Machine room deployment method, device, equipment, storage medium and computer program product

Abstract

The application discloses a machine room deployment method, device, equipment, storage medium and computer program product. The method comprises the following steps: acquiring machine room data center basic physical design unit (POD) networking information and machine room parameters; calculating the number of network devices, the number of servers and the target cabinet number of racks according to the POD networking information and the machine room parameters, and constructing a base minimum construction unit; calculating the number of servers that can be carried by the base and the number of racks that are preoccupied by the base according to the port proportion of the network device and the base minimum construction unit, and constructing a standardized base model; and deploying the machine room according to the computing power resource construction demand and the standardized base model. The application can effectively improve the delivery efficiency of the liquid cooling data center computing power resource project construction engineering by the proactive response of the pre-designed standardized base model to the resource demand, and realizes the rapid response and agile delivery of the computing power resource demand under the intelligent computing scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer technology, and in particular to a computer room deployment method, device, equipment, storage medium and computer program product. BACKGROUND

[0002] With the acceleration of innovation in artificial intelligence, big data and other technologies, the digital economy is increasingly integrated into the real economy. China has entered the era of intelligence and is actively promoting the development of intelligent computing centers. As the infrastructure for computing power, the number of data centers and racks has seen explosive growth, leading to a sharp increase in data center electricity consumption. At the same time, with the increasing demand for single-point computing power, the power consumption of a single computing server is tens of times that of a traditional server, reaching 1000-10000W. In order to reduce data center energy consumption and help achieve China's dual-carbon goals, the existing data center air cooling efficiency has already been unable to meet the cooling and energy consumption needs, and data center computer rooms need to be adapted to mainstream liquid cooling technology. Liquid cooling replaces traditional air cooling by using cooling liquid for heat exchange with servers, which can significantly improve cooling efficiency and reduce data center energy consumption.

[0003] With the development of intelligent computing and the large-scale application of liquid-cooled computer rooms, the demand for computing power resources and liquid-cooled racks is growing rapidly, and the timeliness of delivery is higher. The delivery efficiency of conventional cyclic progressive construction projects cannot effectively meet the actual needs of the rapid development of computing power. In the existing liquid-cooled data center computing power resource delivery process, after receiving a large-scale computing power construction demand, it greatly relies on manual sorting and verification to determine whether the corresponding computer room supporting resources meet the resource demand. Due to the large-scale and sudden nature of many demands, the computer room providing side cannot effectively guarantee the actual number of available racks, resulting in insufficient rack and supporting resource reserves to meet sudden resource demands. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a computer room deployment method, device, equipment, storage medium and computer program product, which can effectively improve the delivery efficiency of liquid-cooled data center computing power resource project construction by pre-designing a standardized base model to actively respond to resource demand, and realize rapid response and agile delivery of computing power resource demand in intelligent computing scenarios.

[0005] To achieve the above purpose, the embodiments of the present application provide a computer room deployment method, comprising:

[0006] Obtaining computer room data center basic physical design unit POD networking information and computer room parameters;

[0007] According to the POD networking information and the computer room parameters, the number of network devices, the number of servers and the target cabinet number of racks are calculated, and the base minimum construction unit is constructed;

[0008] According to the port ratio of the network device and the base minimum construction unit, the number of servers that the base can carry and the number of racks that the base preoccupies are calculated, and a standardized base model is constructed;

[0009] According to the computing resource construction demand and the standardized base model, the computer room is deployed.

[0010] As an improvement of the above-mentioned scheme, the network device includes an access switch and a convergence switch, and the number of network devices, the number of servers, and the target cabinet number of racks are calculated according to the POD networking information and the computer room parameters, and the base minimum construction unit is constructed, including:

[0011] According to the POD networking information, the access switch parameters, the rack parameters, and the server parameters, the number of servers that a single rack can carry is calculated;

[0012] According to the port ratio of the access switch and the port ratio of the server, the number of access switches and the target cabinet number of racks are calculated;

[0013] According to the number of servers that a single rack can carry, the number of access switches, and the target cabinet number of racks, the base minimum construction unit is constructed.

[0014] As an improvement of the above-mentioned scheme, according to the POD networking information, the access switch parameters, the rack parameters, and the server parameters, the number of servers that a single rack can carry is calculated, including:

[0015] According to the POD networking information and the port ratio of the access switch, the number of servers that a single access switch can access is calculated;

[0016] According to the rated power of the rack, the rated power of the access switch, and the rated power of the server, the number of servers that a single rack can carry under power conditions is calculated;

[0017] According to the available U number of the rack, the U number of the access switch, and the occupied U number of a single server, the number of servers that a single rack can carry under space conditions is calculated;

[0018] According to the number of servers that a single access switch can access, the number of servers that a single rack can carry under power conditions, and the number of servers that a single rack can carry under space conditions, the number of servers that a single rack can carry is determined.

[0019] As an improvement of the above-mentioned scheme, according to the port ratio of the network device and the base minimum construction unit, the number of servers that the base can carry and the number of racks that the base preoccupies are calculated, and a standardized base model is constructed, including:

[0020] According to the port proportion of the aggregation switch and the base minimum construction unit, the number of base minimum construction units that the base can carry and the number of racks preoccupied by the base are calculated;

[0021] According to the number of base minimum construction units and the number of servers that a single rack can carry, the number of servers that the base can carry is calculated;

[0022] According to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks preoccupied by the base, and the number of servers that the base can carry, a standardized base model is constructed.

[0023] As an improvement of the above-mentioned scheme, the standardized base model is constructed according to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks preoccupied by the base, and the number of servers that the base can carry, comprising:

[0024] According to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks preoccupied by the base, and the number of servers that the base can carry, a base model is constructed;

[0025] According to the number of servers that the base can carry, it is judged whether it meets the demand of the computer room and the business;

[0026] If it meets, the standardized base model is formed; if it does not meet, the number of aggregation switches is adjusted, and the size of the base is recalculated.

[0027] As an improvement of the above-mentioned scheme, the computer room is deployed according to the computing resource construction demand and the standardized base model, comprising:

[0028] The standardized base model is synchronized to the computer room providing side, the racks are preoccupied, and the computer room supporting resources are transformed, and the network equipment is preinstalled and wired synchronously;

[0029] According to the computing resource construction demand, the network equipment and the server are installed and wired according to the standardized base model.

[0030] As an improvement of the above-mentioned scheme, the method further comprises:

[0031] If the proportion of racks used by the base exceeds a preset threshold, an expansion mechanism is triggered to supplement preoccupied rack resources to a complete base.

[0032] Embodiments of the present application also provide a computer room deployment device, comprising:

[0033] The data acquisition module is configured to acquire POD networking information and room parameters of a basic physical design unit of a data center room.

[0034] The unit construction module is configured to calculate the number of network devices, the number of servers, and the target cabinet number of racks according to the POD networking information and the room parameters, and construct a base minimum construction unit.

[0035] The model construction module is configured to calculate the number of servers that can be carried by the base and the number of racks that can be preoccupied by the base according to the port ratio of the network device and the base minimum construction unit, and construct a standardized base model.

[0036] The room deployment module is configured to deploy the room according to the computing power resource construction demand and the standardized base model.

[0037] The embodiment of the present application further provides a terminal device, which comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the room deployment method of any one of the above when executing the computer program.

[0038] The embodiment of the present application further provides a computer readable storage medium, which comprises a stored computer program, wherein the computer readable storage medium controls a device where the computer readable storage medium is located to execute the room deployment method of any one of the above when the computer program runs.

[0039] The embodiment of the present application further provides a computer program product, which comprises a computer program or computer instructions, and the computer program or the computer instructions implement the room deployment method of any one of the above when executed by a processor.

[0040] The method, device, equipment, storage medium and computer program product provided by the embodiment of the application have the beneficial effects that, by acquiring POD networking information of a basic physical design unit of a data center of a computer room and computer room parameters, the number of network devices, the number of servers and the target number of cabinets of racks are calculated according to the POD networking information and the computer room parameters, and a minimum construction unit of a base is constructed; according to the port ratio of the network devices and the minimum construction unit of the base, the number of servers that can be carried by the base and the number of racks that are preoccupied by the base are calculated, and a standardized base model is constructed; and the computer room is deployed according to the demand of computing power resource construction and the standardized base model. The embodiment of the application actively responds to the resource demand by pre-designing the standardized base model, plays a huge role in the design, supply and delivery stages of a liquid-cooled data center construction project in an intelligent computing scenario, can effectively improve the delivery efficiency of a liquid-cooled data center computing power resource project construction project, and realizes rapid response and agile delivery of the computing power resource demand in the intelligent computing scenario. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flowchart of a preferred embodiment of a computer room deployment method provided by the application;

[0042] Figure 2 is a comparison diagram of the whole process construction period of a computer room deployment method provided by the application and a traditional method;

[0043] Figure 3 is a POD networking diagram of a computer room in a computer room deployment method provided by the application;

[0044] Figure 4 is a structure diagram of a minimum construction unit of a base in a computer room deployment method provided by the application;

[0045] Figure 5 is a structure diagram of a standardized base model in a computer room deployment method provided by the application;

[0046] Figure 6 is a structure diagram of a standardized base model after delivery in a computer room deployment method provided by the application;

[0047] Figure 7 is a structure diagram of a preferred embodiment of a computer room deployment device provided by the application;

[0048] Figure 8 is a structure diagram of a preferred embodiment of a terminal device provided by the application. DETAILED DESCRIPTION

[0049] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those ordinarily skilled in the art without creative work fall within the scope of the present application.

[0050] Please refer to Figure 1 , Figure 1 is a flowchart of a preferred embodiment of a method for deploying a computer room provided by the present application. The method for deploying a computer room comprises the following steps.

[0051] S1, obtaining POD networking information of a computer room data center basic physical design unit and computer room parameters;

[0052] S2, calculating the number of network devices, the number of servers and the target cabinet number of racks according to the POD networking information and the computer room parameters, and constructing a base minimum construction unit;

[0053] S3, calculating the number of servers that the base can carry and the number of racks that the base can occupy according to the port ratio of the network device and the base minimum construction unit, and constructing a standardized base model;

[0054] S4, deploying the computer room according to the computing resource construction demand and the standardized base model.

[0055] Specifically, please refer to Figure 2 , Figure 2 is a comparison diagram of the whole process construction period of a method for deploying a computer room provided by the present application and a traditional method. The intelligent computing data center usually carries self-research training and large-grain computing power demand, and the deployment computing power scale is large, usually above ten thousand cards cluster, which belongs to the super large scale new computing power infrastructure of intensive construction. The large scale and suddenness of computing power resource demand bring many tests to the traditional data center project construction engineering in the aspects of design, supply and delivery at the present stage, and there are the following problems:

[0056] (1) MO manual output stage: after receiving the demand for large-scale construction of computing power, great reliance is placed on manual pre-survey to verify whether the supporting resources of the machine room corresponding to the resource demand meet the requirements (such as the number of racks and ODF transmission required for this resource expansion). Due to the large-scale and sudden nature of many demands, the machine room providing side cannot effectively guarantee the actual available rack quantity in time, resulting in insufficient rack and supporting resource reserves to meet sudden resource demands. For example: involving rack and ODF (Optical Distribution Frame) transmission transformation, the overall transformation time is relatively long and cannot be effectively controlled, usually reaching more than 1.5 months of average transformation time, and the high coefficient of uncertainty of machine room supporting seriously affects the start time of the construction project plan.

[0057] (2) Design stage: after receiving the information about the use of machine rooms and racks output by MO, the construction project plan needs to be prepared in the design stage, the use planning of machine rooms and racks, the output of product server and network equipment configuration and networking, the output of network optimization construction plan, machine room supporting and optical cable transmission construction plan, etc. This is a passive response design method for sudden resource demand, which is not only slow but also inefficient.

[0058] (3) Supply stage: server and liquid cooling rack matching is difficult, the current liquid cooling rack and liquid cooling server and liquid cooling technology standardization level are weak, there is no substantial standard for the connection of server and liquid cooling rack pipeline in the industry, and there is a lack of compatibility connection. In the supply stage, the connection between the server and the liquid cooling rack needs to be matched, and the transformation and matching period is relatively long.

[0059] (4) Delivery stage: in the hard set equipment installation link, large-scale computing power resource equipment needs to be quickly installed and integrated in a short period of time, which is very tedious and time-consuming. Network equipment such as aggregation switches, access switches (TOR) are installed on the rack, and then the cable in the rack and the cabinet line from the aggregation switch to the rack are connected, which seriously affects the work efficiency of the hard set, and further delays the progress of the soft set, affecting the delivery period of the entire demand resource.

[0060] The traditional data center project resource construction full process is from demand to MO artificial output, through design-supply-delivery construction time of more than 9 months. The present application proposes a liquid cooling machine room standardized base (Base: matching computing power resource demand infrastructure modular solution, including machine room, rack, network equipment, general wiring, etc.) design method before resource delivery in the intelligent computing scenario, which can pre-position the MO artificial output capacity before the computing power burst resource demand arrives, and effectively shorten the design-supply-delivery time, i.e. D-S-D construction cycle, greatly improve the delivery efficiency of the liquid cooling data center computing power resource project construction engineering full process in the intelligent computing scenario.

[0061] The embodiment of the present application obtains POD (point of delivery, basic physical design unit of data center) networking information and machine room parameters. Please refer to Figure 3 , Figure 3 is a POD networking schematic diagram in a machine room deployment method provided by the present application. The cloud data center liquid cooling machine room POD hardware device types mainly include servers and various network equipment, and the network equipment mainly includes access switches and aggregation switches. The access switches include service access switches, storage access switches, management access switches and IPMI access switches, and the aggregation switches include service aggregation switches and management aggregation switches. The servers and the access switches, and the access switches and the aggregation switches form a combination through a certain connection relationship and quantity ratio. As shown in Figure 3 , the quantity ratio of the access switches and the servers is usually determined according to the available port quantity of the access switches / the single-side uplink port quantity of the servers, and the quantity ratio of the aggregation switches and the access switches is determined according to the available port quantity of the aggregation switches / the single-side uplink port quantity of the access switches. According to the POD networking information and the machine room parameters, the quantity of the network equipment, the quantity of the servers and the target cabinet quantity of the racks are calculated, and the base minimum construction unit is constructed. According to the port ratio of the network equipment and the base minimum construction unit, the quantity of the servers that can be carried by the base and the quantity of the racks that are pre-occupied by the base are calculated, and the standardized base model is constructed. After receiving the computing power resource construction demand, according to the computing power resource construction demand and the standardized base model, the pre-occupied rack quantity of the base and the matching resources of the machine room are quickly confirmed, the resource demand and supply are matched, the remaining racks and wiring are quickly completed with the base minimum construction unit as the smallest unit, the machine room is deployed, and the agile delivery of the resources is realized.

[0062] The embodiment of the present application plays a great role in the design, supply and delivery stages of the liquid-cooled data center construction project in the intelligent computing scenario through the proactive response of the pre-designed standardized base model to resource demand, and can effectively improve the delivery efficiency of the liquid-cooled data center computing resource project construction, and realize the rapid response and agile delivery of computing resource demand in the intelligent computing scenario.

[0063] In another preferred embodiment, the network device includes an access switch and a convergence switch, and the S2 calculates the number of network devices, the number of servers and the target cabinet number of racks according to the POD networking information and the machine room parameters, and constructs a base minimum construction unit, including:

[0064] S201, calculating the number of servers that a single rack can carry according to the POD networking information, the access switch parameters, the rack parameters and the server parameters;

[0065] S202, calculating the number of access switches and the target cabinet number of racks according to the port ratio of the access switch and the port ratio of the server;

[0066] S203, constructing a base minimum construction unit according to the number of servers that a single rack can carry, the number of access switches and the target cabinet number of racks.

[0067] Specifically, in the embodiment of the present application, the network device includes an access switch and a convergence switch, and the number of servers that a single rack can carry is calculated according to the POD networking information, the access switch parameters, the rack parameters and the server parameters. The number of access switches and the target cabinet number of racks are calculated according to the port ratio of the access switch and the port ratio of the server. Generally, the data center business access switch, storage access switch and management access switch adopt box-type devices, with 40 downstream ports per device, and are deployed in pairs, so that a pair of business access switch / storage access switch / management access switch can access a maximum of 20 servers. The IPMI access switch has 40 downstream ports and can access a maximum of 40 servers. At the same time, in order to facilitate deployment, the business access switch, storage access switch, management access switch and IPMI access switch are deployed in the same cabinet as the server, and the best cabinet mode of the liquid-cooled rack computing resource is determined to be two cabinets, i.e. the target cabinet number is 2, then the number of access switches of the base minimum construction unit Brick is = 2*business access switch + 2*storage access switch + 2*management access switch + IPMI access switch, and the number of servers is = 2*Min(Min(BP1, SP2, MP1, IP), SNp, SNu). In conclusion, the base minimum construction unit Brick model is as shown in Figure 4 Figure 4 ​Figure 1 is a structural schematic diagram of a base minimum construction unit in a machine room deployment method provided by the present application. The gray cable in the figure is a cabinet cable and an inter-cabinet cable reserved according to networking.

[0068] In still another preferred embodiment, the S201 calculates the number of servers that a single rack can carry according to the POD networking information, the access switch parameters, the rack parameters and the server parameters, and includes:

[0069] The S2011 calculates the number of servers that a single access switch can access according to the POD networking information and the port ratio of the access switch.

[0070] The S2012 calculates the number of servers that a single rack can carry under power conditions according to the rated power of the rack, the rated power of the access switch and the rated power of the server.

[0071] The S2013 calculates the number of servers that a single rack can carry under space conditions according to the available U number of the rack, the U number of the access switch and the occupied U number of a single server.

[0072] The S2014 determines the number of servers that a single rack can carry according to the number of servers that a single access switch can access, the number of servers that a single rack can carry under power conditions and the number of servers that a single rack can carry under space conditions.

[0073] Specifically, the embodiment of the present application calculates the number of servers that a single access switch can access according to the POD networking information and the port ratio of the access switch. Exemplarily, the number of product servers that a single business access switch / storage access switch / management access switch / IPMI access switch can access is equal to Min(BP1, SP2, MP1, IP) because the port ratios BP1=BP2 And SP1=SP2 And MP1=MP2. Wherein, BP represents the port ratio of the business access switch and the product server, SP represents the port ratio of the storage access switch and the product server, MP represents the port ratio of the management access switch and the product server, and IP represents the port ratio of the IPMI access switch and the product server.

[0074] The S2012 calculates the number of servers that a single rack can carry under power conditions according to the rated power of the rack, the rated power of the access switch and the rated power of the server. Exemplarily, the number of product servers that a single rack can carry due to power, i.e. the server number SNp=(rack rated power-business access switch rated power-storage access switch rated power-management access switch rated power-IPMI access switch rated power) / server rated power:

[0075]

[0076] Wherein, SNp represents the number of product servers that single rack power can carry, Rp represents single rack power, Sp represents server power, BSWp represents service access switch power, SSWp represents storage access switch power, MSWp represents management access switch power, and IPMISWp represents IPMI access switch power.

[0077] According to the available U number of the rack, the U number of the access switch, and the occupied U number of a single server, the number of servers that a single rack can carry under the space condition is calculated. Exemplarily, the number of product servers that a single rack can carry due to space, that is, the number of servers SNu=(available U number of the rack-access switch U number-storage access switch U number-management access switch U number-IPMI access switch U number) / occupied U number of a single server:

[0078]

[0079] Wherein, SNu represents the number of product servers that single rack space can carry, Ru represents the available U number of single rack space, Su represents the occupied U number of server space, BSWu represents the occupied U number of service access switch space, SSWu represents the occupied U number of storage access switch space, MSWu represents the occupied U number of management access switch space, and IPMISWu represents the occupied U number of IPMI access switch space.

[0080] According to the number of servers that a single access switch can access Min(BP1, SP2, MP1, IP), the number of servers that a single rack can carry under the power condition SNp, and the number of servers that a single rack can carry under the space condition SNu, the number of servers that a single rack can carry is determined to be Min(Min(BP1, SP2, MP1, IP), SNp, SNu).

[0081] In yet another preferred embodiment, S3, according to the port ratio of the network device and the minimum construction unit of the base, calculates the number of servers that the base can carry and the number of racks that the base preoccupies, and constructs a standardized base model, including:

[0082] S301, according to the port ratio of the aggregation switch and the minimum construction unit of the base, calculates the number of the minimum construction units of the base that the base can carry and the number of racks that the base preoccupies;

[0083] S302, according to the number of the minimum construction units of the base and the number of servers that a single rack can carry, calculates the number of servers that the base can carry;

[0084] S303, constructing a standardized base model according to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks that the base preoccupies, and the number of servers that the base can carry.

[0085] Specifically, the embodiment of the present application calculates the number of base minimum construction units that the base can carry and the number of racks that the base preoccupies according to the port ratio of the aggregation switch and the base minimum construction unit based on the POD networking information. Exemplarily, the number of access switches that the service aggregation switch can access is the number of available ports of the aggregation switch / the number of single-side uplink ports of the access switch, i.e., BCP1=BCP2=SCP1=SCP2. Wherein, BCP represents the service aggregation switch and the service access switch connection port ratio, and SCP represents the service aggregation switch and the storage access switch connection port ratio. In combination with Figure 4 The number of bricks contained in a base is Wherein, represents the number of base minimum construction units Brick, and 2 represents the number of service aggregation Brick and management aggregation Brick. The number of bricks that the management aggregation switch (usually a frame device) can access Need to meet Therefore, the number of racks contained in a base is Wherein, 4 represents that the management aggregation Brick and the service aggregation Brick need two additional bricks, and a total of 4 management switch racks; MCP represents the management aggregation switch and the management access switch connection port ratio, and ICP represents the management aggregation switch and the IPMI access switch connection port ratio. According to the number of base minimum construction units and the number of servers that a single rack can carry, the number of servers that the base can carry is calculated. Exemplarily, it can be calculated that a base can carry a product server number: According to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks that the base preoccupies, and the number of servers that the base can carry, a standardized base model is constructed, as shown in Figure 5 Figure 5 is a structure diagram of a standardized base model in a data center deployment method provided by the present application. The yellow cable in the figure is the cabinet line from the service aggregation Brick to the empty cabinet Brick according to the networking reserved, the blue cable is the cabinet line from the management aggregation Brick to the empty cabinet Brick, and the gray cable is the cabinet line in the Brick cabinet and between the cabinets.

[0086] ​In another preferred embodiment, the S303, according to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks that the base preoccupies, and the number of servers that the base can carry, constructs a standardized base model, including:

[0087] S3031, according to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks that the base preoccupies, and the number of servers that the base can carry, constructs a base model;

[0088] S3032, according to the number of servers that the base can carry, determines whether it meets the requirements of the computer room and the business;

[0089] S3033, if it meets, the standardized base model is formed; if it does not meet, the number of converged switches is adjusted, and the size of the base is recalculated.

[0090] Specifically, according to the base minimum construction unit, the number of base minimum construction units that the base can carry, the number of racks that the base preoccupies, and the number of servers that the base can carry, the base model is constructed. According to the number of servers that the base can carry, it is determined whether it meets the requirements of the computer room and the actual business. If it meets, the standardized base model is formed, and the base planning is executed; if it does not meet, the number of business converged switches and management converged switches is increased or decreased, and the size of the base is recalculated.

[0091] In another preferred embodiment, the S4, according to the computing power resource construction demand and the standardized base model, deploys the computer room, including:

[0092] S401, synchronizes the standardized base model to the computer room providing side, preoccupies the rack, and reforms the computer room supporting resources, synchronously preinstalls and wires the network equipment;

[0093] S402, according to the computing power resource construction demand, installs and wires the network equipment and the server according to the standardized base model.

[0094] Specifically, the embodiment of the present application constructs a standardized base model in advance, synchronizes the standardized base model to the machine room providing side, calculates a base pre-occupied rack quantity according to the actual machine room conditions and the calculation method described above, and plans and pre-occupies. In addition to the starting component, the service convergence switch, the management convergence switch, and various access switches are pre-installed according to the networking mode, the network and wiring of the rack are synchronized, and the plug-in delivery of subsequent server resources is realized. The base planning requirements are synchronized to the machine room providing side, the rack and ODF are pre-transformed to ensure the pre-occupation of rack resources, and the pre-delivery of base convergence and access resources is completed. In this way, after receiving the computing power resource construction demand, the servers matching the demand are directly installed on the rack according to the Brick minimum unit in the hard set stage, and the power supply interface and the connection line from the server to the access switch are connected on site, as shown in Figure 6 Figure 6 is a structural schematic diagram after the delivery of the standardized base model in the machine room deployment method provided by the present application, thereby realizing rapid resource delivery.

[0095] As a preferred scheme, the method further comprises:

[0096] If the proportion of base occupied racks exceeds the preset threshold, a capacity expansion mechanism is triggered to supplement pre-occupied rack resources to a complete base.

[0097] Specifically, in actual application of the machine room deployment method provided by the embodiment of the present application, the base planning and pre-occupation are started before the arrival of sudden computing power resource demand, and it is judged whether a new base needs to be built. If a new base needs to be built, the standardized base model is built according to the above method, and resource planning and pre-occupation are performed. If a new base does not need to be built, and the proportion of base occupied racks exceeds the preset threshold, a capacity expansion mechanism is triggered to supplement pre-occupied rack resources to a complete base. Exemplarily, whether the number of base pre-occupied racks is within the safe water level line of 60% (threshold = number of base occupied racks / number of pre-occupied racks of a complete base, which can be set according to actual business demand), if it exceeds the water level line, a capacity expansion mechanism is triggered to supplement rack resources to a complete base. If it is below the safe water level line, the resource demand can be directly matched, and the Brick is used as the minimum delivery unit for delivery.

[0098] Before the arrival of large-scale and sudden computing power resource demand, the embodiment of the present application designs a standardized base model based on machine room information, synchronizes the machine room providing side to plan, pre-occupy and transform the supporting resources of the machine room, quickly responds to the whole process link after the arrival of computing power resource construction demand, and the overall time consumption is 5 months. Compared with the traditional passive response construction method of resource demand, the pre-design method of the base before resource delivery improves the construction project delivery efficiency by nearly 50%, as shown in Figure 2 ​The design method plays a great role in the design, supply and delivery stages of liquid-cooled data center construction projects in the intelligent computing scenario, effectively improves the whole-process delivery efficiency of the construction project, realizes the agile delivery of computing power resource demand, and the specific stage role is as follows:

[0099] (1) MO artificial output: The MO artificial output capacity is pre-positioned before the resource demand arrives, and the calculation standardization base is planned to be pre-occupied, which greatly reduces the artificial sorting time and the uncertainty of the matching resources of the computer room after the sudden demand arrives, realizes the rapid response and matching of the computer room rack and other matching resources to the sudden computing power demand, and guarantees the availability and applicability of the computer room rack.

[0100] (2) Design stage (Design): The base pre-occupied has planned the use of the computer room rack (such as Figure 6 The resource deployment shown by the equipment is put on the shelf in advance, and according to the base planning, the HLD (High-Level Design) and LLD (Low-Level Design) efficiency in the design stage can be significantly improved.

[0101] (3) Supply stage (Supply): After the liquid-cooled computer room is planned, the liquid-cooled server is packaged into a cabinet and delivered, the implementation complexity is minimum, the compatibility is high, the construction project cycle is effectively shortened, and the delivery quality is effectively improved.

[0102] (4) Delivery stage (Delivery): The computer room comprehensive wiring and network equipment such as aggregation switch and access switch are installed and put on the shelf in advance according to the network topology in the design stage, and the cabinet line from the aggregation switch to the rack and the rack line are pre-buried according to the pre-mounted rack. After the server equipment arrives, it is directly and quickly mounted and connected with the cabinet access switch. Compared with the traditional method of mounting and comprehensive wiring after the network equipment and server arrive, the embodiment of the application can greatly improve the delivery efficiency, realize the rapid installation of equipment, and realize the agile delivery.

[0103] Correspondingly, the application also provides a computer room deployment device capable of realizing all processes of the computer room deployment method in the above embodiment.

[0104] Please refer to Figure 7 , Figure 7 is a structure schematic diagram of a preferred embodiment of a computer room deployment device provided by the application. The computer room deployment device comprises:

[0105] The data acquisition module 701 is configured to acquire computer room data center basic physical design unit POD networking information and computer room parameters.

[0106] The unit construction module 702 is configured to calculate the number of network devices, the number of servers, and the target cabinet number of racks according to the POD networking information and the machine room parameters, and construct a base minimum construction unit.

[0107] The model construction module 703 is configured to calculate the number of servers that can be carried by a base and the number of racks that can be preoccupied by the base according to the port ratio of the network device and the base minimum construction unit, and construct a standardized base model.

[0108] The machine room deployment module 704 is configured to deploy a machine room according to a computing resource construction demand and the standardized base model.

[0109] Preferably, the network device includes an access switch and a convergence switch, and the unit construction module 702 includes:

[0110] The first calculation unit 7021 is configured to calculate the number of servers that can be carried by a single rack according to the POD networking information, the access switch parameters, the rack parameters, and the server parameters.

[0111] The second calculation unit 7022 is configured to calculate the number of access switches and the target cabinet number of racks according to the port ratio of the access switch and the port ratio of the server.

[0112] The base construction unit 7023 is configured to construct a base minimum construction unit according to the number of servers that can be carried by a single rack, the number of access switches, and the target cabinet number of racks.

[0113] Preferably, the first calculation unit 7021 is specifically configured to:

[0114] calculate the number of servers that can be accessed by a single access switch according to the POD networking information and the port ratio of the access switch;

[0115] calculate the number of servers that can be carried by a single rack under power conditions according to the rated power of the rack, the rated power of the access switch, and the rated power of the server;

[0116] calculate the number of servers that can be carried by a single rack under space conditions according to the available U number of the rack, the U number of the access switch, and the occupied U number of a single server;

[0117] determine the number of servers that can be carried by a single rack according to the number of servers that can be accessed by a single access switch, the number of servers that can be carried by a single rack under power conditions, and the number of servers that can be carried by a single rack under space conditions.

[0118] Preferably, the model construction module 703 includes:

[0119] The third calculation unit 7031 is configured to calculate the number of the base minimum construction units that can be carried by the base and the number of racks preoccupied by the base according to the port proportion of the convergence switch and the base minimum construction unit.

[0120] The fourth calculation unit 7032 is configured to calculate the number of servers that can be carried by the base according to the number of the base minimum construction units and the number of servers that can be carried by a single rack.

[0121] The model construction unit 7033 is configured to construct a standardized base model according to the base minimum construction unit, the number of the base minimum construction units that can be carried by the base, the number of racks preoccupied by the base and the number of servers that can be carried by the base.

[0122] Preferably, the model construction unit 7033 is specifically configured to:

[0123] construct a base model according to the base minimum construction unit, the number of the base minimum construction units that can be carried by the base, the number of racks preoccupied by the base and the number of servers that can be carried by the base.

[0124] determine whether the number of servers that can be carried by the base meets the demand of the computer room and the business according to the number of servers that can be carried by the base;

[0125] if yes, form the standardized base model; and if no, adjust the number of the convergence switches and recalculate the size of the base.

[0126] Preferably, the computer room deployment module 704 comprises:

[0127] The model synchronization unit 7041 is configured to synchronize the standardized base model to the computer room providing side, preoccupy racks and transform computer room supporting resources, and synchronize preinstallation of the network equipment and wiring.

[0128] The rack installation and wiring unit 7042 is configured to install and wire the network equipment and the servers according to the computing power resource construction demand and according to the standardized base model.

[0129] Preferably, the apparatus is further configured to:

[0130] if the proportion of the used racks of the base exceeds a preset threshold, trigger an expansion mechanism to supplement preoccupied rack resources to a complete base.

[0131] In specific implementations, the working principle, control flow and technical effects of the computer room deployment apparatus provided by the embodiments of the present application correspond to the computer room deployment method in the above embodiments, and thus will not be described herein.

[0132] Please refer toFigure 8 , Figure 8 is a structural schematic diagram of a preferred embodiment of a terminal device provided by the present application. The terminal device comprises a processor 801, a memory 802, and a computer program stored in the memory 802 and configured to be executed by the processor 801, and the processor 801 implements the machine room deployment method of any of the above embodiments when executing the computer program.

[0133] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2, …), which are stored in the memory 802 and executed by the processor 801 to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the terminal device.

[0134] The processor 801 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or the processor 801 can also be any conventional processor. The processor 801 is the control center of the terminal device, and connects various parts of the terminal device through various interfaces and lines.

[0135] The memory 802 mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function, etc., and the data storage area can store related data, etc. In addition, the memory 802 can be a high-speed random access memory, and can also be a non-volatile memory, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., or the memory 802 can also be other volatile solid-state storage devices.

[0136] It should be noted that the above terminal device can include, but is not limited to, a processor and a memory, and those skilled in the art can understand that, Figure 8The structural schematic diagram is merely an example of the terminal device, and does not constitute a limitation on the terminal device, and can include more or fewer components than the diagram, or combine certain components, or different components.

[0137] The embodiment of the present application also provides a computer readable storage medium, which comprises a stored computer program, wherein the computer readable storage medium controls a device where the computer readable storage medium is located to execute the machine room deployment method in any of the above embodiments when the computer program runs.

[0138] The embodiment of the present application also provides a computer program product, which comprises a computer program or computer instructions, and the computer program or the computer instructions are executed by a processor to realize the machine room deployment method in any of the above embodiments.

[0139] The embodiment of the present application provides a machine room deployment method, device, equipment, storage medium and computer program product, by acquiring machine room data center basic physical design unit POD networking information and machine room parameters; according to the POD networking information and the machine room parameters, the number of network devices, the number of servers and the target cabinet number of racks are calculated, and the base minimum construction unit is constructed; according to the port ratio of the network device and the base minimum construction unit, the number of servers that the base can carry and the number of racks that the base preoccupies are calculated, and the standardized base model is constructed; according to the computing power resource construction demand and the standardized base model, the machine room is deployed. The embodiment of the present application plays a great role in the design, supply and delivery stage of liquid cooling data center construction engineering project in the intelligent computing scene through the proactive response of the pre-designed standardized base model to resource demand, can effectively improve the delivery efficiency of the liquid cooling data center computing power resource project construction engineering, and realizes the rapid response and agile delivery of the computing power resource demand in the intelligent computing scene.

[0140] It should be noted that the system embodiments described above are merely illustrative, and the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the connection relationship between the modules in the system embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0141] The above is the preferred embodiment of the present application, it should be pointed out that, for those skilled in the art, without departing from the principles of the present application, can also make a number of improvements and refinements, these improvements and refinements are also considered to be within the scope of the present application.

Claims

1. A method of deploying a machine room, characterized by, The method comprises the following steps: Obtaining the POD networking information and the machine room parameters of the basic physical design unit of the data center; According to the POD networking information and the machine room parameters, the number of network devices, the number of servers and the target cabinet number of racks are calculated, and the minimum construction unit of the base is constructed; According to the port ratio of the network device and the minimum construction unit of the base, the number of servers that the base can carry and the number of racks that the base preoccupies are calculated, and the standardized base model is constructed; According to the computing resource construction demand and the standardized base model, the machine room is deployed; Wherein, the network device includes access switch and aggregation switch, and the number of network devices, the number of servers and the target cabinet number of racks are calculated according to the POD networking information and the machine room parameters, and the minimum construction unit of the base is constructed, including: According to the POD networking information, the access switch parameters, the rack parameters and the server parameters, the number of servers that a single rack can carry is calculated; According to the port ratio of the access switch and the port ratio of the server, the number of access switches and the target cabinet number of racks are calculated; According to the number of servers that a single rack can carry, the number of access switches and the target cabinet number of racks, the minimum construction unit of the base is constructed; According to the port ratio of the aggregation switch and the minimum construction unit of the base, the number of the minimum construction units of the base that the base can carry and the number of racks that the base preoccupies are calculated; According to the number of the minimum construction units of the base and the number of servers that a single rack can carry, the number of servers that the base can carry is calculated; According to the minimum construction unit of the base, the number of the minimum construction units of the base that the base can carry, the number of racks that the base preoccupies and the number of servers that the base can carry, the standardized base model is constructed. According to the POD networking information, the access switch parameters, the rack parameters and the server parameters, the number of servers that a single rack can carry is calculated, including:

2. The machine room deployment method of claim 1, wherein, According to the POD networking information and the port ratio of the access switch, the number of servers that a single access switch can access is calculated; According to the rated power of the rack, the rated power of the access switch and the rated power of the server, the number of servers that a single rack can carry under power condition is calculated; According to the available U number of the rack, the U number of the access switch and the occupied U number of a single server, the number of servers that a single rack can carry under space condition is calculated; According to the number of servers that a single access switch can access, the number of servers that a single rack can carry under power condition and the number of servers that a single rack can carry under space condition, the number of servers that a single rack can carry is determined. ​ 3. The machine room deployment method of claim 1, wherein, The standardized base model is constructed according to the base minimum construction unit, the number of base minimum construction units that can be carried by the base, the number of racks preoccupied by the base and the number of servers that can be carried by the base, and includes: The base model is constructed according to the base minimum construction unit, the number of base minimum construction units that can be carried by the base, the number of racks preoccupied by the base and the number of servers that can be carried by the base; Whether it meets the demand of the computer room and the business is judged according to the number of servers that can be carried by the base; If it meets, the standardized base model is formed; if it does not meet, the number of aggregation switches is adjusted, and the size of the base is recalculated.

4. The machine room deployment method of claim 1, wherein, The computer room is deployed according to the computing resource construction demand and the standardized base model, and includes: The standardized base model is synchronized to the computer room providing side, the preoccupation of the rack and the transformation of the computer room supporting resources are carried out, the preinstallation of the network equipment and the wiring are carried out synchronously; According to the computing resource construction demand, the network equipment and the server are installed and wired according to the standardized base model.

5. The machine room deployment method of claim 4, wherein, The method further includes: If the proportion of racks used by the base exceeds the preset threshold, a capacity expansion mechanism is triggered to supplement the preoccupied rack resources to a complete base.

6. A machine room deployment device, characterized by, It includes: A data acquisition module is configured to acquire POD networking information and computer room parameters; A unit construction module is configured to calculate the number of network devices, the number of servers and the target cabinet number of racks according to the POD networking information and the computer room parameters, and to construct a base minimum construction unit; A model construction module is configured to calculate the number of servers that can be carried by the base and the number of racks preoccupied by the base according to the port ratio of the network device and the base minimum construction unit, and to construct a standardized base model; A computer room deployment module is configured to deploy the computer room according to the computing resource construction demand and the standardized base model; If the network device includes an access switch and an aggregation switch, the unit construction module includes: A first calculation unit is configured to calculate the number of servers that can be carried by a single rack according to the POD networking information, the access switch parameters, the rack parameters and the server parameters; A second calculation unit is configured to calculate the number of access switches and the target cabinet number of racks according to the port ratio of the access switch and the port ratio of the server; A base construction unit is configured to construct a base minimum construction unit according to the number of servers that can be carried by a single rack, the number of access switches and the target cabinet number of racks; The model construction module includes: A third calculation unit is configured to calculate the number of base minimum construction units that can be carried by the base and the number of racks preoccupied by the base according to the port ratio of the aggregation switch and the base minimum construction unit; A fourth calculation unit is configured to calculate the number of servers that can be carried by the base according to the number of base minimum construction units and the number of servers that can be carried by a single rack. A model construction unit is configured to construct a standardized base model according to the base minimum construction unit, the number of base minimum construction units that can be carried by the base, the number of racks preoccupied by the base, and the number of servers that can be carried by the base.

7. A terminal device, characterized by, The computer program is configured to be executed by the processor, and the processor implements the method for deploying a computer room according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and a device where the computer readable storage medium is located implements the method for deploying a computer room according to any one of claims 1 to 5 when executing the computer program.

9. A computer program product, characterised in that, The computer program product includes a computer program or computer instructions, and the computer program or the computer instructions implement the method for deploying a computer room according to any one of claims 1 to 5 when executed by a processor.

Images