A method for calculating the maximum delay of a data center distribution network and planning its configuration

Abstract

The present invention discloses a method for calculating the maximum delay and planning the configuration of a data center distribution network. First, an information element model analogous to an electric circuit is proposed, specifically including information flow, information voltage, information resistance, and information conductance. Secondly, based on the basic element model, an equivalent circuit model of the information system for the data center distribution network is constructed. Thirdly, based on the equivalent circuit model, an information flow and information voltage constraint model is constructed. Then, considering channel blockage and load priority, the maximum delay of the data center distribution network is calculated. Finally, with the goal of minimizing the maximum delay of the entire system, a data center planning and configuration scheme for the distribution network is formulated. The present invention takes into account the user priority and channel blockage problems, and conducts the maximum delay calculation in the form of an analogous electric circuit, with higher calculation accuracy and smaller calculation amount, and can significantly reduce the power flow loss of the distribution network and the maximum delay of the data center.

Description

Technical Field

[0001] The present invention relates to a method for calculating the delay of a data center distribution network, an analog circuit model, and a method for planning and configuring a data center, and particularly to a method for calculating the maximum delay of a data center distribution network and planning and configuring the same. Background Art

[0002] An Internet Data Center (IDC) is a large-scale centralized information processing facility for storing and computing. With the continuous increase in people's data communication and data processing requirements, IDCs play an increasingly important role. Generally, an IDC collects data from an information system, processes user computing tasks, and distributes the computing results to various users. This process will generate computing and transmission delays, and the magnitude of this delay will directly affect the quality of user services. At the same time, the IDC will be powered by a distribution network, and its planned configuration capacity will directly affect the distribution network power flow, thereby exacerbating the energy loss of power lines.

[0003] Currently, existing models cannot accurately and quickly calculate the maximum delay of an IDC, making it difficult to carry out IDC planning and configuration work according to the actual needs of users and the distribution network. Summary of the Invention

[0004] The purpose of the present invention is to quickly and accurately calculate the maximum delay of an IDC distribution network through an analog circuit model, and provide an information system planning and configuration method considering the maximum delay, as described in detail below:

[0005] A method for calculating the maximum delay of an IDC distribution network and planning and configuring the same, the method comprising the following steps:

[0006] Propose an information element model of an analog circuit;

[0007] Based on the basic element model, construct an equivalent circuit model of an information system for a data center distribution network;

[0008] Based on the equivalent circuit model, construct a signal flow and signal voltage constraint model;

[0009] Considering channel blockage and load priority, calculate the maximum delay of a data center distribution network;

[0010] With the goal of minimizing the maximum delay of the entire system, formulate a data center planning and configuration plan for the distribution network.

[0011] Further, the information element model of the analog circuit is specifically:

[0012] Through an analog circuit, the present invention establishes the following information element model:

[0013] Establish a signal flow I' model as:

[0014]

[0015] Wherein, q' is the amount of information generated at time t.

[0016] The information pressure U' model is established as:

[0017]

[0018] Wherein, Y' is the information conductance, and its value represents the information transmission and processing speed of the communication line or IDC; T is the time used to process this information; U' represents the ratio of the information processing time to the generation time; it is stipulated that t is the unit time; when t takes the unit time, U' is numerically the same as the information processing time;

[0019] The information resistance R' model is established as:

[0020]

[0021] Wherein, R' is the time spent on processing a unit of information.

[0022] Furthermore, based on the basic element model, an equivalent circuit model of the information system for the data center distribution network is constructed:

[0023] Suppose there is Figure 3 The IDC distribution network shown, in which there are two users with a priority of 1, one user with a priority of 2, and one IDC in its information system (we stipulate that the IDC will give priority to processing users with a higher priority). Based on the above definition of basic information elements, we can make the following conversions:

[0024] User: Each user can be equivalently regarded as an information flow source and an information resistance. Among them, the information flow source represents the amount of information generated by this user, and the information resistance represents the information upload speed of this user.

[0025] Communication line: Each communication line can be represented as an information resistance, and this information resistance represents the information transmission speed of this communication line.

[0026] IDC: The IDC can be represented as an information resistance, and this information resistance represents the information processing speed of the IDC.

[0027] Furthermore, based on the equivalent circuit model, an information flow and information pressure constraint model is constructed:

[0028] Information flow constraint: For a certain information node, without considering information conversions such as encoding and decoding, the information flow flowing into this node and the information flow flowing out of this node are numerically equal;

[0029] Information pressure constraint: For a given communication network, when there are different communication paths from one node to another, the total time for the information flow to pass through different paths is the same. This indicates that the transmission schemes for different communication paths have been optimized in advance to minimize the total transmission time.

[0030] Furthermore, considering channel blockage and load priority, calculate the maximum delay of the data center distribution network:

[0031] First, based on the method proposed in this patent, convert the Figure 3 information system of the IDC distribution network shown in (a) into the Figure 3 analog circuit model shown in (b), and calculate the information pressure of each node. The calculation result here is the maximum delay that users can withstand under extremely harsh conditions.

[0032] Second, correct the calculation result of the maximum delay of high-priority users. Since the information processing requirements of high-priority users must be satisfied first, when calculating high-priority users 2 and 3, it is necessary to interrupt the communication line of user 1 as shown in Figure 3 (c), recalculate the information pressure, and correct the result of step 1).

[0033] Furthermore, with the goal of minimizing the maximum delay of the entire system, formulate a data center planning and configuration scheme for the distribution network:

[0034] Calculate the maximum delay of the entire system through the method of this patent, and carry out the planning and configuration of the IDC in the distribution network with the goal of minimizing this maximum delay and minimizing the power supply loss.

[0035] The beneficial effects of the technical solution provided by the present invention are:

[0036] 1) Compared with the traditional information system model, the present invention converts the information system model into a circuit model, which helps power dispatching personnel understand the operation law of the information system and can assist in improving the effect of power system optimal dispatching.

[0037] 2) Compared with the traditional maximum transmission delay calculation method, the method proposed by the present invention considers the user priority and channel blockage problems, and calculates the maximum delay in the form of an analog circuit, with higher calculation accuracy and smaller calculation amount.

[0038] 3) Compared with the traditional IDC planning and configuration method, the present invention can significantly reduce the power flow loss of the distribution network and the maximum delay of the IDC. Description of the Drawings

[0039] Figure 1 It is a flowchart of the method for calculating and planning the maximum delay of the IDC distribution network;

[0040] Figure 2 It is a typical IDC distribution network architecture;

[0041] Figure 3 is an information processing model;

[0042] Among them, Figure 3 (a) is an example of an information system, Figure 3 (b) is an analog circuit model, Figure 3 (c) is an analog circuit model considering user priorities;

[0043] Figure 4 is a simulation scenario diagram of an IDC distribution network based on IEEE 33 nodes;

[0044] Figure 5 is the comparison result of the maximum delay calculated by the proposed method and the actual delay range. Specific implementation manners

[0045] To make the objectives, technical solutions and advantages of the present invention clearer, the following further describes in detail the implementation manners of the present invention.

[0046] Example 1:

[0047] A method for calculating the maximum delay and planning the configuration of an IDC distribution network, the method includes the following steps:

[0048] Step 101: Propose an information element model of an analog circuit, specifically including information flow, information voltage, information resistance, and information conductance;

[0049] Through the analog circuit, the present invention proposes the following information element model, as shown in Table 1:

[0050] Table 1 Information element model of the analog circuit

[0051] Circuit element Information element Current I Information current I' Resistance R Information resistance R' Conductance Y Information conductance Y' Voltage U Information voltage U'

[0052] Establish the information flow I' model as:

[0053]

[0054] In the formula, q' is the amount of information generated at time t.

[0055] Establish the information voltage U' model as:

[0056]

[0057] In the formula, Y' is the information conductance, and its value characterizes the information transmission and processing speed of the communication line or IDC; T is the time used to process this information; through mathematical calculation, it can be obtained that U' represents the ratio of the information processing time to the generation time. When t takes the unit time, U' is numerically the same as the information processing time. Therefore, in the present invention, it is stipulated that t takes the unit time.

[0058] The information resistance R' model is established as follows:

[0059]

[0060] In the formula, R' is the time spent on processing unit information.

[0061] Step 102: Based on the basic element model, construct an equivalent circuit model of the information system for the data center distribution network;

[0062] Assume there is Figure 3 The IDC distribution network shown in the figure. There are two users with a priority of 1, one user with a priority of 2, and one IDC in its information system (we stipulate that the IDC will give priority to processing users with a higher priority). Based on the above definition of basic information elements, we can make the following transformation:

[0063] User: Each user can be equivalent to an information flow source and an information resistance. Among them, the information flow source represents the amount of information generated by the user, and the information resistance represents the information upload speed of the user.

[0064] Communication line: Each communication line can be represented as an information resistance, and this information resistance represents the information transmission speed of this communication line.

[0065] IDC: The IDC can be represented as an information resistance, and this information resistance represents the information processing speed of the IDC.

[0066] Through the above method, we can Figure 3 transform the actual information system shown in (a) into Figure 3 the circuit analogy model shown in (b). Figure 3 In (b), U' at node 1' represents the maximum delay of user 1, and the calculation method for other users is similar.

[0067] Step 103: Based on the equivalent circuit model, construct an information flow and information voltage constraint model;

[0068] To solve the information system circuit model established in step 102, the present invention proposes the following information flow and information voltage constraint models:

[0069] Information flow constraint: For a certain information node, without considering information conversions such as encoding and decoding, the information flow flowing into the node and the information flow flowing out of the node are numerically equal. It can be expressed by the following formula:

[0070]

[0071] In the formula, Ω is the collection of information flows flowing into the node; ω is the set of information flows flowing out of the node.

[0072] Information pressure constraint: For a given communication network, when there are different communication paths from one node to another, the total time for the information flow to pass through different paths is the same. This indicates that the transmission schemes for different communication paths have been optimized in advance to minimize the total transmission time. It can be expressed by the following formula:

[0073]

[0074] where φ is the set of information pressures on an information loop.

[0075] Step 104: Considering channel blockage and load priority, calculate the maximum delay of the data center distribution network;

[0076] Considering the priority of user information processing, the calculation of the maximum delay is divided into two steps:

[0077] Calculate Figure 3 the information pressures of each node in (b), and the results are as follows:

[0078]

[0079] The calculation result of formula (6) represents the maximum blocking time that the user can tolerate, that is, for any given user, when it uses any communication line and IDC, other users must execute their tasks prior to this user. In an actual information network, since information from different users is generated at different times, the above extreme scenario may occur.

[0080] Since the information processing requirements of high-priority users must be satisfied first, when calculating high-priority users 2 and 3, it is necessary to interrupt the communication line of user 1 as shown in Figure 3 (c), recalculate the information pressure and correct the result of step 1). The modified results are as follows:

[0081]

[0082] When there are more priority divisions, repeat the above step 2) level by level to calculate the accurate maximum delay time.

[0083] Step 105: With the goal of minimizing the maximum delay of the entire system, formulate a data center planning and configuration scheme for the distribution network;

[0084] Calculate the maximum delay of the entire system through the method of this patent, and carry out the planning and configuration of IDC in the distribution network with the goal of minimizing this maximum delay and minimizing the power supply loss.

[0085] (1) Objective function

[0086]

[0087] In the formula, n is the number of nodes in the IDC distribution network; U ( ' i') is the signal voltage of node i calculated by formula (7), which is numerically equal to the maximum calculation delay of each node; C represents the IDC construction cost; a and b are weight coefficients.

[0088] Among them, the IDC construction cost can be calculated by the following formula:

[0089]

[0090] In the formula, δ i is the node identifier, the value of which can only be 0 or 1. When its value is "0", it means that no IDC is built at this node; when its value is "1", it means that an IDC is built at this node; C 1 represents the infrastructure cost of building an IDC, usually including site cost, building construction cost, etc.; C 2 represents the unit capacity configuration price of the IDC computing device, and Q i represents the IDC capacity built at node i.

[0091] (2) Constraint conditions

[0092] 1) Installation capacity constraint

[0093]

[0094] In the formula, represents the maximum IDC capacity allowed to be installed at node i.

[0095] 2) Site selection plan constraint

[0096]

[0097] In the formula, when IDC is not allowed to be built at node i, δ i can only be 0; when IDC center is allowed at node i, δ i can be 0 or 1.

[0098] Example 2:

[0099] Specific examples are given below to verify the feasibility of the above method. See the following description for details:

[0100] To verify the effectiveness of the maximum delay calculation method, the present invention conducts simulations on the IEEE 33-node network, and the simulation scenario is as Figure 4 shown. At the same time, in order to ensure the data integrity when transmitting information to the IDC, the network is effectively divided into three regions, and each region is managed by an independent IDC.

[0101] In the simulation of the foregoing scenario, the comparison between the maximum delay calculated by the proposed method and the actual delay range is as Figure 5 shown. It can be seen that the maximum delay calculated by the proposed method is very close to the actual maximum delay.

[0102] In addition, calculating the maximum time delay of an information system with n nodes by using the method of the present invention is equivalent to solving a system of linear equations with n variables, and the computational complexity is small.

[0103] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of the present invention are generated in whole or in part.

[0104] The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted through a computer-readable storage medium. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or a data center integrating one or more available media. The available medium can be a magnetic medium or a semiconductor medium, etc.

[0105] For the models of the various devices in the embodiments of the present invention, unless otherwise specifically stated, the models of other devices are not limited, as long as the devices can perform the above functions.

[0106] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred embodiment, and the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0107] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for calculating and planning the maximum delay of a data center distribution network, characterized in that: The method comprises: proposing an information element model of an analog circuit, specifically including signal flow, signal pressure, signal resistance and signal conduction; the proposed information element model of the analog circuit is specifically: By analogy with the circuit, the following information element model is established: Establishing a flow The model is: In the formula, for The amount of information generated at any given moment; Build trust The model is: In the formula, It is a signal conductor, and its value represents the speed of information transmission and processing in a communication line or data center; the time taken to process the information; Represents the ratio of information processing time to generation time; when When taking unit time, It is numerically identical to the information processing time; Establish signal resistance The model is: In the formula, The time required to process unit information; Based on the basic element model, an information system equivalent circuit model for the data center distribution network is constructed, specifically: For the data center distribution network, there are two users with priority 1, one user with priority 2 and one data center in its information system. The data center will give priority to the user with higher priority. Based on the above basic information element definition, the following conversion is performed: User: Each user can be equivalent to a traffic source and a signal barrier, where the traffic source represents the amount of information generated by the user, and the signal barrier represents the information upload speed of the user; Communication line: Each communication line can be represented as a signal impedance, which represents the information transmission speed of this communication line; Data center: A data center can be represented as a signal barrier, which represents the information processing speed of the data center; Based on the equivalent circuit model, a signal flow and signal pressure constraint model is constructed, specifically: Information flow constraint: For a certain information node, regardless of the encoding and decoding information conversion, the information flow into and out of the node is equal in value; Information pressure constraint: For a certain communication network, when there are different communication paths from one node to another, the total time for information flow to pass through different paths is the same, which means that the transmission schemes of different communication paths have been optimized in advance to minimize the total transmission time. Calculate the maximum delay of the data center distribution network considering channel congestion and load priority; With the goal of minimizing the maximum latency of the entire system, a data center planning and configuration plan for the distribution network is developed.

2. A method for calculating and planning the maximum delay of a data center distribution network according to claim 1, characterized in that: The maximum delay of the data center distribution network is calculated by considering channel blocking and load priority as follows: First, the information system of the data center distribution network is converted into an analog circuit model to calculate the signal pressure of each node. The calculation result is the maximum delay that the user can bear under extremely severe conditions. Secondly, correct the maximum delay calculation result of the high priority user. Since the information processing requirements of the high priority user must be met first, when calculating high priority users 2 and 3, the communication line of user 1 must be interrupted, the signal pressure is recalculated and the maximum delay result is corrected.

3. A method for calculating and planning the maximum delay of a data center distribution network according to claim 1, characterized in that: The above-mentioned data center planning and configuration scheme for the distribution network is formulated with the goal of minimizing the maximum delay of the entire system as follows: Calculate the maximum delay of the entire system, and use the minimum maximum delay and power loss as planning goals to carry out planning and configuration of the data center in the distribution network.

Images