A distributed cloud-oriented computing power resource cross-domain collaborative scheduling method and system

Abstract

The application discloses a kind of distributed cloud-oriented computing power resource cross-domain collaborative scheduling method and system, obtain the user load request of different regions different cities;According to cold-temperature-hot classification mechanism, user load is divided;Based on the mathematical optimization model of structure construction computing power resource global scheduling, combined with decomposition coordination mechanism, build the computing power resource cross-domain collaborative scheduling model suitable for distributed cloud computing;The computing power resource cross-domain collaborative scheduling model is solved using alternating direction multiplier method, and the optimal allocation scheme of computing power resource is obtained, and according to the optimal allocation scheme, user load is scheduled to the data center of corresponding region for processing.The application can realize efficient, flexible, economic configuration of global computing power resource, while guaranteeing the data privacy security of each region, improve computing efficiency.

Description

Technical Field

[0001] This invention belongs to the field of cloud computing technology, specifically relating to a cross-domain collaborative scheduling method and system for computing resources in distributed clouds. Background Technology

[0002] The uneven distribution of computing resources across different regions has exacerbated the problem. Addressing the diverse needs of users and facilitating flexible scheduling of various computing resources across the entire region remains a critical challenge for traditional resource scheduling models. Against this backdrop, a comprehensive computing resource scheduling strategy that considers multiple factors is of great significance for promoting my country's digital economy and implementing sustainable development principles. On one hand, such a strategy allows users to apply for computing services on demand, much like purchasing water, electricity, and gas, while ensuring a positive user experience and achieving efficient, flexible, and economical allocation of computing resources. On the other hand, by focusing on computing resources as a demand driver and viewing data centers as the primary energy consumers, it fully leverages the potential of renewable resources in different regions for multi-energy conversion within data centers, promoting the coordinated operation of computing and energy networks. This achieves a win-win situation of economic benefits and energy reduction for data centers, further contributing to the achievement of dual-carbon goals.

[0003] However, the global computing resource scheduling process involves many complex factors such as data transmission latency, network bandwidth, and cross-domain transmission costs. Even slight fluctuations can significantly impact the performance of computing resource scheduling strategies. Moreover, traditional centralized scheduling methods have made data privacy and security issues increasingly prominent. Therefore, this invention presents a cross-domain collaborative scheduling method and system for computing resources in distributed clouds, aiming to provide technical support for scientific decision-making in global computing resource scheduling and promote the low-carbon and sustainable development of my country's digital economy. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a cross-domain collaborative scheduling method and system for computing resources in distributed cloud environments, which addresses the shortcomings of the prior art. This method is used to solve the problem of flexible scheduling of various computing resources across the entire domain under different user needs. It achieves privacy protection of critical computing information in a distributed, multi-entity collaborative optimization manner, while providing a faster computing power scheduling response and enhancing the stability of the entire scheduling system.

[0005] The present invention adopts the following technical solution:

[0006] A method for cross-domain collaborative scheduling of computing resources for distributed clouds includes the following steps:

[0007] S1. Obtain user load requests from different regions and cities;

[0008] S2. Divide the user load obtained in step S1 according to the cold-warm-hot classification mechanism;

[0009] S3. Based on the structure obtained in step S2, construct a mathematical optimization model for global scheduling of computing resources;

[0010] S4. Based on the mathematical optimization model obtained in step S3, and combined with the decomposition and coordination mechanism, construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing.

[0011] S5. Solve the cross-domain collaborative scheduling model of computing resources in step S4 using the alternating direction multiplier method to obtain the optimal allocation scheme of computing resources. Based on the optimal allocation scheme, schedule user loads to the data centers in the corresponding regions for processing.

[0012] Preferably, in step S1, the user load request includes:

[0013] Computing resources, that is, the amount of computing power resources required to complete a user request;

[0014] Latency requirement refers to the maximum acceptable computational latency from responding to a user's request to completing that request.

[0015] Preferably, step S2 specifically includes:

[0016] S201. Estimate the transmission time based on the transmission distance between different regions and cities, divide the delay range according to the magnitude of the transmission time, and design a corresponding cold-temperature-heat load classification mechanism.

[0017] S202. Based on the user's latency requirements, classify the user load according to the cold-warm-hot division mechanism: users with cold loads are allowed to request cross-domain scheduling, users with warm loads are only allowed to request intra-domain scheduling, and users with hot loads are only allowed to request local calculation.

[0018] More preferably, in step S201, the corresponding cold-temperature-heat load classification mechanism is specifically as follows:

[0019] Thermal loads are time-delay sensitive and require rapid response; they can only be calculated locally and cannot be scheduled.

[0020] Cold loads accept long time-span responses, making them suitable for cross-domain scheduling;

[0021] Temperature load, which falls between heat load and cold load, accepts short-term inter-period responses and is capable of intra-domain scheduling.

[0022] Preferably, in step S3, the objective function of the mathematical optimization model for global scheduling of computing resources is as follows:

[0023] min C=C1+C2+C2

[0024] Where C1 is the cross-domain transmission cost, C2 is the intra-domain transmission cost, and C3 is the data center energy consumption cost.

[0025] More preferably, the constraints are as follows:

[0026] (1) Supply and demand balance constraint

[0027] This indicates that the amount of temperature load is equal before and after the transfer in zone i, as detailed below:

[0028]

[0029] in, Let M be the amount of temperature load transferred from city m to city n in region i. The original temperature load for region i; the total cooling load remains the same before and after the transfer, as detailed below:

[0030]

[0031] in, Let M be the amount of cooling load transferred from city m in region i to city n in region j. The original cooling load of region i;

[0032] (2) Bandwidth capacity constraints

[0033] The local area network bandwidth capacity constraints are as follows:

[0034]

[0035]

[0036] The bandwidth capacity constraints of the intranet are as follows:

[0037]

[0038]

[0039] Where α / β is the cold / warm load bandwidth conversion factor, This represents the upper limit of bandwidth capacity for line (i,j) in the regional network. This represents the upper limit of bandwidth capacity for intranet lines within region i.

[0040] (3) Resource capacity constraints

[0041] Resource capacity constraints mainly refer to the fact that the number of cold / warm / hot loads carried by each data center after scheduling does not exceed its capacity limit:

[0042]

[0043]

[0044] in, The heat load of data center m in region i. This represents the upper limit of the operating load capacity of data center m in region i;

[0045] (4) Constraints on the quantity of cooling / warming loads

[0046] The number of cold / warm load transfers for each data center should not exceed its local requested load:

[0047]

[0048]

[0049] in, These refer to the number of cold / warm loads requested locally in data center m of region i, respectively.

[0050] Preferably, in step S4, the objective function of the cross-domain collaborative scheduling model for computing resources is:

[0051]

[0052] in, The transmission cost on the local area network calculated for each region. For the transmission cost on the intranet, I represents the energy consumption cost corresponding to the actual operating load of each region, where I is the set of scheduling regions.

[0053] More preferably, the constraints include:

[0054] (1) Consensus constraints

[0055]

[0056]

[0057] Among them, z ij As an auxiliary variable, x represents the actual amount of cooling load transferred between regions i and j. i [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system for region i. j [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system in region j.

[0058] (2) Cross-regional cooling load transfer constraints

[0059]

[0060]

[0061] in, N represents the number of cooling loads entering / leaving city m in zone i, respectively. i Let i be a row vector of all 1s. x is a unit vector where the i-th element is 1 and all other elements are 0. i J is the cold load transfer matrix calculated by the regional i-cloud data center scheduling system. i This is the set of cities where the data center nodes in region i are located.

[0062] (3) Intra-domain cooling load transfer constraints

[0063]

[0064] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial number of cooling loads for data center m in region i.

[0065] (4) Intra-domain temperature load transfer constraints

[0066]

[0067] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial temperature load for data center m in region i;

[0068] (5) Resource capacity constraints

[0069]

[0070]

[0071] in, The initial temperature load for data center m in region i. Let m be the initial heat load of data center m in region i. The number of cold loads transferred from data center n in region i to data center m in region i. The number of temperature loads transferred from data center n in region i to data center m in region i. Q represents the number of temperature loads transferred from data center m in region i to data center n in region i. im The total number of loads running after scheduling in region i data center m;

[0072] (6) Bandwidth capacity constraints

[0073]

[0074]

[0075]

[0076] in, Let α be the upper limit of the bandwidth capacity of the regional network line (i,j), α be the cold load bandwidth conversion factor, and b be the lower limit of the bandwidth capacity. im β is the upper limit of intra-domain network bandwidth capacity from the central city of region i to the data center m, and β is the temperature load bandwidth conversion coefficient;

[0077] (7) Cooling / warming load quantity constraints

[0078]

[0079]

[0080]

[0081]

[0082] in, Let J be the initial cold load number of data center m in region i, I be the set of scheduling regions, and J be the initial cold load number of data center m in region i. i This refers to the set of cities where the data center nodes in region i are located. The number of temperature loads transferred from data center m in region i to data center n in region i. Let be the initial temperature load number for data center m in region i.

[0083] Secondly, embodiments of the present invention provide a cross-domain collaborative scheduling system for computing resources in distributed clouds, comprising:

[0084] The data module retrieves user load requests from different regions and cities.

[0085] The user load obtained from the data modules is divided into modules according to a cold-warm-hot classification mechanism;

[0086] Build modules, and construct a mathematical optimization model for global scheduling of computing resources based on the structure obtained from the module division;

[0087] The optimization module, based on the mathematical optimization model obtained from the construction module, combines a decomposition and coordination mechanism to construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing.

[0088] The scheduling module uses the alternating direction multiplier method to solve the cross-domain collaborative scheduling model of computing resources in the optimization module, and obtains the optimal allocation scheme of computing resources.

[0089] Preferably, in the optimization module, the objective function of the cross-domain collaborative scheduling model for computing resources is:

[0090]

[0091] in, The transmission cost on the local area network calculated for each region. For the transmission cost on the intranet, Energy consumption cost corresponding to the actual operating load of each region;

[0092] The constraints include:

[0093] (1) Consensus constraints

[0094]

[0095]

[0096] Among them, z ij As an auxiliary variable, x represents the actual amount of cooling load transferred between regions i and j. i [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system for region i. j [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system in region j.

[0097] (2) Cross-regional cooling load transfer constraints

[0098]

[0099]

[0100] in, N represents the number of cooling loads entering / leaving city m in zone i, respectively. i Let i be a row vector of all 1s. x is a unit vector where the i-th element is 1 and all other elements are 0. i J is the cold load transfer matrix calculated by the regional i-cloud data center scheduling system. i This is the set of cities where the data center nodes in region i are located.

[0101] (3) Intra-domain cooling load transfer constraints

[0102]

[0103] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial number of cooling loads for data center m in region i.

[0104] (4) Intra-domain temperature load transfer constraints

[0105]

[0106] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial temperature load for data center m in region i;

[0107] (5) Resource capacity constraints

[0108]

[0109]

[0110] in, The initial temperature load for data center m in region i. Let m be the initial heat load of data center m in region i. The number of cold loads transferred from data center n in region i to data center m in region i. The number of temperature loads transferred from data center n in region i to data center m in region i. Q represents the number of temperature loads transferred from data center m in region i to data center n in region i. im The total number of loads running after scheduling in region i data center m;

[0111] (6) Bandwidth capacity constraints

[0112]

[0113]

[0114]

[0115] in, Let α be the upper limit of the bandwidth capacity of the regional network line (i,j), α be the cold load bandwidth conversion factor, and b be the lower limit of the bandwidth capacity. im β is the upper limit of intra-domain network bandwidth capacity from the central city of region i to the data center m, and β is the temperature load bandwidth conversion coefficient;

[0116] (7) Cooling / warming load quantity constraints

[0117]

[0118]

[0119]

[0120]

[0121] in, Let J be the initial cold load number of data center m in region i, I be the set of scheduling regions, and J be the initial cold load number of data center m in region i.i This refers to the set of cities where the data center nodes in region i are located. The number of temperature loads transferred from data center m in region i to data center n in region i. Let be the initial temperature load number for data center m in region i.

[0122] Compared with the prior art, the present invention has at least the following beneficial effects:

[0123] A cross-domain collaborative scheduling method for computing resources in distributed clouds breaks away from the traditional structure of centralized computing scheduling systems, providing a solution to the global computing resource scheduling problem under a multi-agent collaborative optimization model. First, based on user latency requirements, user loads are divided into different schedulable ranges using a cold-warm-hot classification mechanism, and corresponding scheduling strategies are adopted. Then, considering factors such as network bandwidth, resource capacity, and transmission costs, with the goal of minimizing operating costs, optimal allocation of computing resources is achieved through information exchange between the scheduling systems of cloud data centers in different regions, in a collaborative optimization manner. While ensuring the privacy of critical computing information, the large-scale optimization problem is rationally decomposed according to the physical logical structure, effectively improving the implementation efficiency of cross-domain computing in distributed clouds. In engineering applications, relevant technicians can formulate appropriate computing scheduling strategies based on user computing power requests in different regions using the method and system provided by this invention, satisfying various user requirements while minimizing the total operating cost of the computing system within a given time window. This invention relates to a collaborative optimization problem of global computing resource scheduling for engineering needs. It can effectively improve the economy and overall efficiency of the global scheduling system, help data centers achieve low-carbon transformation, and provide a practical and effective solution for the flexible scheduling of global computing resources.

[0124] Furthermore, to ensure the allocation of computing resources before the user request deadline, it is first necessary to estimate the transmission time to different data centers. Combining this with the processing time of different computing tasks, the maximum latency from transmission to deployment and final processing is calculated. Because massive user requests have varying latency requirements, considering them individually would make the entire model cumbersome and complex. Therefore, a cold-warm-hot load classification mechanism was designed. Based on the estimated maximum latency, latency gradients are divided for scheduling to data centers in different regions and cities, and a corresponding cold-warm-hot load classification mechanism is designed. User requests belonging to cold loads are allowed cross-domain scheduling, user requests belonging to warm loads are only allowed intra-domain scheduling, and user requests belonging to hot loads are only allowed local computation. Finally, based on the latency requirements of all users within the time window, user loads are classified according to the cold-warm-hot classification mechanism, and then global computing resource scheduling is performed based on this data. This approach ensures both the quality of service for user requests and the economic efficiency of global scheduling.

[0125] Furthermore, users typically lack an overall understanding of the global deployment and network topology of cloud services. Therefore, they usually decide to purchase computing resources in a specific region based on the resource distribution, pricing, and current usage presented by the cloud vendor. However, this is often not the optimal scheduling solution for service providers, as differentiated user choices can lead to uneven distribution of computing resources across different regions and data centers. The proposed centralized computing resource global scheduling model breaks the constraints of regional services, introducing global scheduling capabilities. Based on optimizing computing costs, latency for specific cloud services and business loads, and communication coupling relationships between applications / application groups, it provides users with the best choice. The specific geographical region to which cloud service resource instances are allocated is dynamically determined by the cloud's intelligent scheduling system. This approach ensures both a good user experience and smooth allocation of computing resources, guaranteeing a uniform distribution of computing resources.

[0126] Furthermore, the distributed computing resource cross-domain collaborative scheduling system, transformed from the centralized computing resource global scheduling model, breaks the traditional structure of previous centralized computing resource scheduling systems and provides a solution to the global computing resource scheduling problem under a multi-entity collaborative optimization model. Each regional cloud scheduling system is responsible for scheduling warm loads within its own region, and negotiates the amount of cold load transfer between different regions by exchanging cold load transfer matrices with neighboring clouds. Neighboring clouds can only know the amount of cold load transfer between different regions through the cold load transfer matrix and cannot obtain the computing power distribution within other regions. While ensuring data privacy and security, this system can allocate computing resources more rationally, accelerate the response speed of the entire system, and improve overall system stability.

[0127] It is understandable that the beneficial effects of the second aspect mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0128] In summary, this invention enables efficient, flexible, and economical allocation of computing resources across the entire domain, while ensuring data privacy and security in each region and improving computing efficiency.

[0129] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0130] Figure 1 This is a schematic diagram of the system of the present invention;

[0131] Figure 2 This is a schematic diagram of the traditional centralized framework of the global computing power resource scheduling framework of the present invention;

[0132] Figure 3 This is a schematic diagram of the method flow of the present invention;

[0133] Figure 4 This is a schematic diagram of a typical scenario of the present invention;

[0134] Figure 5 A schematic diagram of a computer device provided in an embodiment of the present invention;

[0135] Figure 6 This is a block diagram of a chip provided according to an embodiment of the present invention. Detailed Implementation

[0136] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0137] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0138] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0139] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0140] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0141] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0142] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0143] This invention provides a cross-domain collaborative scheduling method for computing resources in distributed cloud environments. Targeting the characteristics of high throughput and low latency in computing requests, it minimizes the total operating cost through a distributed computing resource scheduling system, taking into account constraints such as transmission latency, network bandwidth, and resource capacity.

[0144] Please see Figure 1 Each geographical region bearing computing demands corresponds to at least one cloud data center, equipped with a scheduling system and communication facilities, serving as the input port for local computing loads. Each regional scheduling system, through information exchange with neighboring regional scheduling systems, coordinates the flexible and optimized allocation of computing demands across different regions—a process known as cross-domain collaborative scheduling of computing resources. The specific scheduling process is as follows: First, the information layer acquires computing demand information for each region and uploads it to the corresponding regional cloud. Then, the regional cloud, through bidirectional information exchange with neighboring clouds, generates a scheduling strategy through local optimization decisions and issues it to the corresponding data center. Local interactive information supports the coordinated iteration of scheduling decisions, and an alternating direction multiplier update mechanism helps the cloud scheduling decision approach global optimality. Finally, the data center, based on the received strategy information, utilizes the computing network to achieve cross-domain scheduling of computing resources.

[0145] Please see Figure 2Traditional centralized cloud data center scheduling systems need to handle scheduling requests from all geographical regions. This involves acquiring and comprehensively processing computing resource information from all regions to make optimal global scheduling decisions. Compared to distributed cloud scheduling systems, centralized scheduling faces many challenges. First, the information collection process is complex, requiring extensive coordination and communication. Furthermore, it places extremely high demands on the stability and security of the central scheduler. A failure of the central scheduler will paralyze the entire scheduling system, preventing not only the scheduling of computing resources in various regions but also posing a potential risk of sensitive data leakage.

[0146] The distributed cloud resource scheduling system breaks away from the traditional structure of centralized computing power schedulers, providing a multi-agent collaborative optimization solution for the global scheduling problem of computing power resources. Through information exchange between regional clouds and neighboring clouds, collaborative computing power requests are transferred between regions. While ensuring data privacy and security, computing power resources can be allocated more rationally, thereby effectively reducing the overall system energy consumption and transmission costs. The mechanism by which the proposed scheduling system achieves data privacy protection is illustrated below. Figure 1 In actual operation, since there is no information exchange between Cloud1 and Cloud3, the distribution of computing power between corresponding regions can be shielded to a certain extent. On the other hand, both of them, along with Cloud2 and Cloud4, are neighboring clouds, and can indirectly obtain information through Cloud2 and Cloud4 as intermediaries, thereby promoting the achievement of global optimization scheduling decisions.

[0147] Please see Figure 3 This invention provides a cross-domain collaborative scheduling method for computing resources in distributed clouds, comprising the following steps:

[0148] S1. Obtain user load requests from different regions and cities;

[0149] Computing resources, that is, the amount of computing power resources required to complete a user request;

[0150] Latency requirement refers to the maximum acceptable computational latency from responding to a user's request to completing that request.

[0151] S2. Divide the user load obtained in step S1 according to the cold-warm-hot classification mechanism;

[0152] S201. Estimate the transmission time based on the transmission distance between different regions and cities, divide the delay range according to the transmission time, and design a corresponding cold-temperature-heat load classification mechanism:

[0153] Thermal loads are sensitive to latency, require rapid response, can only be calculated locally, and cannot be scheduled, such as requests for disaster warnings, financial transactions, online video, etc.

[0154] Cold loads have a high tolerance for latency and can accept longer response times, making them suitable for cross-domain scheduling, such as offline simulation, backup and archiving, and other service requests.

[0155] Temperature loads, falling between heat loads and cooling loads, can tolerate short-term, cross-period responses because they can be scheduled within the domain.

[0156] S202. Based on the user's latency requirements, classify the user load according to the cold-warm-hot division mechanism: users with cold loads are allowed to request cross-domain scheduling, users with warm loads are only allowed to request intra-domain scheduling, and users with hot loads are only allowed to request local calculation.

[0157] S3. Based on the cold-warm-hot load partitioning mechanism, construct a mathematical optimization model for global scheduling of computing resources;

[0158] 1) Objective function

[0159] The goal of global computing resource scheduling is to minimize the total system operating cost within a unit time window. The corresponding objective function is expressed as follows:

[0160] The total operating cost of user load within a unit time window includes cross-domain transmission cost C1, intra-domain transmission cost C2, and data center energy consumption cost C3, as detailed below:

[0161] min C=C1+C2+C2 (1)

[0162] Cross-domain transfer costs include intra-domain cold / warm load transfer costs, as well as costs incurred by the intra-domain network portion of cross-domain cold load transfers, as detailed below:

[0163]

[0164] Where I is the set of scheduling regions, and i / j is the index of the scheduling region. For regional network lines<m,i> Transmission cost, in yuan / Gbps / year, B ij For regional network lines<m,i> Bandwidth usage.

[0165] Intra-domain transmission costs are as follows:

[0166]

[0167] Among them, J i Let m be the set of cities within region i, where m is the city number. The transmission cost of intranet lines in region i, in yuan / Gbps / year, b im Bandwidth usage on the network line from the central node of region i to city m

[0168] Data center energy consumption costs, which are the energy costs incurred by the i-zone data center after scheduling to meet computing demands, are as follows:

[0169]

[0170] Where μ is the conversion factor between the number of data center operating loads and the total operating power, γ is the annual conversion factor, and Q im η represents the actual load quantity in city m of region i after scheduling. im Let m be the unit total power cost for city m in region i.

[0171] 2) Constraints

[0172] (1) Supply and demand balance constraint

[0173] The basic principle of cross-domain demand scheduling is that the number of cold / warm loads before and after scheduling is equal.

[0174] Equation (5) indicates that the amount of temperature load is equal before and after the transfer in zone i, as follows:

[0175]

[0176] in, Let M be the amount of temperature load transferred from city m to city n in region i. The original temperature load quantity for region i.

[0177] Equation (6) indicates that the total cooling load is equal before and after the transfer, as follows:

[0178]

[0179] in, Let M be the amount of cooling load transferred from city m in region i to city n in region j. The original cooling load of region i.

[0180] (2) Bandwidth capacity constraints

[0181] Bandwidth capacity constraints mainly refer to the requirement that the number of cold / warm load transfers actually carried by the regional network and intra-domain network during the scheduling process must not exceed their bandwidth capacity limit.

[0182] Equations (7) and (8) represent the bandwidth capacity constraints of the regional network, as detailed below:

[0183]

[0184]

[0185] Equations (9) and (10) represent the bandwidth capacity constraints of the intra-domain network, as detailed below:

[0186]

[0187]

[0188] Where α / β is the cold / warm load bandwidth conversion factor, This represents the upper limit of bandwidth capacity for line (i,j) in the regional network. This represents the upper limit of bandwidth capacity for intranet lines within region i.

[0189] (3) Resource capacity constraints

[0190] Resource capacity constraints mainly refer to the fact that the number of cold / warm / hot loads carried by each data center after scheduling does not exceed its capacity limit:

[0191]

[0192]

[0193] in, The heat load of data center m in region i. This represents the upper limit of the operating load capacity of data center m in region i.

[0194] (4) Constraints on the quantity of cooling / warming loads

[0195] The number of cold / warm load transfers for each data center should not exceed its local requested load:

[0196]

[0197]

[0198] in, These refer to the number of cold / warm loads requested locally in data center m of region i, respectively.

[0199] S4. Based on the mathematical optimization models (1) to (14), and combined with the decomposition and coordination mechanism, a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing is constructed.

[0200] 1) Objective function

[0201] The objective function of the cross-domain collaborative scheduling model for computing resources is expressed as follows:

[0202] The total operating cost of user load within a unit time window is the transmission cost calculated for each region. and energy consumption costs sum:

[0203]

[0204] Transmission costs on the local area network calculated for each region for:

[0205]

[0206] Transmission costs on intranet for:

[0207]

[0208] Energy consumption cost corresponding to the actual operating load of each region for:

[0209]

[0210] Where θ1 / θ2 is the basic unit price of bandwidth for the regional network / internal domain network, in yuan / Gbps / year; x i(n×n) (n=|I|) represents the regional network cold load flow matrix replicated in region i, N i(1×n) It is a one-dimensional vector consisting of the unit price coefficients of line bandwidth in adjacent areas.

[0211] 2) Constraints

[0212] (1) Consensus constraints

[0213] Consensus constraints are the core of the distributed scheduling framework. Because cold load transfers between different regions are interdependent, consensus on neighboring cold load flows is needed to ensure independent operation of each region.

[0214]

[0215] Among them, z ij As an auxiliary variable, it represents the actual amount of cold load transferred between regions i and j.

[0216] (2) Cross-regional cooling load transfer constraints

[0217] For each region, the cross-regional transfer portion of the cooling load and the portion transferred via the intra-regional network are considered separately;

[0218] First, construct the consensus matrix x representing the number of cross-domain transferred cold loads. i The correlation equations between them are as follows: Equation (20) represents the cold load conservation constraint transferred to region i, and Equation (21) represents the cold load conservation constraint transferred out of region i.

[0219]

[0220]

[0221] in, These represent the number of cooling loads entering / leaving city m in zone i, respectively.

[0222] (3) Intra-domain cooling load transfer constraints

[0223] Construct intra-domain cooling load transfer constraints to describe the destination of cooling load transfer within the domain:

[0224]

[0225] in, This represents the number of cold loads transferred from data center m in region i to data center n in region i.

[0226] (4) Intra-domain temperature load transfer constraints

[0227] Construct intra-domain temperature load transfer constraints to describe the destination of temperature load transfer within the domain:

[0228]

[0229] in, This represents the number of cold loads transferred from data center m in region i to data center n in region i.

[0230] (5) Resource capacity constraints

[0231] Resource capacity constraints mainly refer to the fact that the number of cold / warm / heat loads carried by each data center after scheduling does not exceed its capacity limit:

[0232]

[0233] (6) Bandwidth capacity constraints

[0234] Bandwidth capacity constraints mainly refer to the requirement that the number of cold / warm load transfers actually carried by the regional network and intra-domain network during the scheduling process must not exceed their bandwidth capacity limit.

[0235]

[0236]

[0237] (7) Cooling / warming load quantity constraints

[0238] The number of cold / warm load transfers for each data center should not exceed its local requested load:

[0239]

[0240]

[0241]

[0242]

[0243] in, Let J be the initial cold load number of data center m in region i, I be the set of scheduling regions, and J be the initial cold load number of data center m in region i. i This refers to the set of cities where the data center nodes in region i are located. The number of temperature loads transferred from data center m in region i to data center n in region i. Let be the initial temperature load number for data center m in region i.

[0244] S5. Solve the cross-domain collaborative scheduling model of computing resources in step S4 using the alternating direction multiplier method to obtain the optimal allocation scheme of computing resources. Based on the allocation scheme, schedule the user load to the data center of the corresponding region for processing.

[0245] In another embodiment of the present invention, a cross-domain collaborative scheduling system for computing resources in distributed cloud environments is provided. This system can be used to implement the above-mentioned cross-domain collaborative scheduling method for computing resources in distributed cloud environments. Specifically, the cross-domain collaborative scheduling system for computing resources in distributed cloud environments includes a data module, a partitioning module, a construction module, an optimization module, and a scheduling module.

[0246] The data module acquires user load requests from different regions and cities.

[0247] The user load obtained from the data modules is divided into modules according to a cold-warm-hot classification mechanism;

[0248] Build modules, and construct a mathematical optimization model for global scheduling of computing resources based on the structure obtained from the module division;

[0249] The optimization module, based on the mathematical optimization model obtained from the construction module, combines a decomposition and coordination mechanism to construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing.

[0250] In the optimization module, the objective function of the cross-domain collaborative scheduling model for computing resources is:

[0251]

[0252] in, The transmission cost on the local area network calculated for each region. For the transmission cost on the intranet, Energy consumption cost corresponding to the actual operating load of each region;

[0253] The constraints include:

[0254] (1) Consensus constraints

[0255]

[0256]

[0257] Among them, z ij As an auxiliary variable, x represents the actual amount of cooling load transferred between regions i and j. i [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system for region i. j [i,j] represents the number of cold loads transmitted between regions i and j, calculated by the cloud data center scheduling system in region j.

[0258] (2) Cross-regional cooling load transfer constraints

[0259]

[0260]

[0261] in, N represents the number of cooling loads entering / leaving city m in zone i, respectively. i Let i be a row vector of all 1s. x is a unit vector where the i-th element is 1 and all other elements are 0. i J is the cold load transfer matrix calculated by the regional i-cloud data center scheduling system. i This is the set of cities where the data center nodes in region i are located.

[0262] (3) Intra-domain cooling load transfer constraints

[0263]

[0264] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial number of cooling loads for data center m in region i.

[0265] (4) Intra-domain temperature load transfer constraints

[0266]

[0267] in, This represents the amount of cold load transferred from data center m in region i to data center n in region i. Let m be the initial temperature load for data center m in region i;

[0268] (5) Resource capacity constraints

[0269]

[0270]

[0271] in, The initial temperature load for data center m in region i. Let m be the initial heat load of data center m in region i. The number of cold loads transferred from data center n in region i to data center m in region i. The number of temperature loads transferred from data center n in region i to data center m in region i. Q represents the number of temperature loads transferred from data center m in region i to data center n in region i. im The total number of loads running after scheduling in region i data center m;

[0272] (6) Bandwidth capacity constraints

[0273]

[0274]

[0275]

[0276] in, Let α be the upper limit of the bandwidth capacity of the regional network line (i,j), α be the cold load bandwidth conversion factor, and b be the lower limit of the bandwidth capacity. im β is the upper limit of intra-domain network bandwidth capacity from the central city of region i to the data center m, and β is the temperature load bandwidth conversion coefficient;

[0277] (7) Cooling / warming load quantity constraints

[0278]

[0279]

[0280]

[0281]

[0282] in, Let J be the initial cold load number of data center m in region i, I be the set of scheduling regions, and J be the initial cold load number of data center m in region i. i This refers to the set of cities where the data center nodes in region i are located. The number of temperature loads transferred from data center m in region i to data center n in region i. Let be the initial temperature load number for data center m in region i.

[0283] The scheduling module uses the alternating direction multiplier method to solve the cross-domain collaborative scheduling model of computing resources in the optimization module, and obtains the optimal allocation scheme of computing resources.

[0284] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve corresponding method flows or corresponding functions. The processor described in this embodiment can be used for the operation of a cross-domain collaborative scheduling method for computing resources in distributed clouds, including:

[0285] The system acquires user load requests from different regions and cities; classifies user loads according to a cold-warm-hot classification mechanism; constructs a mathematical optimization model for global scheduling of computing resources based on the structure; combines a decomposition and coordination mechanism to construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing; solves the cross-domain collaborative scheduling model for computing resources using the alternating direction multiplier method to obtain the optimal allocation scheme for computing resources; and schedules user loads to data centers in the corresponding regions for processing according to the optimal allocation scheme.

[0286] Please see Figure 5 The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the fluid composition calculation method in the reservoir stimulation wellbore of this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the cross-domain collaborative scheduling system for distributed cloud computing resources of this embodiment. To avoid repetition, these details are not elaborated here.

[0287] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 5This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0288] The processor 61 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0289] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on the computer device 60.

[0290] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0291] Please see Figure 6 The terminal device is a chip. In this embodiment, the chip 600 includes a processor 622, which may be one or more, and a memory 632 for storing computer programs executable by the processor 622. The computer program stored in the memory 632 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processor 622 may be configured to execute the computer program to perform the aforementioned cross-domain collaborative scheduling method for computing resources in distributed clouds.

[0292] Additionally, chip 600 may also include a power supply component 626 and a communication component 650. The power supply component 626 can be configured to perform power management of chip 600, and the communication component 650 can be configured to enable communication of chip 600, such as wired or wireless communication. Furthermore, chip 600 may also include an input / output (I / O) interface 658. Chip 600 can operate on an operating system stored in memory 632.

[0293] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0294] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the cross-domain collaborative scheduling method for computing resources in the above embodiments for distributed clouds; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:

[0295] The system acquires user load requests from different regions and cities; classifies user loads according to a cold-warm-hot classification mechanism; constructs a mathematical optimization model for global scheduling of computing resources based on the structure; combines a decomposition and coordination mechanism to construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing; solves the cross-domain collaborative scheduling model for computing resources using the alternating direction multiplier method to obtain the optimal allocation scheme for computing resources; and schedules user loads to data centers in the corresponding regions for processing according to the optimal allocation scheme.

[0296] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0297] Please see Figure 4 The system is divided into four regions: A, B, C, and D. Each region has a regional cloud data center scheduling system responsible for coordinating its computing power information and exchanging information with neighboring clouds. Regions are connected via backbone networks with different lines; for simplicity, a fully connected structure is adopted overall. In addition to the core cities with high computing power demand (A1, B1, C1, D1), each region has 1-2 computing power cities (A2, A3, etc.), each with a data center of appropriate size and load data. Cities are connected via intra-domain networks with different lines, all in a star topology centered on the core city.

[0298] For detailed parameters, please refer to Tables 1 to 3. Table 1 shows the architecture information of the backbone network and intranet. Table 2 shows the data center parameter information and corresponding load data of different cities. Other conversion parameters and cost information are shown in Table 3.

[0299] Table 1. Architecture parameters of backbone network and intranet

[0300]

[0301] Table 2. Relevant parameters of data centers in various cities

[0302]

[0303]

[0304] Table 3. Other required parameters for this embodiment.

[0305]

[0306] Note: Cost for each route = base unit price * unit price coefficient;

[0307] Data center operating costs = Total power of data center * Electricity price * Annual conversion factor.

[0308] Based on the above parameters, the basic models (1) to (14) under the centralized framework were solved using the commercial solver gurobi, and the equivalent transformation models (15) to (31) under the distributed framework were solved using the decomposition optimization algorithm. The load scheduling situation of each city point was obtained, and the calculation results of the two were compared with the local calculation results without global load scheduling. The results are detailed in Tables 4 to 7. Tables 4 and 5 show the specific flow direction of warm load and cold load of each city point during the scheduling process, Table 6 shows the total operating load of each city point after the scheduling is completed, and Table 7 shows the comparison of the operating costs under the two scheduling frameworks and the local calculation results without global scheduling.

[0309] Table 4. Specific flow of point temperature load in each city (unit: 10,000 cores)

[0310]

[0311]

[0312] Table 5. Specific flow of cooling load in various cities (unit: 10,000 units)

[0313]

[0314] Table 6 Total Operating Load of Each City (Unit: 10,000 Cores)

[0315]

[0316]

[0317] Table 7 Comparison of Operating Costs (Unit: Yuan)

[0318]

[0319] As shown in Table 7, under the parameters of this embodiment, since the transmission cost is much lower than the energy consumption cost, the total operating cost of global scheduling is reduced by about 14% compared to the local computing cost without scheduling. With the increase in load in developed regions and the widening gap in energy consumption unit prices between regions, the economic benefits will be even more significant. This highlights the superiority of the global scheduling strategy that leverages regional advantages. Furthermore, the computing power scheduling results of multi-regional cloud collaborative optimization are consistent with the centralized framework, demonstrating the accuracy of the proposed distributed framework and the high-performance solution algorithm used.

[0320] In summary, the present invention provides a cross-domain collaborative scheduling method and system for computing resources in distributed cloud environments, which can achieve efficient, flexible, and economical allocation of computing resources across the entire domain while ensuring data privacy and security in each region, decomposing problem scale, and improving computing efficiency, thus providing a new solution for global scheduling strategies of computing resources.

[0321] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0322] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0323] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0324] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0325] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0326] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0327] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random-access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0328] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0329] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0330] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0331] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A cross-domain collaborative scheduling method for computing resources in distributed clouds, characterized in that, Includes the following steps: S1. Obtain user load requests from different regions and cities; S2. Divide the user load obtained in step S1 according to the cold-warm-hot classification mechanism; S3. Based on the structure obtained in step S2, construct a mathematical optimization model for global scheduling of computing resources. The objective function of the mathematical optimization model for global scheduling of computing resources is as follows: in, To reduce cross-domain transmission costs, For intra-domain transmission costs, For data center energy consumption costs; Specifically as follows: in, A set of scheduling regions, For the index of the scheduling region, For regional network lines Transmission costs, in yuan / Gbps / year. For regional network lines Bandwidth usage; Intra-domain transmission costs are as follows: in, For the region The collection of cities within, Number the city For the region The transmission cost of intranet lines within the domain, in yuan / Gbps / year. For the region From the central node to the city Bandwidth usage on network lines; Data center energy consumption costs, after scheduling The energy costs incurred by the district data center in meeting computing demands are as follows: in, This is a conversion factor between the number of data center operating loads and the total operating power. This is the annual conversion factor. For the region after scheduling City The actual number of loads in operation. For the region City The unit total power cost; The specific constraints are as follows: (1) Supply and demand balance constraints express The temperature load before and after the zone transfer is the same, as detailed below: in, For the region City Move to the city The number of temperature loads, For the region The original temperature load quantity; The total cooling load remains the same before and after the transfer, as detailed below: in, For the region City Transfer to the region City The amount of cooling load, For the region The original cooling load quantity; (2) Bandwidth capacity constraints The local area network bandwidth capacity constraints are as follows: The bandwidth capacity constraints of the intranet are as follows: in, This is the bandwidth conversion factor for cooling / warm loads. For regional network lines The upper limit of bandwidth capacity, For the region The maximum bandwidth capacity of intranet lines; (3) Resource capacity constraints Resource capacity constraints refer to the fact that the number of cold / warm / hot loads carried by each data center after scheduling does not exceed its capacity limit: in, For the region Data Center Heat load quantity, For the region Data Center The upper limit of the operating load capacity; (4) Constraints on the quantity of cooling / warming loads The number of cold / warm load transfers for each data center should not exceed its local requested load: in, / Refers to the regions respectively Data Center The number of locally requested cooling / warm loads; S4. Based on the mathematical optimization model obtained in step S3, and combined with the decomposition and coordination mechanism, construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing. S5. Solve the cross-domain collaborative scheduling model of computing resources in step S4 using the alternating direction multiplier method to obtain the optimal allocation scheme of computing resources. Based on the optimal allocation scheme, schedule user loads to the data centers in the corresponding regions for processing.

2. The cross-domain collaborative scheduling method for computing resources in distributed cloud environments according to claim 1, characterized in that, In step S1, the user load request includes: Computing resources, that is, the amount of computing power resources required to complete a user request; Latency requirement refers to the maximum acceptable computational latency from responding to a user's request to completing that request.

3. The cross-domain collaborative scheduling method for computing resources in distributed cloud environments according to claim 1, characterized in that, Step S2 is as follows: S201. Estimate the transmission time based on the transmission distance between different regions and cities, divide the delay range according to the magnitude of the transmission time, and design a corresponding cold-temperature-heat load classification mechanism. S202. Based on the user's latency requirements, classify the user load according to the cold-warm-hot division mechanism: users with cold loads are allowed to request cross-domain scheduling, users with warm loads are only allowed to request intra-domain scheduling, and users with hot loads are only allowed to request local calculation.

4. The cross-domain collaborative scheduling method for computing resources in distributed cloud environments according to claim 3, characterized in that, In step S201, the corresponding cooling-temperature-heating load classification mechanism is as follows: Thermal loads are time-delay sensitive and require rapid response; they can only be calculated locally and cannot be scheduled. Cold loads accept long time-span responses, making them suitable for cross-domain scheduling; Temperature load, which falls between heat load and cold load, accepts short-term inter-period responses and is capable of intra-domain scheduling.

5. The cross-domain collaborative scheduling method for computing resources in distributed cloud environments according to claim 1, characterized in that, In step S4, the objective function of the cross-domain collaborative scheduling model for computing resources is: in, The transmission cost on the local area network calculated for each region. For the transmission cost on the intranet, The energy cost corresponding to the actual operating load of each region This is a set of scheduling regions.

6. The cross-domain collaborative scheduling method for computing resources in distributed cloud environments according to claim 5, characterized in that, The transmission cost on the regional network calculated for each region is as follows: Transmission costs on intranet for: Energy consumption cost corresponding to the actual operating load of each region for: in, The basic unit price for bandwidth in a regional network / internal domain network is expressed in yuan / Gbps / year. Indicates the area The replicated regional network cold load flow matrix , A one-dimensional vector consisting of the unit price coefficients of line bandwidth in adjacent areas; The constraints include: (1) Consensus constraints in, As an auxiliary variable, it represents the region. , The actual amount of cooling load transferred between them For the region The cloud data center scheduling system calculates the area , The amount of cooling load transferred between them For the region The cloud data center scheduling system calculates the area , The amount of cold load transferred between them; (2) Constraints on cross-regional cooling load transfer in, Representing in / out respectively district The city's cooling load. for A row vector of all 1s. Let i be a unit vector whose i-th element is 1 and all other elements are 0. For the region The cold load transfer matrix calculated by the cloud data center scheduling system. For the region The set of cities where the data center nodes are located; (3) Intra-domain cooling load transfer constraints in, Indicates the area Data Center Transfer to the region Data Center The amount of cooling load, For the region Data Center The initial cooling load; (4) Intra-domain temperature load transfer constraints in, Indicates the area Data Center Transfer to the region Data Center The amount of cooling load, For the region Data Center The initial temperature load number; (5) Resource capacity constraints in, For the region Data Center The initial temperature load number, For the region Data Center The initial heat load number, For the region Data Center Transfer to the region Data Center The number of cooling loads, For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center The total number of loads running after scheduling; (6) Bandwidth capacity constraints in, For regional network lines The upper limit of bandwidth capacity, This is the bandwidth conversion factor for cooling load. For the region From central cities to data centers The upper limit of intranet bandwidth capacity, This is the temperature load bandwidth conversion factor; (7) Cooling / warming load quantity constraints in, For the region Data Center The initial cooling load number, For the set of scheduling regions, For the region The set of cities where the data center nodes are located. For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center The initial temperature load number.

7. A cross-domain collaborative scheduling system for computing resources in distributed cloud environments, used to execute the cross-domain collaborative scheduling method for computing resources in distributed cloud environments as described in any one of claims 1 to 6, characterized in that, include: The data module retrieves user load requests from different regions and cities. The user load obtained from the data modules is divided into modules according to a cold-warm-hot classification mechanism; The construction module is used to build a mathematical optimization model for global scheduling of computing resources based on the structure obtained from the module division. The objective function of the mathematical optimization model for global scheduling of computing resources is as follows: in, To reduce cross-domain transmission costs, For intra-domain transmission costs, For data center energy consumption costs; The specific constraints are as follows: (1) Supply and demand balance constraints express The temperature load before and after the zone transfer is the same, as detailed below: in, For the region City Move to the city The number of temperature loads, For the region The original temperature load quantity; The total cooling load remains the same before and after the transfer, as detailed below: in, For the region City Transfer to the region City The amount of cooling load, For the region The original cooling load quantity; (2) Bandwidth capacity constraints The local area network bandwidth capacity constraints are as follows: The bandwidth capacity constraints of the intranet are as follows: in, This is the bandwidth conversion factor for cooling / warm loads. For regional network lines The upper limit of bandwidth capacity, For the region The maximum bandwidth capacity of intranet lines; (3) Resource capacity constraints Resource capacity constraints refer to the fact that the number of cold / warm / hot loads carried by each data center after scheduling does not exceed its capacity limit: in, For the region Data Center Heat load quantity, For the region Data Center The upper limit of the operating load capacity; (4) Constraints on the quantity of cooling / warming loads The number of cold / warm load transfers for each data center should not exceed its local requested load: in, / Refers to the regions respectively Data Center The number of locally requested cooling / warm loads; The optimization module, based on the mathematical optimization model obtained from the construction module, combines a decomposition and coordination mechanism to construct a cross-domain collaborative scheduling model for computing resources suitable for distributed cloud computing. The scheduling module uses the alternating direction multiplier method to solve the cross-domain collaborative scheduling model of computing resources in the optimization module, and obtains the optimal allocation scheme of computing resources.

8. The cross-domain collaborative scheduling system for computing resources in distributed cloud environments according to claim 7, characterized in that, In the optimization module, the objective function of the cross-domain collaborative scheduling model for computing resources is: in, The transmission cost on the local area network calculated for each region. For the transmission cost on the intranet, The energy cost corresponding to the actual operating load of each region.

9. The cross-domain collaborative scheduling system for computing resources in distributed cloud environments according to claim 8, characterized in that, Transmission costs on the local area network calculated for each region for: Transmission costs on intranet for: Energy consumption cost corresponding to the actual operating load of each region for: in, The basic unit price for bandwidth in a regional network / internal domain network is expressed in yuan / Gbps / year. Indicates the area The replicated regional network cold load flow matrix , A one-dimensional vector consisting of the unit price coefficients of line bandwidth in adjacent areas; The constraints include: (1) Consensus constraints in, As an auxiliary variable, it represents the region. , The actual amount of cooling load transferred between them For the region The cloud data center scheduling system calculates the area , The amount of cooling load transferred between them For the region The cloud data center scheduling system calculates the area , The amount of cold load transferred between them; (2) Constraints on cross-regional cooling load transfer in, Representing in / out respectively district The city's cooling load. Let i be a row vector of all 1s. Let i be a unit vector whose i-th element is 1 and all other elements are 0. For the region The cold load transfer matrix calculated by the cloud data center scheduling system. For the region The set of cities where the data center nodes are located; (3) Intra-domain cooling load transfer constraints in, Indicates the area Data Center Transfer to the region Data Center The amount of cooling load, For the region Data Center The initial cooling load; (4) Intra-domain temperature load transfer constraints in, Indicates the area Data Center Transfer to the region Data Center The amount of cooling load, For the region Data Center The initial temperature load number; (5) Resource capacity constraints in, For the region Data Center The initial temperature load number, For the region Data Center The initial heat load number, For the region Data Center Transfer to the region Data Center The number of cooling loads, For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center The total number of loads running after scheduling; (6) Bandwidth capacity constraints in, For regional network lines The upper limit of bandwidth capacity, This is the bandwidth conversion factor for cooling load. For the region From central cities to data centers The upper limit of intranet bandwidth capacity, This is the temperature load bandwidth conversion factor; (7) Cooling / warming load quantity constraints in, For the region Data Center The initial cooling load number, For the set of scheduling regions, For the region The set of cities where the data center nodes are located. For the region Data Center Transfer to the region Data Center Temperature load number, For the region Data Center The initial temperature load number.

Images