an increase in load, increasing the computing power resources in the second computing network resources, or when the second trend is a decrease in load, decreasing the computing power resources in the second computing network resources.

[0012] The third method involves determining the computing power resources for communication-related accelerated computations in the mobile network and the fixed access network, and scheduling at least a portion of the remaining computing power resources for non-communication-related computations in the mobile network and the fixed access network based on the first feature data and the second feature data. The remaining computing power resources are the computing network resources shared by the mobile network and the fixed access network, excluding the computing power resources for communication.

[0013] In practical applications, if the computing resources shared by the mobile network and fixed access network are sufficient to meet the needs of both networks, then the first and second methods described above can be used to schedule the shared computing resources. In this third method, priority can be given to ensuring the computing resources required for communication-related accelerated computations, thereby guaranteeing network performance.

[0014] Optionally, based on the first feature data and the second feature data, a first computing network resource is scheduled from the computing network resources to provide services to the mobile network, and a second computing network resource is scheduled to provide services to the fixed access network. This can be achieved by: determining the traffic distribution in the target network based on the first feature data and the second feature data, where the traffic distribution indicates the traffic load of different sub-networks within the target network, and the different sub-networks have different coverage areas; and scheduling computing network resources for each sub-network in the target network based on the traffic distribution; wherein the target network is either the fixed access network or the mobile network.

[0015] Optionally, the computing resources scheduled for each sub-network may include communication resources and / or computing resources.

[0016] In some examples, network resources include communication resources, and the communication resources include wavelengths of an optical transport network. The optical transport network includes a first ring network and a second ring network connected to the same optical transport network device. The available wavelengths of both the first and second ring networks belong to a wavelength pool, and the available wavelengths of the first and second ring networks do not overlap. The wavelengths in this wavelength pool are shared by multiple ring networks containing the optical transport network device.

[0017] In this case, allocating communication resources to each sub-network in the target network according to the traffic distribution includes: when the traffic distribution indicates that the total traffic load of each sub-network connected to the first ring network is greater than the total traffic load of each sub-network connected to the second ring network, allocating a first available wavelength to the first ring network and allocating a second available wavelength to the second ring network, and the number of the first available wavelengths is greater than the number of the second available wavelengths.

[0018] The more available wavelengths, the greater the available bandwidth for transmission. Allocating available wavelengths means adjusting the available bandwidth corresponding to each ring network. By allocating more available wavelengths to the ring networks corresponding to sub-networks with higher traffic loads, these sub-networks can have greater available bandwidth, which helps reduce latency within the sub-networks. When traffic distribution is predicted based on the first and second characteristic data, the available wavelengths of the ring networks corresponding to the sub-networks can be adjusted in advance to ensure a stable and optimal user experience.

[0019] In other examples, communication resources include transmission paths in an optical transport network. Allocating communication resources to each sub-network within the target network based on the traffic distribution includes: selecting a transmission path for the target sub-network with a bandwidth margin greater than the bandwidth required by the traffic load, based on the traffic load of the target sub-network; wherein the target sub-network is any sub-network within the target network. By selecting a transmission path with a bandwidth margin greater than the bandwidth required by the traffic load, the performance of the end-to-end connection can be optimized.

[0020] In one possible implementation, the traffic distribution in the target network can be a predicted traffic distribution based on the first feature data and the second feature data. This allows for advance adjustment of the available wavelengths of the ring network corresponding to the sub-network based on the predicted traffic distribution, ensuring a stable and optimal user experience.

[0021] In another possible implementation, the traffic distribution in the target network can be the actual traffic distribution determined based on the first feature data and the second feature data.

[0022] Optionally, the method further includes: obtaining the end-to-end latency requirement of the target terminal; determining the transmission latency required for the migration of the target terminal's computing data between computing resources based on the access location of the target terminal; and allocating target computing resources to the target terminal based on the transmission latency and the end-to-end latency requirement, wherein the target computing resources are used to provide services to the target terminal.

[0023] For example, the target terminal can be a rapidly moving terminal accessed via a mobile network, such as a drone. Because drones move very quickly, their access locations change rapidly. Following the principle of selecting computing resources based on proximity (i.e., prioritizing computing resources closer to the access location), the location of the selected computing resources may also change. In this case, allocating target computing resources to the target terminal based on the transmission latency required for the migration of the target terminal's computational data between computing resources can ensure that the end-to-end latency of the target terminal is always within the end-to-end latency requirements, thus improving the reliable end-to-end service experience.

[0024] Optionally, the method further includes: when the computing power resources in the computing network are insufficient to meet the load of the fixed access network and the mobile network, calling other computing power resources outside the computing network resources to provide services to the fixed access network and / or the mobile network. By calling other computing power resources, the communication quality of users in the fixed access network and the mobile network can be guaranteed in extreme cases. Extreme cases refer to a sudden increase in traffic load in the fixed access network and the mobile network in a certain location, exceeding the capacity of their shared computing network resources.

[0025] Optionally, other computing resources can be invoked in the following way: a request is sent to the control device of the other computing resources, the request being for the use of the other computing resources; dedicated communication resources are configured for the request from the computing network resources, the dedicated communication resources being used to transmit the computational data related to the request. By allocating dedicated communication resources for the computational data related to the request, service latency can be reduced when invoking other computing resources.

[0026] Optionally, the first feature data includes at least one of first traffic data and first connection data, wherein the first traffic data represents the change in traffic load in the mobile network over time and / or space, and the first connection data represents the number of connections in the mobile network. The second feature data includes at least one of second traffic data and second connection data, wherein the second traffic data represents the change in traffic load in the fixed access network over time and / or space, and the second connection data represents the number of connections in the fixed access network.

[0027] Optionally, the load of the mobile network and the load of the fixed access network are correlated, for example, negatively correlated during periods of high user activity. Applying this method to mobile and fixed access networks where load is correlated is beneficial for improving the utilization rate of network resources (especially computing power resources).

[0028] Optionally, the computing resources shared by the mobile network and the fixed access network are located in a first network device, which belongs to the mobile network; or, the computing resources shared by the mobile network and the fixed access network are located in a second network device, which belongs to the fixed access network; or, a portion of the computing resources shared by the mobile network and the fixed access network are located in a first network device and another portion are located in a second network device, where the first network device belongs to the mobile network and the second network device belongs to the fixed access network.

[0029] Optionally, both the first network device and the second network device are access network devices. Deploying computing resources in the access network devices, closer to the terminal, helps reduce latency.

[0030] Secondly, a resource scheduling device is also provided. This resource scheduling device has the function of implementing the method described in the first aspect. The function can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function.

[0031] Thirdly, a computer device is also provided, including a processor and a memory, wherein the memory stores program code; the processor is used to read and execute the program code stored in the memory to implement the resource scheduling method provided in the first aspect.

[0032] Optionally, the processor may be one or more, and the processor may be a multi-core processor, and the memory may be one or more.

[0033] Optionally, the memory may be integrated with the processor, or the memory may be separated from the processor.

[0034] In the specific implementation process, the memory can be a non-transitory memory, such as read-only memory (ROM), which can be integrated with the processor on the same chip or set on different chips. This application does not limit the type of memory or the way the memory and processor are set.

[0035] Fourthly, a computer-readable storage medium is also provided, wherein a software program is stored therein, and the software program, when read and executed by one or more processors, can implement the resource scheduling method provided in the first aspect.

[0036] Fifthly, a computer program (product) is provided, the computer program (product) comprising: computer program code, wherein when the computer program code is run by a computer device, the computer device executes the resource scheduling method provided in the first aspect.

[0037] Sixthly, a chip is provided, the chip including a processor and a communication interface. The processor is used to execute instructions to cause the chip to perform the resource scheduling method provided in the first aspect.

[0038] In a seventh aspect, a communication system is provided, the communication system comprising a cooperative network controller, multiple access network devices, and multiple transport network devices; the multiple access network devices include wireless access network devices and fixed access network devices; the wireless access network devices belong to a mobile network, the fixed access network devices belong to a fixed access network, the fixed access network and the mobile network share computing network resources, and the cooperative network controller is used to implement the resource scheduling method provided in the first aspect above. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the structure of a communication system provided in an embodiment of this application;

[0040] Figure 2This is a schematic diagram of the FTTR system architecture;

[0041] Figure 3 This is a schematic diagram illustrating traffic changes in different networks within a communication system provided in an embodiment of this application;

[0042] Figure 4 This is a flowchart illustrating a resource scheduling method provided in an embodiment of this application;

[0043] Figure 5 This is a flowchart illustrating a resource scheduling method provided in an embodiment of this application;

[0044] Figure 6 This is a schematic diagram illustrating an application scenario of a resource scheduling method provided in an embodiment of this application;

[0045] Figure 7 This is a schematic diagram illustrating another application scenario of a resource scheduling method provided in the embodiments of this application;

[0046] Figure 8 This is a schematic diagram illustrating another application scenario of a resource scheduling method provided in the embodiments of this application;

[0047] Figure 9 This is a schematic diagram of the structure of a resource scheduling device provided in an embodiment of this application;

[0048] Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0050] Figure 1 This is a schematic diagram of the structure of a communication system provided in an embodiment of this application. For example... Figure 1 As shown, the communication system includes multiple access network devices and multiple optical transport network (OTN) devices.

[0051] The access network equipment includes radio access network (RAN) equipment and fixed access network equipment. RAN equipment belongs to the mobile network, and fixed access network equipment belongs to the fixed access network. In this embodiment, the mobile network includes cellular communication networks and / or satellite communication networks, and the fixed access network includes, but is not limited to, fiber optic networks (also known as optical networks). The following description will use a cellular communication network as the mobile network and a fiber optic network as the fixed access network as an example.

[0052] Wireless access network equipment may include base stations. Figure 1 The example provided illustrates the use of active antenna units (AAUs) or remote radio units (RRUs), centralized units (CUs), and distributed units (DUs) in base stations in China and Israel.

[0053] The AAU is deployed at base station sites to provide wireless signal coverage so that terminal devices can access the mobile network and communicate; the DU is deployed near the AAU (e.g., at base station sites) and is responsible for low-level protocol processing and real-time sensitive tasks; the CU is deployed in the cloud or regional data center (RDC) and is responsible for non-real-time high-level protocol processing and global resource scheduling.

[0054] The AAU connects to the DU via a front-haul network. The front-haul network can use a star, chain, or ring topology. In a star topology, the AAU connects directly to the corresponding DU via fiber optic or wireless connections, for example... Figure 1 AAU1 and AAU2 are directly connected to DU1 via optical fiber. In a ring topology, multiple AAUs and one DU form a ring network, for example... Figure 1 AAU3-AAU5 and DU2 form a ring network. DUs are connected to CUs via a mid-haul network, and CUs are connected to core network equipment via a back-haul network. CUs can support collaborative management of multiple DUs, enabling efficient scheduling of network resources. The mid-haul network can adopt a ring or mesh topology, and the back-haul network can also adopt a ring or mesh topology. Both the mid-haul and back-haul networks can be carried by OTN or other wired transmission networks. Figure 1 In this network, the midhaul network and backhaul network are carried by OTN. The metropolitan area core network can adopt a mesh topology, enabling data center interconnection (DCI).

[0055] Fixed access network equipment includes user-side equipment and network-side equipment. For fiber optic networks, user-side equipment includes optical network units (ONUs) or optical network terminals (ONTs). Network-side equipment includes optical line terminals (OLTs). The OLT connects to multiple ONUs via an ODN (Optical Distribution Network). The ODN includes passive optical splitters for optical power distribution, a backbone fiber connecting the passive optical splitter and the OLT, and branch fibers connecting the passive optical splitter and the ONUs. The ONUs, OLT, and ODN together form a passive optical network (PON). When transmitting downlink signals, the downlink signal sent by the OLT is transmitted to each ONU through the splitter, and the ONUs selectively receive downlink data belonging to themselves from the downlink signal. When transmitting uplink signals, the uplink signals sent by multiple ONUs are combined into a single optical signal by the splitter and transmitted to the OLT.

[0056] Building upon fiber-to-the-home (FTTH) or fiber-to-the-office (FTTO) networks, to address signal coverage issues (such as wireless local area network (WLAN) signals) in home or office networks, the fiber optic cable can be extended further into the room. Optical terminal equipment providing WLAN signals is then installed inside the room, thus reducing the distance between the user terminal and the wireless access point (AP) and improving signal quality. This technology is called fiber-to-the-room (FTTR).

[0057] Figure 2 This is a schematic diagram of the FTTR system architecture. (Example) Figure 2 As shown, in FTTH / O, the OLT is deployed in the central equipment room, while the ONUs are deployed in homes or offices. The master device in the FTTR network acts as both an ONU in the FTTH network and an upstream device for the FTTR slave devices, managing them. Slave devices in the FTTR network can be deployed in various rooms of homes or offices. Slave devices possess ONU functionality and can also function as access points (APs) to provide WLAN signals to user terminals. Both master and slave devices can connect to user terminals through a user network interface (UNI).

[0058] In an FTTR network, multiple slave devices can be deployed, each connected to the master device via an optical splitter. The master device can centrally manage and configure all slave devices. The master device can also be called a "master gateway," "master optical modem," or "master FTTR unit (or master fiber unit, MFU)," etc., while slave devices can be called "slave gateways," "slave optical modems," or "slave FTTR units (or sub-fiber units, SFU)," etc.

[0059] See you again Figure 1 OTN equipment includes, but is not limited to, OTN switches, optical cross-connects (OXCs), and reconfigurable optical add-drop multiplexers (ROADMs). Based on their location, multiple ONT devices can be categorized into metro-access network devices, metro-aggregation network devices, and metro-core network devices. Metro-access network devices connect to the OLT and CU, while metro-aggregation network devices are located between the metro-access network devices and the metro-core network devices, serving to connect them.

[0060] Mobile networks and fixed access networks can share communication resources in an OTN, such as transmission paths (e.g., fiber optic links), bandwidth, and wavelength.

[0061] To improve network performance, at least one of the DU and CU in the RAN, as well as the OLT, can be configured with certain computing resources to perform communication-related accelerated calculations or provide artificial intelligence services. These computing resources can be deployed not only in access network equipment but also in other locations, such as data centers (DCs). Figure 1In this context, the Data Center (DC) includes access data centers (DCs) connected to metropolitan area access network equipment, aggregation DCs connected to metropolitan area aggregation network equipment, and core DCs connected to metropolitan area core network equipment. Access DCs can act as Core Units (CUs), providing multi-access edge computing (MEC) services for handling ultra-reliable and low-latency communication (URLLC) services. Aggregation DCs can act as content delivery networks (CDNs) for enhanced mobile broadband (eMBB) user plane (UE) services or massive machine-type communication (mMTC) services. Core DCs can act as CDNs or Gateways (GWs) for enhanced mobile broadband (eMBB) control plane (UE) services. Figure 1 The primary purpose of data centers (DCs) in this context is to provide services for mobile networks, such as 5G mobile communication networks. In addition, computing resources can also be deployed in other DCs besides those serving mobile networks, such as enterprise data centers (EDCs), regional data centers (RDCs), or cloud data centers (CDCs).

[0062] Computing resources include at least computing power resources, and may also include storage resources. Computing power resources include, but are not limited to, accelerators such as graphics processing units (GPUs), central processing units (CPUs), data processing units (DPUs), tensor processing units (TPUs), and neural network processing units (NPUs), and may include one or more of these. Storage resources include, but are not limited to, hard disks and memory.

[0063] In the embodiments of this application, computing resources and OTN communication resources are collectively referred to as computing network resources.

[0064] It should be noted that, Figure 1The system architecture shown is for illustrative purposes only and can be modified according to actual needs. For example, the topology of the fronthaul network in a RAN can be changed; the topology in an OTN can also be changed; and so on. Figure 1 In this context, AI represents computing resources. The more AIs there are, the more abundant the deployed computing resources are. For example, the core network has more computing resources deployed than the aggregation network.

[0065] also, Figure 1 The communication system shown also includes a network controller, such as the cooperative network controller described below for scheduling computing resources for mobile networks and fixed access networks.

[0066] Figure 3 This is a schematic diagram of traffic changes in different networks in a communication system provided in this application embodiment, illustrating the process of traffic load changing over time in different networks in a specific area. Figure 3 The horizontal axis represents time, and the vertical axis represents the standardized traffic load. The standardized traffic load can be calculated as the product of the ratio of actual traffic to maximum total traffic and 100%. Total traffic is the sum of traffic in the radio access network (RAN), FTTO, and FTTH at the same time, while maximum total traffic is the maximum value among the sums at different times.

[0067] like Figure 3 As shown, the traffic load of the radio access network (RAN) is higher in the morning (6-9 am), noon (11 am-2 pm), and evening (5 pm-8 pm), and lower in other time periods. This is because these time periods correspond to the commute to work, lunch break, and commute home, respectively, during which most people need to use mobile network access for communication.

[0068] For FTTO, its traffic load gradually increases in the morning (6-9 am), remains high throughout the day, and gradually decreases after 8 pm. This is because people arrive at their workplaces between 6 am and 9 am and start accessing the network via FTTO; they need to work at their workplaces during the day and go home after get off work in the evening.

[0069] For FTTH, the traffic load is low during the day (6 am - 6 pm), and gradually increases from the evening (starting at 5 pm) until it reaches its peak.

[0070] RAN belongs to the mobile network, while FTTO and FTTH belong to the fixed access network. Figure 3It can be seen that the load on the mobile network and the load on the fixed access network are basically negatively correlated during the peak user activity period (6:00 AM to 12:00 AM). This is because during the day, user devices mostly generate services in commercial and office areas, while at night, user devices mostly generate services in residential areas. Therefore, within this specific area, the total number of communication users is relatively fixed, but the bandwidth requirements for the mobile and fixed access networks vary at different times.

[0071] If the computing resources of fixed access networks and mobile networks are configured independently, each network needs to be configured with computing resources greater than or equal to its respective peak demand to meet those demands. This not only results in high hardware costs but also high energy consumption and maintenance costs. Furthermore, most computing resources remain idle for the majority of the time, leading to low utilization of computing resources (especially computing resources).

[0072] from Figure 3 It can be seen that the traffic load of different types of networks fluctuates significantly over time, and the peak traffic load of different types of networks occurs at different times. Therefore, in this embodiment, the fixed access network and the mobile network share computing resources. Thus, the total amount of computing resources configured for the mobile network and the fixed access network only needs to be greater than or equal to the peak value of the sum of their resource requirements. When the peak traffic load of different types of networks occurs at different times, and the traffic load of different types of networks shows a trend of ebb and flow, the total amount of computing resources required for sharing computing resources is significantly less than the sum of the peak computing resource requirements of each type of network. Furthermore, when the traffic load of the fixed access network is low, the traffic load of the mobile network is high, so more computing resources can be allocated to the mobile network; conversely, when the traffic load of the fixed access network is high, the traffic load of the mobile network is low, so more computing resources can be allocated to the fixed access network. The overall utilization rate of the computing resources shared by the mobile and fixed access networks is high. Therefore, sharing computing resources between the fixed access network and the mobile network not only reduces costs but also improves the utilization rate of computing resources.

[0073] Based on this, embodiments of this application provide a resource scheduling method that can schedule computing network resources for mobile networks and fixed access networks sharing computing network resources, so as to meet the needs of mobile networks and fixed access networks.

[0074] Figure 4 This is a flowchart illustrating a resource scheduling method provided in an embodiment of this application. The method is applied to... Figure 1The communication system shown can be executed by a collaborative network controller (also known as an overall network controller or a shared network controller), which can be a standalone device or integrated into... Figure 1 This can be found in the OLT, OTN, or CU, or integrated with the network controller of the mobile network or the network controller of the fixed access network. For example... Figure 4 As shown, the method includes the following steps.

[0075] In step 401, the first feature data of the mobile network and the second feature data of the fixed access network are obtained.

[0076] The first feature data is used to characterize the load of the mobile network, and the second feature data is used to characterize the load of the fixed access network. The mobile network and the fixed access network share computing network resources.

[0077] Optionally, the first feature data includes at least one of first traffic data and first connection data. The first traffic data represents the variation of traffic load in the mobile network over time and / or space; the first connection data represents the number of connections in the mobile network. For example, the first traffic data includes multiple traffic values ​​and a traffic identifier and time corresponding to each traffic value. The traffic identifier indicates the location corresponding to the traffic value, the sub-network of the corresponding mobile network, or the network device in the mobile network corresponding to the traffic value. The first connection data may include the number of user connections.

[0078] For mobile networks, a higher traffic load results in a greater network load and requires more computing resources. Conversely, a lower traffic load results in a lower network load and requires fewer computing resources. A large number of connections indicates a potentially larger computational load and requires more computing resources; conversely, a smaller number of connections indicates a potentially smaller computational load and requires fewer computing resources. Therefore, using at least one of the first traffic data and the first connection data as the first characteristic data can reflect the mobile network load, or in other words, the mobile network's demand for computing resources. This application does not limit the specific content of the first characteristic data, as long as it reflects the mobile network's demand for computing resources.

[0079] Optionally, the second feature data includes at least one of second traffic data and second connection data. The second traffic data represents the variation of traffic load in the fixed access network over time and / or space; the second connection data represents the number of connections in the fixed access network.

[0080] For example, the second traffic data includes multiple traffic values ​​and a traffic identifier and time corresponding to each traffic value. The traffic identifier is used to indicate the location corresponding to the traffic value, the subnetwork of the corresponding fixed access network, or the network device in the fixed access network corresponding to the traffic value. The second connection data may include the number of user connections.

[0081] For fixed access networks, a higher traffic load results in a greater network load and requires more computing resources. Conversely, a lower traffic load results in a lower network load and requires fewer computing resources. A large number of connections indicates a potentially larger computational load and requires more computing resources; conversely, a smaller number of connections indicates a potentially smaller computational load and requires fewer computing resources. Therefore, using at least one of the second traffic data and the second connection data as the second feature data can accurately reflect the load of the fixed access network, or in other words, the network resource requirements. This application does not limit the specific content of the second feature data, as long as it reflects the network resource requirements of the fixed access network.

[0082] In this embodiment, a connection can refer to a connection at the medium access control (MAC) layer, a network layer, a session layer, or an application layer. MAC layer connections include, for example, Ethernet or OTN connections; network layer connections include, for example, Transmission Control Protocol / Internet Protocol (TCP / IP) connections; session layer connections include, for example, socket connections; and application layer connections include, for example, Hypertext Transfer Protocol (HTTP) connections. In this embodiment, the connection can refer to a network layer connection.

[0083] In implementation, first traffic data can be obtained from the network controller of the mobile network, and second traffic data can be obtained from the network controller of the fixed access network. The network controller of the mobile network can obtain the first traffic data using any suitable method, for example, by having network devices in the mobile network report the first traffic data in real time. The network controller of the fixed access network can obtain the second traffic data using any suitable method, for example, by having network devices in the fixed access network report the second traffic data in real time.

[0084] Connection data can be obtained from the corresponding control device based on the layer at which the connection exists. For example, MAC layer connection data can be obtained through an Ethernet controller or OTN device; network layer connection data can be obtained through a server providing network services.

[0085] In step 402, based on the first feature data and the second feature data, the first computing network resources are scheduled to provide services to the mobile network and the second computing network resources are scheduled to provide services to the fixed access network from the computing network resources shared by the mobile network and the fixed access network.

[0086] Among them, the first computing network resource and the second computing network resource are different computing network resources shared by the mobile network and the fixed access network.

[0087] As mentioned above, the computing resources shared by the mobile network and the fixed access network include one or more of the aforementioned communication resources and computing resources. In the embodiments of this application, sharing computing resources means that any computing resource can be scheduled for use by the mobile network or the fixed access network at any time as needed. That is, any computing resource shared by the mobile network and the fixed access network can be used as a first computing resource in one time period and as a second computing resource in another time period.

[0088] In this embodiment, the shared computing resources of the mobile network and the fixed access network are dedicated to providing computing services for both. Optionally, the computing services may include communication-related accelerated computing services and non-communication-related computing services. Communication-related accelerated computing may be network performance-related computations below layer 3 in the OSI 7-layer model. Non-communication-related computing services may include artificial intelligence (AI) services such as fault diagnosis and location. Communication-related accelerated computing in the mobile network includes, but is not limited to, physical layer signal processing and beam optimization, while communication-related accelerated computing in the fixed access network includes, but is not limited to, bandwidth allocation and path optimization.

[0089] In this embodiment, by enabling the mobile network and fixed access network to share computing network resources, the total amount of computing network resources only needs to meet the maximum sum of the resource demands of the mobile network and fixed access network at any given time. Then, based on the load of the mobile network and the load of the fixed access network, corresponding computing network resources are scheduled to provide services to the mobile network and fixed access network. In this way, while meeting the needs of the mobile network and fixed access network, the total amount of computing network resources required can be effectively reduced, costs can be lowered, and the utilization rate of computing network resources can be improved.

[0090] In implementation, the computing resources in the shared computing network resources of mobile networks and fixed access networks can be arranged in one of the following three ways:

[0091] Method 1: The computing resources in the shared computing network resources of the mobile network and fixed access network are set in the first network device, which belongs to the mobile network;

[0092] Method 2: The computing resources shared by the mobile network and the fixed access network are set in the second network device, which belongs to the fixed access network;

[0093] Method 3: In the shared computing resources of the mobile network and the fixed access network, a portion of the computing resources are located in the first network device and the other portion is located in the second network device. The first network device belongs to the mobile network and the second network device belongs to the fixed access network.

[0094] Both the first network device and the second network device are access network devices. For example, the first network device includes the aforementioned DU and / or CU, and the second network device includes the aforementioned OLT. Prioritizing the selection of the first and second network devices that are geographically close, especially co-located first and second network devices, for sharing computing resources can help reduce transmission latency between them.

[0095] In this embodiment, the network load is positively correlated with the amount of scheduled computing network resources. For example, the greater the load of the mobile network, the more first computing network resources are scheduled; the smaller the load of the mobile network, the fewer first computing network resources are scheduled. Similarly, the greater the load of the fixed access network, the more second computing network resources are scheduled; the smaller the load of the fixed access network, the fewer second computing network resources are scheduled.

[0096] Optionally, the first computing resource can be located on one or more network devices. The second computing resource can be located on one or more network devices.

[0097] When the first or second computing resource to be scheduled is located in multiple network devices, bandwidth adjustment can be performed through the in-band channels of the OTN device (such as the optical channel data unit, ODUk), thereby achieving simple, low-latency, and high-bandwidth-efficiency request and result transfer. For example, bandwidth adjustment can be performed through the in-band channels of the fgOTN (fine-grained OTN) device.

[0098] Optionally, in step 402, when the computing network resources include computing power resources, the methods for scheduling computing power resources include, but are not limited to, the following:

[0099] The first method involves determining a first trend of change in the load of the mobile network within a target time period based on the first characteristic data; when the first trend is an increase in load, increasing the computing power resources in the first computing network resources, or when the first trend is a decrease in load, decreasing the computing power resources in the first computing network resources.

[0100] In this first approach, the first trend of change in the load of the mobile network within the target time period can be determined by comparing the first characteristic data at the start time of the target time period with the first characteristic data at the end time of the target time period to obtain the first trend of change.

[0101] For example, when the first characteristic data is the first traffic data, if the first difference between the first traffic data at the start of the target time period and the first traffic data at the end of the target time period is greater than a1, the first trend can be determined as an increase in load; if the first difference between the first traffic data at the start of the target time period and the first traffic data at the end of the target time period is less than b1, the first trend can be determined as a decrease in load. Here, a1 is greater than or equal to 0, and b1 is less than or equal to 0. The method for determining the first trend when the first characteristic data is the first connection data is similar to that for the first traffic data, and a detailed description is omitted here.

[0102] Optionally, the increase or decrease of computing resources in the first computing network can be determined based on the first difference. The larger the absolute value of the first difference, the greater the change in computing resources; the smaller the absolute value of the first difference, the smaller the change in computing resources. In implementation, a first correspondence between the difference and the change in computing resources can be pre-set, and then the change in computing resources corresponding to the first difference can be determined according to the first correspondence. Then, the computing resources in the first computing network can be increased or decreased according to the change in computing resources.

[0103] The second method involves determining the second trend of load change in the fixed access network within the target time period based on the second characteristic data; when the second trend is an increase in load, increasing the computing power resources in the second computing network resources, or when the second trend is a decrease in load, decreasing the computing power resources in the second computing network resources.

[0104] In this second method, the second trend of load variation of the fixed access network within the target time period can be determined by comparing the second characteristic data at the start and end of the target time period. For example, when the second characteristic data is second connection data, if the second difference between the second connection data at the start and end of the target time period is greater than a2, the second trend can be determined to be an increase in load; if the second difference between the second connection data at the start and end of the target time period is less than b2, the first trend can be determined to be a decrease in load. Here, a2 is greater than or equal to 0, and b2 is less than or equal to 0. The method for determining the second trend when the second characteristic data is second traffic data is similar to that for second connection data, and a detailed description is omitted here.

[0105] Optionally, the increase or decrease of computing resources in the second computing network can be determined based on the second difference. The larger the absolute value of the second difference, the greater the change in computing resources; the smaller the absolute value of the second difference, the smaller the change in computing resources. In implementation, a second correspondence between the difference and the change in computing resources can be pre-set, and then the change in computing resources corresponding to the second difference can be determined according to the second correspondence. Then, the computing resources in the second computing network can be increased or decreased according to the change in computing resources.

[0106] In practical applications, if the computing resources shared by the mobile network and the fixed access network are sufficient to meet the needs of both networks, then the first and second methods described above can be used to schedule the shared computing resources.

[0107] The third method involves determining the computing power resources for communication-related accelerated computation in mobile networks and fixed access networks, and scheduling at least a portion of the remaining computing power resources for non-communication-related computation in mobile networks and fixed access networks based on the first and second feature data. The remaining computing power resources are the computing network resources shared by mobile networks and fixed access networks, excluding the computing power resources for communication.

[0108] In a first possible implementation of this third approach, the mobile network's communication computing resources corresponding to the first traffic data are determined based on a first mapping relationship between the traffic range and the amount of communication computing resources in the mobile network; the fixed access network's communication computing resources corresponding to the second traffic data are determined based on a second mapping relationship between the traffic range and the amount of communication computing resources in the fixed access network; then, a first amount of non-communication computing resources in the mobile network corresponding to the first connection data is determined based on a third mapping relationship between the connection number range and the amount of non-communication computing resources; and a second amount of non-communication computing resources in the mobile network corresponding to the second connection data is determined based on a fourth mapping relationship between the connection number range and the amount of non-communication computing resources. If the remaining computing resources are sufficient, the non-communication computing resources are scheduled according to the determined first and second resource amounts. If the remaining computing resources are insufficient, the non-communication computing resources are scheduled according to a set rule based on the determined first and second resource amounts.

[0109] The rules include, but are not limited to, scheduling non-communication computing resources based on the priority of computing tasks. Optionally, the cooperative network controller can determine the priority based on the network layer related to the computing task, with higher priority at lower layers; or the cooperative network controller can determine the priority of the computing task based on the priority carried in the resource scheduling request sent by the network controller of the mobile network or fixed access network.

[0110] The first to fourth mapping relationships mentioned above can all be obtained from historical data on the use of computing resources.

[0111] In a second possible implementation of this third approach, at least a portion of the shared computing resources can be reserved for communication purposes. The reserved proportion can be set according to actual needs, and this application embodiment does not impose any limitations on this; for example, it can be all shared computing resources or 80% of the shared computing resources. The reserved computing resources are periodically determined to be in a low-load state. If it is determined that the reserved computing resources are in a low-load state, communication-related calculations are concentrated in the first portion of the determined reserved computing resources; then, the second portion of the reserved computing resources is scheduled for non-communication-related calculations. The methods for determining whether computing resources are in a low-load state include, but are not limited to, comparing the utilization rate of the computing resources within a period with a utilization rate threshold. If the utilization rate of the computing resources within a period is lower than the utilization rate threshold, it indicates that the computing resources are in a low-load state; if the utilization rate of the computing resources within a period is higher than or equal to the utilization rate threshold, it indicates that the computing resources are not in a low-load state.

[0112] Optionally, the reserved computing resources can be used as computing resources for communication in the mobile network instead of the fixed access network, or as computing resources for communication in both the mobile and fixed access networks. For example, if the impact of computing resources for communication in the fixed access network (especially fiber optic networks) on network performance is relatively small, the reserved computing resources can be used only as computing resources for communication in the mobile network.

[0113] In the third possible implementation of this third approach, since the impact of communication computing resources on network performance is more significant in mobile networks, while the impact of communication computing resources on network performance is relatively small in fixed access networks (especially fiber optic networks), the step of determining the communication computing resources of the fixed access network corresponding to the second traffic data can be omitted based on the first possible implementation, according to the second mapping relationship between the traffic range and the amount of communication computing resources of the fixed access network.

[0114] In this third approach, priority can be given to ensuring the computing power resources required for communication-related accelerated computing, so as to guarantee network performance.

[0115] The fourth method involves pre-setting a first AI model and a second AI model to determine the computing resource requirements. The first AI model takes first feature data as input and outputs the computing resource requirements corresponding to the first feature data. The second AI model takes second feature data as input and outputs the computing resource requirements corresponding to the second feature data. Computing resources in the first computing network are scheduled according to the computing resource requirements output by the first AI model; computing resources in the second computing network are scheduled according to the computing resource requirements output by the second AI model.

[0116] Optionally, the smallest granularity (or smallest unit) for scheduling computing resources can be an accelerator, such as a GPU; or, a computing power partition of an accelerator, such as in multi-instance GPU (MIG) technology, where a single GPU is divided into multiple instances, and each instance is a computing power partition.

[0117] Figure 5 This is a flowchart of another resource scheduling method provided in an embodiment of this application. This method is applied to... Figure 1 The communication system shown can consist of a collaborative network controller (also called an overall network controller or a shared network controller), which can be a standalone device or integrated into... Figure 1 The network resources are located in the OLT, OTN, or CU, or integrated with the network controller of the mobile network or the network controller of the fixed access network. In this embodiment, the network resources include communication resources and / or computing resources. Figure 5 As shown, the method includes the following steps.

[0118] In step 501, the first feature data of the mobile network and the second feature data of the fixed access network are obtained.

[0119] The first characteristic data includes at least one of first traffic data and first connection data. The second characteristic data includes at least one of second traffic data and second connection data. For details on obtaining the first and second traffic data and the first and second connection data, please refer to step 301 above; they will not be described in detail here.

[0120] In step 502, the traffic distribution in the target network is determined based on the first feature data and the second feature data.

[0121] The target network comprises multiple sub-networks, each with a different coverage area. Here, the target network can be a mobile network or a fixed access network. Traffic distribution is used to indicate the traffic load of different sub-networks within the target network.

[0122] In one possible implementation, the traffic distribution in the target network can be predicted based on the first feature data and the second feature data.

[0123] This application does not limit the prediction method based on the traffic distribution in the target network; any suitable method can be selected according to actual needs.

[0124] In some examples, the target network is a fixed access network. Predicting the traffic distribution of the fixed access network can be done as follows: predict the traffic changes of each subnetwork of the fixed access network based on access point data; determine the actual traffic of the target subnetwork of the target network based on the second feature data; and finally, obtain the predicted traffic load of the target subnetwork based on the predicted traffic changes and the actual traffic of the target subnetwork. Here, the target subnetwork is any subnetwork within the fixed access network.

[0125] In this case, the first feature data includes, in addition to at least one of the aforementioned first traffic data and first connection data, the access point data. This access point data includes multiple access point sets, each containing identifiers of multiple historical base stations accessed by a terminal device. Optionally, each access point set can be reported by the terminal device to the network controller of the mobile network, and then obtained from the network controller.

[0126] In other examples, traffic distribution in a target network can be inferred based on patterns in historical data. For instance, a neural network model of the target network can be trained based on historical data, with the input of the first feature data and the second feature data, and the output being the traffic distribution.

[0127] In other examples, the traffic distribution in the target network can be inferred by combining the first feature data, the second feature data, and real-time information about the public event. The real-time information about the public event includes the location, time, and scale of the event. For example, based on the location of the public event, the traffic of one or more sub-networks in the mobile network covering the location of the public event and the surrounding area is statistically analyzed based on the first feature data within a set time period before the start of the public event; the actual traffic distribution in the fixed access network is determined based on the second feature data; then, according to a set proportion of the statistically analyzed traffic, the amount of traffic change that will be transferred to the fixed access network is determined; based on this actual traffic distribution and traffic change, the traffic load of the sub-network corresponding to the location of the public event is predicted.

[0128] In another possible implementation, the actual traffic distribution in the target network can be determined based on the first feature data and the second feature data.

[0129] In step 503, computing resources are allocated to each sub-network in the target network according to the traffic distribution.

[0130] In step 503, network resources include communication resources and / or computing resources.

[0131] In some examples, network resources include computing resources. Based on traffic distribution, network resources are allocated to each subnetwork within the target network, including allocating computing resources to each subnetwork according to the proportion of traffic in each subnetwork to the total traffic in the target network. This way, as traffic migrates between subnetworks, the allocation of computing resources changes accordingly, which helps ensure the network service performance of the target network.

[0132] Figure 6 yes Figure 5 A schematic diagram illustrating another application scenario of the method shown. (For example...) Figure 7 As shown, the target network is a fixed access network, and the sub-networks of the fixed access network include P2MP network, point-to-point (P2P) network and PON network.

[0133] In this network, the PON network covers area 1, and the OLT in the PON network is connected to the OTN device located in ring network 2. The P2MP network and point-to-point (P2P) network cover area 2, and the terminals in both networks are connected to the OTN device located in ring network 3. When the traffic load of the subnetwork in area 1 increases, more computing resources can be allocated to the subnetwork in area 1, while fewer computing resources can be allocated to the subnetwork in area 2.

[0134] In other examples, network resources include communication resources, which include wavelengths of the optical transport network. The optical transport network includes a first ring network and a second ring network, both connected to the same optical transport network equipment. The available wavelengths of both the first and second ring networks belong to a wavelength pool, and these available wavelengths do not overlap. This wavelength pool includes multiple wavelengths, which can be dynamically allocated to multiple ring networks where the optical transport network equipment resides. That is, the multiple ring networks where the optical transport network equipment resides share multiple wavelengths in the wavelength pool; this technique can be called pooled wavelength division. The first ring network connects to one or more sub-networks in the target network; the second ring network connects to one or more sub-networks in the target network; and the sub-networks connected to the first and second ring networks are different. Here, the optical transport network mainly refers to a metropolitan area network (MAN).

[0135] When the traffic distribution indicates that the total traffic load of each sub-network connected to the first ring network is greater than the total traffic load of each sub-network connected to the second ring network, a first available wavelength is allocated to the first ring network and a second available wavelength is allocated to the second ring network, and the number of first available wavelengths is greater than the number of second available wavelengths.

[0136] Since each wavelength supports a certain bandwidth of data transmission, the bandwidth is directly proportional to the number of wavelengths. The more available wavelengths, the greater the available bandwidth for transmission. Allocating available wavelengths means adjusting the available bandwidth corresponding to each ring network. By allocating more available wavelengths to the ring networks corresponding to sub-networks with higher traffic loads, these sub-networks can have greater available bandwidth, which helps reduce latency within the sub-networks. When traffic distribution is predicted based on the first and second characteristic data, the available wavelengths of the ring networks corresponding to the sub-networks can be adjusted in advance to ensure a stable and optimal user experience.

[0137] Optionally, each wavelength in the metropolitan core network corresponds to a bandwidth of 400Gb / s, 800Gb / s, or 1.6Tb / s, while the bandwidth corresponding to each wavelength in the metropolitan aggregation network can be less than or equal to the bandwidth corresponding to the wavelength in the core network.

[0138] See you again Figure 6When the traffic load of the subnetwork corresponding to region 1 increases while the traffic load of the subnetwork corresponding to region 2 decreases, the available wavelengths of ring network 2 can be increased, while the available wavelengths of ring network 3 can be decreased.

[0139] Figure 7 yes Figure 5 A schematic diagram illustrating another application scenario of the method shown. (For example...) Figure 7 As shown, the optical transport network is a metropolitan area network (MAN), comprising a metropolitan area aggregation network and a metropolitan area access network. The metropolitan area aggregation network includes four aggregation devices (all OTN devices), forming a mesh topology. Aggregation devices 1 and 2 are located in three ring networks, respectively. The three ring networks associated with the same aggregation device share multiple wavelengths in a wavelength pool. For example, if ring networks 1-3 are associated with aggregation device 1, and the total traffic load of the sub-networks connected to ring network 3 increases, more wavelengths are allocated to ring network 3, i.e., the number of available wavelengths in ring network 3 is increased (e.g., selecting unallocated wavelengths from the wavelength pool or allocating available wavelengths from other ring networks to ring network 3). Conversely, if the total traffic load of the sub-networks connected to ring network 3 decreases, the number of available wavelengths in ring network 3 is reduced (e.g., at least one available wavelength in ring network 3 is released as an unallocated wavelength or allocated to other ring networks).

[0140] Figure 7 The example provided illustrates a fixed access network as the target network, where all subnetworks of the fixed access network are point-to-multipoint (P2MP). Figure 7 In this context, an architecture where one aggregation device is located in multiple ring networks can be called a metropolitan P2MP network.

[0141] In this embodiment, since the collaborative network controller can simultaneously acquire the first feature data and the second feature data, it can perform overall planning and dynamic adjustment of computing and communication resources based on the first feature data and the second feature data, which is beneficial to improving the user experience.

[0142] In other examples, network resources include communication resources, which include transmission paths within the OTN. Here, a transmission path can refer to an optical fiber link. An optical fiber link includes a single optical fiber or multiple optical fibers connected sequentially, and may also include functional devices such as optical amplifiers. Step 503 includes: selecting a transmission path for the target sub-network with a bandwidth margin greater than the bandwidth required for the traffic load, based on the traffic load of the target sub-network. Here, the target sub-network is any sub-network within the target network.

[0143] Combination Figure 1 This section provides an example illustrating resource scheduling when communication resources include transmission paths within an OTN. For instance... Figure 1As shown, if the traffic distribution determined based on the first feature data indicates a sudden increase in traffic in the fronthaul network where AAU3-5 is located, the traffic that needs to be transmitted through the metropolitan area access network to DU2+CU may exceed the remaining bandwidth of the transmission path pre-configured for that fronthaul network in the OTN. In this case, the mobile network controller can send a bandwidth request carrying indication information to the cooperating network controller in advance. After receiving the bandwidth request, the cooperating network controller can select a transmission path with a bandwidth margin greater than the bandwidth required by the traffic corresponding to that fronthaul network.

[0144] In practical applications, the transmission path selected for the target sub-network may be a transmission path pre-configured for the fronthaul network within the OTN or an alternative transmission path. This is because the bandwidth available to the fronthaul network is limited within the pre-configured transmission path. In this case, the cooperative network controller can configure a larger bandwidth for the fronthaul network.

[0145] In one possible implementation, the OTN includes a first transmission path and a second transmission path, where the transmission rate of the optical fiber used in the first transmission path is greater than that of the optical fiber used in the second transmission path. For example, the first transmission path uses conventional optical fiber, while the second transmission optical fiber uses hollow-core fiber (HCF). Unlike conventional optical fibers, which use solid materials such as glass as the core, HCF has a hollow region composed of air or other gases at its center, where light primarily propagates. Compared to conventional optical fibers, HCF has lower latency, potentially reducing it by more than one-third, thus better meeting the demands of high-speed transmission.

[0146] In this scenario, when the bandwidth and computing resources allocated to the target sub-network have reached their limits, the transmission path containing the hollow-core fiber can be selected to further reduce transmission latency. That is, in addition to selecting a transmission path based on the target sub-network's bandwidth requirements, a transmission path can also be selected based on the target sub-network's latency requirements.

[0147] Optionally, the latency requirement of the target sub-network can be the latency requirement for scheduling computing resources to complete computing tasks.

[0148] It should be noted that, Figure 5 The method shown and Figure 3 The methods shown can be used in combination or individually.

[0149] The above explanation uses the example of deploying shared computing resources between mobile and fixed access networks in network-side equipment of the access network. In other embodiments, the shared computing resources between mobile and fixed access networks can be deployed in user-side equipment of the access network (e.g., in the main FTTR device) or in the metropolitan area network or core network (e.g., in EDC, RDC, or CDC). For example, when global data analysis is required and / or a large amount of computing resources are needed, the shared computing resources between mobile and fixed access networks can be deployed in the core network. The deployment locations of CDC, RDC, and EDC can be found in [reference needed]. Figure 6 .

[0150] The following section explains the deployment of shared computing resources between mobile and fixed access networks in the core network.

[0151] Figure 8 This is a schematic diagram illustrating another application scenario of the resource scheduling method provided in the embodiments of this application. Figure 8 In this context, root cause analysis systems are used to analyze data from mobile networks to determine the root cause of problems (such as malfunctions). Figure 8 As shown, when the root cause system is deployed near the edge, although the transmission latency is low, the computing resources are insufficient to support the large computing demands of the root cause system. When the root cause system is deployed in the core network, computing resources are sufficient, but because the transmission between the core network and the access network devices connected to the data acquisition equipment is multi-hop, the transmission latency is greater than 10ms, which cannot meet the latency requirements. However, when using... Figure 5 Under the resource scheduling method shown, the transmission path between the root cause system and the access network device connected to the data acquisition device can be optimized, so that the transmission between the root cause system and the access network device connected to the data acquisition device is single-hop transmission. This can reduce the transmission latency to less than 5ms, thus meeting the latency requirements.

[0152] It should be noted that in practical applications, some of the computing resources shared by the mobile network and the fixed access network can be deployed in the core network, and the other part can be deployed in the access network.

[0153] For rapidly mobile terminals (such as drones) that access mobile networks, their small size necessitates smaller battery capacities and weaker computing power from their chips. Therefore, computational tasks (such as image processing) need to be offloaded to nearby computing resources. Furthermore, their high speed makes them highly sensitive to latency. In this context, low latency requirements must be met for these terminals.

[0154] Optionally, the method may further include the following steps:

[0155] The first step is to obtain the end-to-end latency requirements of the target terminal;

[0156] The second step is to determine the transmission latency required for the migration of the target terminal's computing data between computing resources based on the target terminal's access location.

[0157] The third step is to allocate target computing resources to the target terminal based on the transmission latency and end-to-end latency requirements. The target computing resources are used to provide services to the target terminal.

[0158] Optionally, the end-to-end latency requirement can refer to the latency between the target terminal transmitting data to the computing resources and the computing resources returning the result to the drone after completing the computation task. When the target terminal is a drone, the end-to-end latency requirement can be 0.1ms-20ms. For example, the end-to-end latency requirement can be 0.1ms, 1ms, 5ms, 10ms, or 20ms, etc.

[0159] Optionally, in this second step, the access location of the target terminal can be represented by the AAU to which the target terminal is connected.

[0160] The second step may include: when the access location of the target terminal changes from the first AAU to the second AAU, if the computing resources available to the first AAU and the second AAU are located in the same access network device (which may be the aforementioned DU, CU, or OLT), then the transmission delay required for the migration of the target terminal's computing data between the computing resources is determined to be 0; if the computing resources available to the first AAU are located in the first access network device and the computing resources available to the second AAU are located in the second access network device, then the transmission delay required for the migration of the target terminal's computing data between the computing resources is determined based on the available transmission bandwidth between the first access network device and the second access network device and the amount of data to be migrated. For example, the transmission delay is obtained by dividing the available transmission bandwidth by the amount of data to be migrated.

[0161] In this third step, the difference between the end-to-end latency requirement and the transmission latency can be calculated first, and then the target computing resources can be allocated to the target terminal based on this difference. The larger the difference, the longer the available computing time, and the less target computing resources are available; conversely, the smaller the difference, the shorter the available computing time, and the more target computing resources are available.

[0162] In this way, by determining the access location of the target terminal, the location of the computing resources used by the target terminal (e.g., access network equipment) can be determined. When the location of the computing resources used by the target terminal changes, the transmission latency required for the migration of computing data between computing resources is determined first. Then, based on the transmission latency and end-to-end latency requirements, sufficient computing resources are allocated to the target terminal, which helps ensure that the target terminal's computing tasks are completed within the end-to-end latency requirements. Thus, during the movement of the target terminal, computation can be completed within the end-to-end latency requirements, thereby guaranteeing the target terminal's service needs.

[0163] Optionally, the method further includes: when the computing resources in the computing network are insufficient to meet the total load of the fixed access network and the mobile network, calling other computing resources outside the computing network to provide services to the fixed access network and / or the mobile network.

[0164] The other computing resources can be, for example, the computing resources of a data center or the computing resources in the main FTTR device. Data centers include, but are not limited to, training data centers dedicated to training various models.

[0165] By utilizing other computing resources, communication quality for users in fixed and mobile networks can be guaranteed even in extreme circumstances. Extreme circumstances refer to a sudden increase in traffic load on fixed and mobile networks in a particular location, exceeding the capacity of their shared computing resources. For example, during peak tourist season, a large number of users gather in a tourist city and access the network through that city's fixed and mobile networks.

[0166] Optionally, the step of calling upon other computing resources to provide services to the fixed access network and / or mobile network includes: sending a request to the control device of the other computing resources, the request being for requesting the use of the other computing resources; configuring dedicated communication resources for the request from the computing network resources, the dedicated communication resources being used to transmit computational data related to the request. The dedicated communication resources include dedicated bandwidth in the OTN.

[0167] By allocating dedicated communication resources to the computational data related to the request, service latency can be reduced when other computing resources are invoked.

[0168] In practice, latency-insensitive computational tasks can be processed using other computing resources. This way, when those resources are available in the training data center, the latency-insensitive task can utilize idle computing power during training. For the training data center, the execution of high-intensity tasks such as model training exhibits peak and trough periods. Power consumption is extremely high during peak periods and low during trough periods. Such large fluctuations in power consumption can damage the power supply station. Therefore, utilizing the training data center's computing resources for latency-insensitive computational tasks can also help stabilize the power supply to the training data center.

[0169] Figure 9 This is a schematic diagram of the structure of a resource scheduling device provided in an embodiment of this application. Figure 9 As shown, the resource scheduling device 900 includes an acquisition module 901 and a scheduling module 902. The acquisition module 901 acquires first characteristic data of a mobile network and second characteristic data of a fixed access network. The first characteristic data characterizes the load of the mobile network, and the second characteristic data characterizes the load of the fixed access network. The mobile network and the fixed access network share computing network resources. The scheduling module 902, based on the first characteristic data and the second characteristic data, schedules first computing network resources to provide services to the mobile network and schedules second computing network resources to provide services to the fixed access network. The first computing network resources and the second computing network resources are different computing network resources among the computing network resources.

[0170] Optionally, the computing network resources include computing power resources, and the scheduling module 902 is used to schedule the first computing network resources and the second computing network resources using one or more of the following methods:

[0171] Based on the first feature data, a first trend of load change of the mobile network within the target time period is determined; when the first trend is an increase in load, the computing power resources in the first computing network resources are increased, or when the first trend is a decrease in load, the computing power resources in the first computing network resources are decreased.

[0172] Based on the second feature data, determine the second trend of load change of the fixed access network within the target time period; when the second trend is that the load increases, increase the computing power resources in the second computing network resources, or when the second trend is that the load decreases, decrease the computing power resources in the second computing network resources.

[0173] The computing power resources for communication purposes are determined to meet the communication-related accelerated computing needs in the mobile network and the fixed access network. At least a portion of the remaining computing power resources are scheduled for non-communication-related computing needs in the mobile network and the fixed access network based on the first feature data and the second feature data. The remaining computing power resources are the computing network resources shared by the mobile network and the fixed access network, excluding the computing power resources for communication purposes.

[0174] Optionally, the network resources include communication resources. The scheduling module 902 includes a determining submodule and an allocating submodule. The determining submodule is used to determine the traffic distribution in the target network based on the first feature data and the second feature data. The traffic distribution is used to indicate the traffic load of different subnetworks in the target network, and the coverage areas of different subnetworks are different. The allocating submodule is used to allocate communication resources to each subnetwork in the target network based on the traffic distribution. The target network is either the fixed access network or the mobile network.

[0175] Optionally, the communication resources include a wavelength pool of an optical transport network, the optical transport network including a first ring network and a second ring network, the first ring network and the second ring network being connected to the same optical transport network equipment, the available wavelengths of the first ring network and the second ring network both belonging to the wavelength pool, and the available wavelengths of the first ring network and the second ring network not overlapping; the allocation submodule is used to allocate a first available wavelength to the first ring network and a second available wavelength to the second ring network when the traffic distribution indicates that the total traffic load of each sub-network connected to the first ring network is greater than the total traffic load of each sub-network connected to the second ring network, and the number of the first available wavelengths is greater than the number of the second available wavelengths.

[0176] Optionally, the communication resources include the transmission path of the optical transport network; the allocation submodule is used to select a transmission path with a bandwidth margin greater than the bandwidth required by the traffic load of the target subnetwork.

[0177] Optionally, the determining submodule is used to predict the traffic distribution in the target network based on the first feature data and the second feature data.

[0178] Optionally, the acquisition module 901 is further configured to acquire the end-to-end latency requirement of the target terminal; the scheduling module 902 is further configured to determine the transmission latency required for the migration of the target terminal's computing data between computing resources based on the access location of the target terminal; and allocate target computing resources to the target terminal based on the transmission latency and the end-to-end latency requirement, wherein the target computing resources are used to provide services to the target terminal.

[0179] Optionally, the scheduling module 902 is further configured to call other computing resources outside the computing network resources to provide services to the fixed access network and / or the mobile network when the computing resources in the computing network resources are insufficient to meet the load corresponding to the fixed access network and the mobile network.

[0180] Optionally, the scheduling module 902 is configured to invoke other computing resources outside the computing network resources to provide services to the fixed access network and / or mobile network in the following manner: sending a request to the control device of the other computing resources, the request being used to request the use of the other computing resources; configuring dedicated communication resources for the request from the computing network resources, the dedicated communication resources being used to transmit computing data related to the request.

[0181] Optionally, the first feature data includes at least one of the following: first traffic data, which represents the variation of traffic load in the mobile network over time and / or space; and first connection data, which represents the number of connections in the mobile network. The second feature data includes at least one of the following: second traffic data, which represents the variation of traffic load in the fixed access network over time and / or space; and second connection data, which represents the number of connections in the fixed access network.

[0182] Optionally, the load of the mobile network and the load of the fixed access network are negatively correlated during the user's active time period.

[0183] It should be noted that the resource scheduling device provided in the above embodiments is only illustrated by the division of the above functional modules when performing resource scheduling. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the resource scheduling device and the resource scheduling method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0184] The descriptions of the processes corresponding to the above-mentioned figures each have their own emphasis. For parts of a process that are not described in detail, please refer to the relevant descriptions of other processes.

[0185] This application also provides a computer device 1000. For example... Figure 10 As shown, the computer device 1000 includes a bus 1002, a processor 1004, a memory 1006, and a communication interface 1008. The processor 1004, the memory 1006, and the communication interface 1008 communicate with each other via the bus 1002. It should be understood that this application does not limit the number of processors and memories in the computer device 1000.

[0186] Bus 1002 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 10 The bus 1002 may be represented by a single line, but this does not mean that there is only one bus or one type of bus. The bus 1002 may include a path for transmitting information between various components of the computer device 1000 (e.g., memory 1006, processor 1004, communication interface 1008).

[0187] The processor 1004 may include any one or more processors such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0188] The memory 1006 may include volatile memory, such as random access memory (RAM). The processor 1004 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0189] The memory 1006 stores executable program code, and the processor 1004 executes the executable program code to implement the functions of the aforementioned modules, thereby implementing the resource scheduling method. That is, the memory 1006 stores instructions for executing the resource scheduling method.

[0190] The communication interface 1008 uses transceiver modules such as, but not limited to, network interface cards and transceivers to enable communication between the computer device 1000 and other devices or communication networks.

[0191] This application also provides a computer program product containing instructions. The computer program product may be a software or program product containing instructions, capable of running on a computer device or stored on any usable medium. When the computer program product is run on at least one computer device, it causes the at least one computer device to perform the aforementioned resource scheduling method.

[0192] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computer device can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computer device to execute the aforementioned resource scheduling method.

[0193] This application also provides a chip. The chip includes a processor and a communication interface, the communication interface being connected to the processor; the processor is used to execute instructions so that the chip performs the aforementioned resource scheduling method.

[0194] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “first,” “second,” “third,” and similar terms used in this patent application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms “an” or “a” and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms “comprising” or “including” and similar terms mean that the elements or objects preceding “comprising” or “including” encompass the elements or objects listed following “comprising” or “including” and their equivalents, and do not exclude other elements or objects. The “multiple” mentioned in the embodiments of this application refers to two or more. “A and / or B” includes the following three cases: A, B, or A and B.

[0195] The above is merely one embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A resource scheduling method, characterized in that, include: Acquire first characteristic data of a mobile network and second characteristic data of a fixed access network. The first characteristic data is used to characterize the load of the mobile network, and the second characteristic data is used to characterize the load of the fixed access network. The mobile network and the fixed access network share computing network resources. Based on the first feature data and the second feature data, a first computing network resource is scheduled to provide services to the mobile network, and a second computing network resource is scheduled to provide services to the fixed access network. The first computing network resource and the second computing network resource are different computing network resources among the computing network resources.

2. The method according to claim 1, characterized in that, The computing network resources include computing power resources. The step of scheduling first computing network resources to provide services to the mobile network and scheduling second computing network resources to provide services to the fixed access network based on the first feature data and the second feature data includes one or more of the following methods: Based on the first feature data, a first trend of load change of the mobile network within the target time period is determined; when the first trend is an increase in load, the computing power resources in the first computing network resources are increased, or when the first trend is a decrease in load, the computing power resources in the first computing network resources are decreased. Based on the second feature data, determine the second trend of load change of the fixed access network within the target time period; when the second trend is that the load increases, increase the computing power resources in the second computing network resources, or when the second trend is that the load decreases, decrease the computing power resources in the second computing network resources. The computing power resources for communication purposes are determined to meet the communication-related accelerated computing needs in the mobile network and the fixed access network. At least a portion of the remaining computing power resources are scheduled for non-communication-related computing needs in the mobile network and the fixed access network based on the first feature data and the second feature data. The remaining computing power resources are the computing power resources in the computing network other than the computing power resources for communication purposes.

3. The method according to claim 1, characterized in that, The step of scheduling a first computing network resource to provide services to the mobile network and scheduling a second computing network resource to provide services to the fixed access network based on the first feature data and the second feature data includes: Based on the first feature data and the second feature data, the traffic distribution in the target network is determined. The traffic distribution is used to indicate the traffic load of different sub-networks in the target network, and the coverage areas of the different sub-networks are different. Based on the traffic distribution, schedule computing resources for each sub-network in the target network; The target network is either the fixed access network or the mobile network.

4. The method according to claim 3, characterized in that, The computing network resources include communication resources, the communication resources include wavelengths of the optical transport network, the optical transport network includes a first ring network and a second ring network, the first ring network and the second ring network are connected to the same optical transport network equipment, the available wavelengths of the first ring network and the second ring network both belong to the wavelength pool, and the available wavelengths of the first ring network and the second ring network do not overlap. The step of allocating communication resources to each sub-network in the target network according to the traffic distribution includes: When the traffic distribution indicates that the total traffic load of each sub-network connected to the first ring network is greater than the total traffic load of each sub-network connected to the second ring network, a first available wavelength is allocated to the first ring network and a second available wavelength is allocated to the second ring network, and the number of the first available wavelengths is greater than the number of the second available wavelengths.

5. The method according to claim 3, characterized in that, The communication resources include transmission paths in optical transport networks; The step of allocating communication resources to each sub-network in the target network according to the traffic distribution includes: Based on the traffic load of the target sub-network, select a transmission path for the target sub-network with a bandwidth margin greater than the bandwidth required by the traffic load; The target sub-network can be any sub-network within the target network.

6. The method according to claim 4 or 5, characterized in that, Determining the traffic distribution in the target network based on the first feature data and the second feature data includes: Based on the first feature data and the second feature data, predict the traffic distribution in the target network.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Obtain the end-to-end latency requirements of the target terminal; Based on the access location of the target terminal, determine the transmission delay required for the migration of the target terminal's computing data between computing resources; Based on the transmission delay and the end-to-end delay requirement, target computing resources are allocated to the target terminal, and the target computing resources are used to provide services to the target terminal.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: When the computing resources in the computing network are insufficient to meet the load of the fixed access network and the mobile network, other computing resources outside the computing network are called to provide services to the fixed access network and / or the mobile network.

9. The method according to claim 8, characterized in that, The provision of services to the fixed access network and / or mobile network by utilizing computing resources other than those in the computing network includes: Send a request to the control device of the other computing resources, the request being used to request the use of the other computing resources; Configure dedicated communication resources for the request from the computing network resources. The dedicated communication resources are used to transmit computing data related to the request.

10. The method according to any one of claims 1 to 9, characterized in that, The first feature data includes at least one of the following: First traffic data, which is used to represent the change of traffic load in the mobile network over time and / or space; and First connection data, which is used to represent the number of connections in the mobile network; The second feature data includes at least one of the following: Second traffic data, which is used to represent the change of traffic load in the fixed access network over time and / or space; and The second connection data is used to represent the number of connections in the fixed access network.

11. The method according to any one of claims 1 to 10, characterized in that, The load on the mobile network and the load on the fixed access network are negatively correlated during periods of user activity.

12. The method according to any one of claims 1 to 11, characterized in that, The computing resources shared by the mobile network and the fixed access network are located in a first network device, which belongs to the mobile network; or... The computing resources shared by the mobile network and the fixed access network are set in the second network device, which belongs to the fixed access network. A portion of the computing resources shared by the mobile network and the fixed access network are located in the first network device, and another portion is located in the second network device. The first network device belongs to the mobile network, and the second network device belongs to the fixed access network.

13. The method according to claim 12, characterized in that, Both the first network device and the second network device are access network devices.

14. A resource scheduling device, characterized in that, include: The acquisition module is used to acquire first feature data of the mobile network and second feature data of the fixed access network. The first feature data is used to characterize the load of the mobile network, and the second feature data is used to characterize the load of the fixed access network. The mobile network and the fixed access network share computing network resources. The scheduling module is configured to, based on the first feature data and the second feature data, schedule a first computing network resource from the computing network resources to provide services to the mobile network, and schedule a second computing network resource to provide services to the fixed access network.

15. A computer device, characterized in that, The method includes a processor and a memory, wherein the memory stores program code; the processor is configured to read and execute the program code stored in the memory to implement the method as claimed in any one of claims 1 to 13.

16. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a software program that, when read and executed by one or more processors, can implement the method as described in any one of claims 1 to 13.

17. A communication system, characterized in that, The communication system includes a cooperative network controller, multiple access network devices, and multiple transport network devices; the multiple access network devices include wireless access network devices and fixed access network devices; the wireless access network devices belong to a mobile network, the fixed access network devices belong to a fixed access network, the fixed access network and the mobile network share computing network resources, and the cooperative network controller is used to implement the method as described in any one of claims 1 to 13.

Images