CN107729141B

CN107729141B - Service distribution method, device and server

Info

Publication number: CN107729141B
Application number: CN201710888856.0A
Authority: CN
Inventors: 张鸣
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2017-09-27
Filing date: 2017-09-27
Publication date: 2022-06-10
Anticipated expiration: 2037-09-27
Also published as: CN107729141A

Abstract

The invention provides a service distribution method, a device and a server, wherein the method and the device are used for the server, the server comprises service distribution equipment and service processing equipment, the service processing equipment comprises a plurality of service processing module groups, the service distribution equipment distributes the services to be processed to specific service processing modules in the service processing module groups in a centralized manner, so that the service processing modules are in a high power consumption state to the maximum extent, and other service processing modules in the same service processing module group are in a low power consumption state. The method provided by the invention does not need all the service processing modules to work simultaneously, so that all the radiating assemblies do not need to be started, and only the radiating assemblies corresponding to the service processing modules in the working state need to be started, thereby improving the overall radiating efficiency of the server.

Description

Service distribution method, device and server

Technical Field

The present application relates to the field of servers, and in particular, to a service allocation method, an apparatus, and a server.

Background

The resource pooling server comprises at least one service module group, and different service module groups are used for processing services corresponding to different service models. Each service module group includes a plurality of service treatment cooling zones. Each business process cooling zone includes a business process module and a heat dissipation assembly. Each service processing cooling zone processes service through the service processing module it comprises. The heat dissipation assembly is used for dissipating heat of the service processing module in the service processing cooling area.

One conventional service allocation mode is a polling allocation mode, i.e., when the service is allocated for polling, a plurality of services having the same service model are allocated to different service processing cooling zone processes in the service module group. And when each service processing cooling area in the service module group is allocated to one service, ending the service allocation polling and starting the next service allocation polling. The processing mode of the next service distribution polling is the same as that of the current service distribution polling, and the polling distribution mode of the service is realized in a circulating way.

The power consumption of the resource pooling server comes from two aspects, namely the power consumption generated by the operation of the service processing module and the power consumption generated by the operation of the heat dissipation assembly. If the service is distributed according to the polling distribution mode, when the number of the services is large, each service processing cooling area in the service processing module group is distributed to the service, and the service processing module in each service processing cooling area is in a working state. Since the service processing modules start to operate and generate power consumption, the service processing modules in operation in each service processing cooling area need the heat dissipation assemblies in the service processing cooling areas to dissipate heat of the service processing modules. That is, all the heat dissipation assemblies in the service processing module group need to be activated, which results in large total power consumption of the service processing module group, and further results in large total power consumption of the resource pooling server and low heat dissipation efficiency.

Disclosure of Invention

The application provides a service distribution method, a service distribution device and a server, and aims to solve the problems of high power consumption and low overall heat dissipation efficiency of the server.

The application provides a service distribution method, which is used for a server, wherein the server comprises service distribution equipment and service processing equipment, the service processing equipment comprises at least one service processing module group, and the service processing module group comprises at least two service processing modules.

The method comprises the following steps: the service distribution equipment determines a service to be processed; the service distribution equipment determines a service processing module group for processing the service to be processed; and the service distribution equipment distributes the service to be processed to the service processing module with the maximum power consumption and without reaching the upper limit in the service processing module group.

The server sets a plurality of mutually independent service processing module groups, such as a calculation module group, a storage module group and an interface module group, in the service processing equipment according to the service model. The service distribution equipment distributes the service to be processed to a specific service processing module of the service processing module group in a centralized manner, so that the service processing module is in a high power consumption state to the maximum extent, and other service processing modules in the same service processing module group are in a low power consumption state. According to the method, all the service processing modules do not need to work simultaneously, so that all the radiating components do not need to be started, and only the radiating components corresponding to the working service processing modules need to be started, so that the radiating power consumption can be reduced, and the overall radiating efficiency of the server is improved.

In one implementation, the determining, by the service allocating device, the pending service includes: the service distribution equipment acquires a service request, wherein the service request comprises a service model identifier and a service to be processed corresponding to the service model identifier; and the service distribution equipment analyzes and obtains the service to be processed from the service request according to the service model identifier included in the service request.

Optionally, the method further includes querying, by the service distribution device, a value of a working duration parameter of the service processing module, where the value of the working duration parameter indicates a duration that the service processing module is in a working state in a current working period; if the value of the working duration parameter of the service processing module reaches the working duration preset by a working period, the service distribution equipment updates the working period number of the service processing module, the updated working period number is greater than the working period number before updating, and the updated working period number represents that the service processing module is in the next working period.

Optionally, if the number of the service processing modules in the service processing module group that consume the most power and do not reach the upper limit is greater than one, the method further includes: the service distribution equipment confirms a service processing module with the minimum working period number in all the service processing modules which have the maximum power consumption and do not reach the upper limit and are confirmed from the service processing module group; and the service distribution equipment distributes the service to be processed to the service processing module with the minimum working period number.

The application further provides another service allocation method, where the service allocation method is used for a server, the server includes a service allocation device and a service processing device, the service processing device includes at least one service processing module group, and the service processing module group includes at least two service processing modules.

The method comprises the following steps: the service distribution equipment determines a service to be processed; the service distribution equipment determines a service processing module group for processing the service to be processed; the service distribution equipment determines a candidate processing module, wherein the candidate processing module is the service processing module with the minimum working cycle number in the service processing module group; and if the number of the candidate processing modules is more than 1, the service distribution equipment distributes the service to be processed to the service processing module which has the maximum power consumption and does not reach the upper limit in the candidate processing modules.

According to the business allocation method, the allocation of the to-be-processed business is centralized and balanced, so that the business processing module for processing the to-be-processed business of the same business model can intensively process the to-be-processed business to the greatest extent, thereby reducing the power consumption of the business processing equipment and improving the heat dissipation efficiency.

Optionally, the method further comprises: and if the number of the candidate processing modules is equal to 1, the service distribution equipment distributes the service to be processed to the candidate processing modules.

Optionally, the determining, by the service distribution device, the service to be processed includes: the service distribution equipment acquires a service request, wherein the service request comprises a service model identifier and a service to be processed corresponding to the service model identifier; and the service distribution equipment analyzes and obtains the service to be processed from the service request according to the service model identifier included in the service request.

Optionally, the method further comprises: the service distribution equipment inquires the value of a working duration parameter of the service processing module, wherein the value of the working duration parameter indicates the duration of the service processing module in a working state in the current working period; if the value of the working duration parameter of the service processing module reaches the working duration preset by a working period, the service distribution equipment updates the working period number of the service processing module, the updated working period number is greater than the working period number before updating, and the updated working period number represents that the service processing module is in the next working period.

The present application further provides a service distribution apparatus, including: a service determining unit, configured to determine a service to be processed; a module group determining unit, configured to determine a service processing module group for processing the service to be processed; and the service distribution unit is used for distributing the service to be processed to the service processing module with the maximum power consumption and without reaching the upper limit in the service processing module group.

In one implementation, the traffic determination unit includes: a service request obtaining subunit, configured to obtain a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier; and the service analysis subunit is used for analyzing and obtaining the service to be processed from the service request according to the service model identifier included in the service request.

In another implementation manner, the apparatus further includes a duty cycle counting unit, configured to query a value of a working duration parameter of the service processing module, where the value of the working duration parameter indicates a duration that the service processing module is in a working state in a current duty cycle; if the value of the working time length parameter of the service processing module reaches the working time length preset by a working period, the working period counting unit is further configured to update the working period number of the service processing module, where the updated working period number is greater than the working period number before updating, and the updated working period number indicates that the service processing module is in the next working period.

Optionally, the service distribution unit further includes: a module determining subunit, configured to determine, if the number of service processing modules in the service processing module group that consume the most power and do not reach the upper limit is greater than one, a service processing module with the smallest number of working cycles among all the service processing modules that have been determined from the service processing module group that consume the most power and do not reach the upper limit; and the service distribution subunit is used for distributing the service to be processed to the service processing module with the minimum working cycle number.

The present application also provides another service distribution apparatus, including: a service determining unit, configured to determine a service to be processed; a module group determining unit, configured to determine a service processing module group for processing the service to be processed; the service distribution unit is used for determining a candidate processing module, and the candidate processing module is the service processing module with the minimum working cycle number in the service processing module group; and if the number of the candidate processing modules is more than 1, distributing the service to be processed to the service processing module which has the maximum power consumption and does not reach the upper limit in the candidate processing modules.

Optionally, the service distribution unit is further configured to allocate the service to be processed to the candidate processing module if the number of candidate processing modules is equal to 1.

Optionally, the service determination unit includes: a service request obtaining subunit, configured to obtain a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier; and the service analysis subunit is used for analyzing and obtaining the service to be processed from the service request according to the service model identifier included in the service request.

Optionally, the apparatus further includes a duty cycle counting unit, configured to query a value of a working duration parameter of the service processing module, where the value of the working duration parameter indicates a duration that the service processing module is in a working state in a current duty cycle; if the value of the working time length parameter of the service processing module reaches the working time length preset by a working period, the working period counting unit is further configured to update the working period number of the service processing module, where the updated working period number is greater than the working period number before updating, and the updated working period number indicates that the service processing module is in the next working period.

The application further provides a service allocation device, which includes a data interface and a service allocation control module, where the data interface is configured to receive a service request sent by a client, and the service allocation control module is configured to: determining a service to be processed; determining a service processing module group for processing the service to be processed; and distributing the service to be processed to the service processing module with the maximum power consumption and without reaching the upper limit in the service processing module group.

The present application further provides another service allocation device, including a data interface and a service allocation control module, where the data interface is configured to receive a service request sent by a client, and the service allocation control module is configured to: determining a service to be processed; determining a service processing module group for processing the service to be processed; determining a candidate processing module, wherein the candidate processing module is the service processing module with the minimum working cycle number in the service processing module group; and if the number of the candidate processing modules is more than 1, the service distribution control module is used for distributing the service to be processed to the service processing module which has the maximum power consumption and does not reach the upper limit in the candidate processing modules.

The application also provides a server, which comprises the service distribution equipment and the service processing equipment. The service processing equipment comprises at least one service processing module group, and the service processing module group comprises at least two service processing modules.

The present application also provides a computer-readable storage medium comprising instructions which, when run on a computer, cause the computer to perform the method as described above.

The present application also provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method as described above.

Drawings

Fig. 1 is a schematic structural diagram of a resource pooling server according to this embodiment;

fig. 2 is a schematic structural diagram of a service processing device provided in this embodiment;

fig. 3 is a flowchart of a service allocation method provided in this embodiment;

FIG. 4 is a graph illustrating the relationship between the additional power consumption increase caused by the temperature increase of each CPU;

fig. 5 is a flowchart of another service allocation method according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a service distribution apparatus according to an embodiment;

fig. 7 is a schematic structural diagram of a service distribution device provided in this embodiment;

Fig. 8 is a schematic structural diagram of the service determination unit provided in this embodiment.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention.

The present embodiment provides a service allocation method, where the method is applied to a server, and the service allocation method provided in the present embodiment is described below with reference to a resource pooling server as an example.

Referring to fig. 1, fig. 1 is a schematic structural diagram of a resource pooling server provided in this embodiment. As shown in fig. 1, in the present embodiment, the resource pooling server includes a service allocating apparatus 100 and a service processing apparatus 200. The service distribution apparatus 100 is connected to the service processing apparatus 200. The service distribution device 100 is configured to distribute the service to be processed to the service processing device 200.

In the resource pooling server, the service processing device 200 includes at least one service processing module group 201, and different service processing module groups are used for processing services to be processed corresponding to different service models. In the service processing device 200, each service processing module group 201 includes a service processing switching module 211, at least two service processing modules 212, and heat dissipation assemblies 213 corresponding to each service processing module 212 one to one. The service processing module group 201 sends the service to be processed to the service processing module 212 through the service processing switching module 211. The service processing module 212 is configured to process the service to be processed. Each heat dissipation assembly 213 is used to dissipate heat for the corresponding business processing module 212. The service processing module 212 and the heat dissipation assembly 213 corresponding to the service processing module 212 form a service processing cooling area 214.

In each service processing cooling area 214, each service processing module 212 is provided with a set of heat dissipation assemblies 213. The heat sink assembly 213 includes a heat sink controller, a power consumption sensor, a temperature sensor, and a heat sink device.

The power consumption sensor in the service processing cooling area 214 is configured to monitor the power consumption of the service processing module 212 in the service processing cooling area 214 in real time, and report the monitored power consumption to the service distribution device 100 and the heat dissipation controller in real time through the service processing switching module 211. The temperature sensor in the service processing cooling area 214 is configured to monitor a temperature value of the service processing module 212 in the service processing cooling area 214 in real time, and report the temperature value to the heat dissipation controller. The heat dissipation controller is used for controlling the heat dissipation device to perform heat dissipation operation, and the heat dissipation device is used for example as a fan, a cooling device and the like.

The heat dissipation controller adjusts the heat dissipation intensity of the heat dissipation device, such as adjusting the rotation speed of the fan or adjusting the flow rate of the cooling liquid in the cooling device, according to the operation state of the service processing module 212, such as power consumption and/or temperature, so as to dissipate heat from the service processing module 212 in the service processing cooling area 214, thereby achieving targeted differential heat dissipation.

Optionally, in an embodiment, the service processing device 200 may include three service processing module groups, for example, please refer to fig. 2, and fig. 2 shows a schematic structural diagram of the service processing device provided in this embodiment. As shown in fig. 2, the service processing apparatus 200 includes three service processing module groups, which are a computing module group 221, a storage module group 222, and an interface module group 223. The computing module group 221 is for processing computing traffic, the storage module group 222 is for processing storage traffic, and the interface module group 223 is for processing interface traffic. The interface service refers to a service for performing network intercommunication between the client and the external network.

The computing module group 221 includes a business process switching module 2211 and three computing cooling zones 2214 connected thereto. And, the storage module set 222 includes a service processing switching module 2221 and three storage cooling sections 2224 connected thereto. The interface module group 223 includes a service processing switching module 2231 and three interface cooling zones 2234 connected thereto.

In the set of computing modules 221, a computing module 2212 and a set of heat sinks 2213 corresponding to the computing module 2212 form a computing cooling area 2214. In the storage module group 222, one storage module 2222 and a set of heat dissipation assemblies 2223 corresponding to the storage module 2222 constitute one storage cooling area 2224. In the interface module group 223, one interface module 2232 and a group of heat dissipation assemblies 2233 corresponding to the interface module 2232 form one interface cooling zone 2234.

Each computing module 2212 includes a Central Processing Unit (CPU) and a Double Data Rate synchronous dynamic random access memory (DDR) connected thereto. The CPU and DDR are used together for compute interface type traffic. In particular, the CPU is used to interpret computer instructions and process data in computer software. The DDR is used to temporarily store data processed by the CPU. The heat dissipation assembly 2213 is used to dissipate heat from the computing modules 2212 within the same cooling zone. The heat sink assembly 2213 includes a first heat sink, a first temperature sensor, a first heat sink controller, and a first power consumption sensor. The first heat sink may be a device for performing a heat dissipating operation, for example, a fan, a liquid cooling device. The first temperature sensor may be a device for monitoring the temperature of the calculation module 2212 in the same calculation cooling area in real time and feeding back the monitored temperature value to the first heat sink controller in real time, or may be another device capable of achieving the same function. The first power consumption sensor is configured to monitor power consumption of the calculation module 2212 in the same calculation cooling area in real time, and feed back the monitored power consumption value to the service distribution device 100 in real time.

Each storage module 2222 includes three connected Solid State Drives (SSDs). The SSD is used to handle storage type traffic, in particular, to store data information. The heat dissipation assembly 2223 serves to dissipate heat of the memory module 2222. The heat dissipation assembly 2223 includes a second heat sink, a second temperature sensor, a second heat sink controller, and a second power consumption sensor. The second heat sink may be a device for performing a heat dissipating operation, such as a fan, a liquid cooling device. The second temperature sensor may be a device for monitoring the temperature of the storage module 2222 in the same storage cooling area in real time, and feeding back the monitored temperature value to the second heat sink controller in real time, or may be another device capable of achieving the same function. The second power consumption sensor is configured to monitor power consumption of the storage modules 2222 in the same storage cooling area in real time, and feed back a monitored power consumption value to the service distribution apparatus 100 in real time.

Each interface module 2232 includes two connected network adapters (NICs). The NIC is configured to process an interface type service, and specifically, by being connected to a switching module in the resource pooling server, information intercommunication between any service processing modules and between all service processing modules and the service allocation module is achieved. The heat sink assembly 2233 is used to dissipate heat from the interface module 2232. The heat sink assembly 2233 includes a third heat sink, a third temperature sensor, a third heat sink controller, and a third power consumption sensor. The third heat sink may be a device for performing a heat dissipating operation, for example, a fan, a liquid cooling device. The third temperature sensor may be a device for monitoring the temperature of the interface module 2232 in the same interface cooling area in real time and feeding back the monitored temperature value to the third heat sink controller in real time, or other devices capable of achieving the same function. The third power consumption sensor is configured to monitor power consumption of the interface module 2232 in the same interface cooling area in real time, and feed back the monitored power consumption value to the service distribution apparatus 100 in real time.

Referring to fig. 3, fig. 3 is a flowchart of a service allocation method provided in this embodiment. As shown in fig. 3, the present embodiment provides a service allocation method, which is executed by the service allocation device 100 shown in fig. 1, and specifically includes:

step S100: the service distribution device 100 determines a service to be processed.

The service model to which the service to be processed belongs may be any one of a calculation service model, a storage service model or an interface service model, or may be another service model.

Illustratively, a specific implementation manner of step S100 is that the service distribution device 100 obtains a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier. The service distribution device 100 analyzes the service request according to the service model identifier to obtain the service to be processed corresponding to the service model identifier. Optionally, the service distribution device 100 parses the service request to obtain one or more to-be-processed services corresponding to the service model identifier.

For example, the service request may include one or more to-be-processed services corresponding to any one of the multiple service models, or the service request includes a combination of one or more to-be-processed services corresponding to at least two of the multiple service models, respectively. For example, the service request may include one or more to-be-processed services corresponding to any one of a computation service model, a storage service model, or an interface service model, and may further include a combination of one or more to-be-processed services corresponding to at least two of the computation service model, the storage service model, and the interface service model.

If the service request is a combination of one or more to-be-processed services corresponding to multiple service models respectively, and the service request includes one or more to-be-processed services corresponding to different service models respectively, the service distribution device 100 parses the service request according to different service model identifiers included in the service request to obtain one or more to-be-processed services corresponding to different service models. For example, if the service request includes both one or more computing services corresponding to the computing service model identifier and one or more interface services corresponding to the interface service model identifier, the service distribution device 100 parses the service request to obtain the one or more computing services and obtain the one or more interface services. The service distribution device 100 takes each obtained service as a service to be processed. If the service request includes one or more computation services corresponding to the computation service model identifier, one or more interface services corresponding to the interface service model identifier, and one or more storage services corresponding to the storage service model identifier, the service distribution device 100 parses the service request to obtain one or more computation services, one or more storage services, and one or more interface services.

Step S200: the service distribution apparatus 100 determines a service processing module group for processing the service to be processed.

And selecting a corresponding service processing module group according to the service model of the service to be processed, wherein the service model of the service to be processed is the same as the service model of the service which can be processed by the service processing module group. For example, based on different service processing module groups in the service processing apparatus 200 shown in fig. 2, if the service to be processed is a computing service, determining that the computing module group 221 is used for processing the computing service; if the service to be processed is a storage service, determining that the storage module group 222 is used for processing the storage service; if the service to be processed is an interface service, the interface module 223 is configured to process the interface service.

The service distribution device 100 pre-stores a corresponding relationship between a service model identifier and the service processing module group identifier, where the service processing module group identifier is used to identify a service processing module group that processes services corresponding to different service models. The service distribution device 100 determines the service processing module group for processing the service to be processed according to the corresponding relationship between the pre-stored service model identifier and the service processing module group identifier.

Step S300: the service distribution device 100 distributes the service to be processed to the service processing module with the largest power consumption and not reaching the upper limit in the service processing module group.

After determining the service processing module group, the service distribution device 100 determines a service processing module with the maximum power consumption and not reaching the upper limit, in an implementation manner, that is, the service distribution device 100 monitors the power consumption of all the service processing modules in the service processing device 200, queries the service processing module with the maximum power consumption from the service processing module group, selects the service processing module with the power consumption not reaching the upper limit from the service processing module with the maximum power consumption, and distributes the service to be processed to the service processing module. If there are multiple service processing modules with the same power consumption and the largest power consumption in the same service module group, the service allocation apparatus 100 may randomly allocate the service to be processed to any one of the multiple service processing modules with the same power consumption and the largest power consumption. It is also possible to select one of the traffic processing modules with the smallest number of cycles among all the traffic processing modules that have been identified from the group of traffic processing modules that have the largest power consumption and that have not reached the upper limit. Then, the service allocating device 100 allocates the service to be processed to the service processing module with the minimum number of working cycles. Alternatively, if the number of the service processing modules with the minimum number of the operation cycles is more than one, the service allocating apparatus 100 may allocate the service to be processed to one with the maximum value of the operation duration parameter.

The upper limit may be a preset power consumption upper limit for the service processing module. And if the power consumption of the service processing module with the maximum power consumption does not reach the upper limit, distributing the service to be processed to the service processing module.

And if the power consumption of the service processing module with the maximum power consumption reaches the upper limit, suspending the distribution of the service to be processed.

If the number of the service processing modules with the largest power consumption is multiple, the service to be processed is randomly allocated to one of the service processing modules. Optionally, the service allocating device 100 allocates the service to be processed to the service processing module with the largest working time parameter value among the plurality of service processing modules with the largest power consumption.

The upper limit is much larger than the power consumption of the traffic to be processed. Optionally, the upper limit of the power consumption is at least four orders of magnitude greater than the service to be processed, so that the power consumption of the service to be processed is a negligible value with respect to the upper limit. In this case, as long as the power consumption of the candidate processing module does not reach the preset upper limit, when a new service to be processed needs to be processed, the new service to be processed is allocated to the candidate processing module for processing, without affecting the normal processing capability of the candidate processing module.

The traditional service allocation mode is a polling allocation mode, if the to-be-processed services are allocated according to the polling allocation mode, when the number of the to-be-processed services is large, each service processing cooling area in the service processing module group is allocated to the to-be-processed services, and the service processing module in each service processing cooling area is in a working state. Since the service processing modules start to operate and generate power consumption, the service processing modules operating in each service processing cooling area need a heat dissipation assembly to dissipate heat of the service processing modules. That is, all the heat dissipation assemblies in the service processing module group need to be started, which results in large total power consumption of the service processing module group, and therefore, the total power consumption of the server is large, and the heat dissipation efficiency is low.

The embodiment of the present invention provides different service allocation modes, where after determining a service to be processed, the service allocation device 100 determines a service processing module group for processing the service to be processed. Then, the service allocation device 100 allocates the service to be processed to the service processing module with the largest power consumption and not reaching the upper limit in the service processing module group. Therefore, the to-be-processed services of the same service model are intensively distributed to the same service processing module to the greatest extent, and are not distributed to other service processing modules for processing the to-be-processed services of the same service model, so that all the service processing modules do not need to work simultaneously, all the radiating assemblies do not need to be started simultaneously, and only the radiating assemblies corresponding to the working service processing modules need to be started, so that the total power consumption of the service processing modules is reduced, the extra power consumption is reduced, and the radiating efficiency of the server is improved.

The service distribution mode provided by the implementation can obviously reduce extra power consumption. Taking the computing module group 221 as an example, the computing services to be processed may be allocated according to the method provided by this embodiment, that is, all the computing services are allocated to the service processing module with the largest power consumption and not reaching the upper limit (for example, when the service to be processed is a computing service, the service processing module is a computing module for processing the computing service), or may be allocated according to a polling allocation mode and allocated to one computing module respectively.

If the computing service is allocated in the manner provided in this embodiment, assuming that the computing service to be processed increases the power consumption of the server by nW, the temperature of the CPU in the computing module processing the computing service will increase by T1, so that the additional power consumption caused by the CPU increases by P1, and if the service is allocated in the polling allocation mode, the temperature of the CPU in each computing module will increase by T2, so that the additional power consumption caused by each CPU increases by P2. Since the temperature rise value has a great influence on the additional power consumption increase value brought by each CPU, the greater the temperature rise value is, the greater the additional power consumption increase value brought by each CPU is, and the additional power consumption increase value brought by each CPU increases with the increase of the temperature rise value, similar to exponential increase, specifically, see fig. 4, which is a schematic diagram illustrating a relationship change of the additional power consumption increase caused by each CPU due to the temperature rise. As can be seen from fig. 4, the larger the value of n is, the easier it is to determine that n × P2> P1, and therefore, compared with the polling distribution mode, the distribution mode provided by the present embodiment distributes the pending traffic, and the total power consumption of the server is low.

It should be noted that, in this application, the number of work cycles indicates the number of accumulated work cycles of the business processing module. One working period represents that the accumulated working time of the service processing module reaches the preset working time. For example, with day or 24 hours as the work cycle unit, if the work cycle number is 1, it means that the business processing module is in the first work cycle, and the work duration reaches or exceeds 24 hours, but does not exceed 2 days or 48 hours. The value of the working duration parameter indicates the duration of the service processing module in the working state in the current working period. For example, if the value of the operating duration parameter is 12, it indicates that the duration of the operating state of the service processing module in the current operating period is 12 hours. Optionally, the working state of the service processing module represents a state when the service processing module processes a service, and the service processing module may process the service continuously or discontinuously.

In an implementation manner, the method for counting the number of working cycles of all the service processing modules in this embodiment includes: if the value of the working duration parameter of the service processing module reaches the preset accumulated continuous working duration, the service distribution device 100 updates the working cycle number of the service processing module, wherein the updated working cycle number is greater than the working cycle number before updating, and the updated working cycle number indicates that the service processing module is in the next working cycle.

Specifically, if the value of the working duration parameter of the service processing module reaches the accumulated continuous working duration preset in one working period, the working period number of the service processing module is increased by a preset granularity, where the preset granularity may be 1 or 2 or any value, and is not limited by this embodiment. For example, the number of the working cycles of the service processing module is increased by 1, and the service to be processed is suspended from being allocated to the service processing module, if the continuous working duration of the service processing module does not reach the preset continuous working duration, the number of the working cycles of the service processing module keeps the current value, and the service processing module keeps continuing to work. The continuous work keeping can be in a work keeping state, the service is allowed to be processed completely, and the service to be processed is allowed to be continuously distributed for the service.

For example, a duty cycle is set to be 24h, and a counting method of the duty cycle is described by taking a calculation module as an example: starting timing from the first computing service processed by the computing module, distributing all computing services to the computing module without distributing all computing services to other computing modules until the computing module accumulatively and continuously works for 24h, increasing 1 for the number of working cycles of the computing module, and resetting the timing to zero after the number of working cycles is updated. When the cumulative continuous work of the computing module is less than 24h, the working state of the computing module is kept unchanged, namely, the computing service which is allowed to be processed by the computing module is naturally ended, and a new computing service is also allowed to be distributed to the computing module.

For the service processing modules for processing different service model services, the preset accumulated continuous working time lengths corresponding to one working period can be the same or different. For example, one duty cycle of both the computing module and the storage module may be 24 h; it is also possible that one duty cycle of the calculation module is 24h and one duty cycle of the storage module is 48 h. The service life of the different types of service processing modules is maximized by adapting to the full-load operation capacity of the different types of service processing modules.

The method of the present embodiment is described below by taking an example of processing a service corresponding to a computational service model. In this embodiment, the computing module group 221 includes eight computing modules, which are respectively a computing module S1, a computing module S2, a computing module S3, a computing module S4, a computing module S5, a computing module S6, a computing module S7, and a computing module S8, a unit of a working cycle of each computing module is 24 hours, an upper limit of power consumption is 300W, and current working cycle numbers and current power consumption values of the computing modules are shown in table 1 below, where the working cycle numbers of the computing modules S1, S5, and S6 are all T1, which indicates that the computing modules S1, S5, and S6 are currently in a first working cycle, the working cycle numbers of the computing modules S2, S4, and S7 are T2, which indicate that the computing modules S2, S4, and S7 are currently in a second working cycle, the accumulated working cycle numbers of the computing modules S3 and S8 are T3, which indicate that the computing modules S3 and S8 are currently in a third working cycle.

TABLE 1

For the example of table 1, according to the flow chart of the service allocation method shown in fig. 3, in the method of the present embodiment, the calculation modules that determine the maximum power consumption are the calculation module S7 and the calculation module S8. Pending computing traffic may then be assigned to any of the computing module S7 and the computing module S8. Alternatively, the computing service to be processed may be allocated to the computing block S7 with the smaller number of working cycles in the computing block S7 and the computing block S8.

If the service processing modules in the service processing module group are in a working state for a long time, a heat dissipation fault or a temperature rise is easily caused, and further, power consumption of the server is additionally increased, so that another service allocation method is provided in the embodiment of the present invention, as shown in fig. 5, including:

step S400: the service distribution device 100 determines a service to be processed.

For a specific implementation of this step, refer to step S100, and specific details are not described herein.

Step S500: the service distribution apparatus 100 determines a service processing module group for processing the service to be processed. For a specific implementation manner of this step, refer to step S200, and specific details are not described herein again.

Step S600: the service distribution apparatus 100 determines a candidate processing module, which is a service processing module with the minimum number of working cycles in the service processing module group.

One way to determine candidate processing modules is that the service distribution apparatus 100 counts the working cycles of all the service processing modules in the service processing apparatus 200, and counts the working cycles of all the service processing modules. After determining the service processing module group, inquiring the service processing module with the minimum working cycle number in the service processing module group, and taking the service processing module with the minimum working cycle number as a candidate processing module. In this step, the implementation manner of the working cycle number may refer to the implementation manner of the working cycle number of all the service processing modules in the service processing module group in step S300.

Step S700: and if the number of the candidate processing modules is more than 1, the service distribution unit distributes the service to be processed to the candidate processing modules of which the power consumption does not reach the upper limit.

For a specific implementation manner of this step, reference may be made to the method in which the service allocation device 100 allocates the service to be processed to the service processing module with the largest power consumption and not reaching the upper limit in the service processing module group in step S300, and specifically, details of determining that the power consumption is the largest and not reaching the upper limit are not described herein again.

Optionally, if the number of candidate processing modules is equal to 1, the to-be-processed service is allocated to the candidate processing module.

In this embodiment, the computing module group 22 includes three computing modules, which are respectively a first computing module, a second computing module, and a third computing module, the initial values of the working cycles of the first computing module, the second computing module, and the third computing module are all 0, each working cycle is 24 hours, and the upper limit of the power consumption is 300W.

The service distribution unit in the server firstly analyzes and obtains a service model from the received service request according to the service model identification, wherein the service model is a to-be-processed service of the computing service.

If the power consumption of only one computing module in the computing module group 221 does not reach the upper limit, for example, the first computing module, the to-be-processed computing service is allocated to the first computing module. When the power consumption of the first computing module reaches 300W (that is, the power consumption reaches the upper limit), if a new computing service needs to be processed, since the power consumption of each computing module reaches the upper limit, the subsequent new computing service is allocated according to the polling allocation mode, that is, the service allocation device 100 allocates the new computing service to the second computing module.

As time goes on, some computing services in the first computing module are processed, so that the power consumption of the first computing module is less than 300W (i.e., the first computing module is recovered to be in an unloaded state), and the accumulated continuous working time of the first computing module has not reached one working period (24 hours), at this time, two computing modules whose power consumption has not reached the upper limit exist in the computing module group 221. Since the speed of ending the original service is often much less than the growth speed of the new service, the power consumption of the first computing module is usually greater than the power consumption of the second computing module, for example, the power consumption of the first computing module is 290W, and the power consumption of the second computing module is 200W, so if a new computing service needs to be allocated, the new computing service is allocated to the first computing module. If the power consumption of the first computing module is equal to that of the second computing module, new computing services are randomly distributed to the first computing module or the second computing module. Or new calculation services are distributed to the first calculation module with the larger value of the working time length parameter.

When the first calculation module accumulates continuous work for 24 hours, the work cycle number of the first calculation module is updated to be 1, the work cycle numbers of the other calculation modules are still 0, and when the service distribution equipment 100 receives new calculation service needing distribution, a calculation module for processing the new calculation service is selected from the calculation modules with the work cycle number of 0. Therefore, the service to be processed is distributed to the same service processing module to the maximum extent, and the centralized distribution of the service to be processed is realized.

The method of the embodiment is described below by taking an example of processing a service corresponding to the computational service model. The service allocation method shown in fig. 5 is described with reference to the current number of working cycles and the current power consumption values of the eight computing modules in the computing module group 221 shown in table 1. For the example in table 1, the method shown in fig. 5 first determines that there are three calculation modules with the smallest number of working cycles, namely, the calculation module S1, the calculation module S5, and the calculation module S6, then determines the calculation module S6 with the largest power consumption and not reaching the upper limit from the three calculation modules, and allocates a new calculation service to the calculation module S6.

Corresponding to the service allocation method shown in fig. 3 or fig. 5 provided in the embodiment of the present application, the present application further provides a service allocation apparatus 600. The service distribution apparatus 600 is implemented based on the service distribution method shown in fig. 3 or fig. 5. Specifically, please refer to fig. 6, where fig. 6 is a schematic structural diagram of a service allocating apparatus 600 provided in this embodiment. As shown in fig. 6, the service distribution apparatus 600 includes a data interface 121 and a service distribution control module 122.

The data interface 121 is configured to receive a service request sent by a client.

The service allocation control module 122 is configured to determine a service to be processed according to the service request received by the data interface 121, and allocate the service to be processed to the service processing device 200 for processing.

In an implementable embodiment of the service allocating apparatus 600 implemented based on the service allocating method shown in fig. 3, the service allocating control module 122 is configured to: determining a service to be processed, determining a service processing module group for processing the service to be processed, and distributing the service to be processed to a service processing module with the maximum power consumption and without reaching an upper limit in the service processing module group. Optionally, if the number of the service processing modules with the largest power consumption and not reaching the upper limit is greater than one, the service allocation control module 122 is further configured to identify a service processing module with the smallest number of working cycles among all the service processing modules with the largest power consumption and not reaching the upper limit, which have been identified from the service processing module group; and distributing the service to be processed to the service processing module with the minimum number of working cycles.

In particular, in an implementable embodiment of the service allocation device 600 implemented based on the service allocation method shown in fig. 3, the service allocation control module 122 comprises a processor 1221 and a memory 1222. The data interface 121, the processor 1221, and the memory 1222 are connected by a communication bus. The aforementioned process of determining the service to be processed, determining the set of service processing modules, and allocating the service to be processed to the service processing modules by the service allocation control module 122 can be implemented by the processor 1221 in the service allocation module 122.

In another implementable embodiment of the service allocating apparatus 600 implemented based on the service allocating method shown in fig. 5, the service allocating device 122 is further configured to: determining a service to be processed, determining a service processing module group for processing the service to be processed, determining candidate processing modules, and if the number of the candidate processing modules is greater than 1, the service allocation unit allocates the service to be processed to the candidate processing module with the power consumption not reaching the upper limit. The candidate processing module is the service processing module with the minimum working period number in the service processing module group. Alternatively, if the number of candidate processing modules is equal to 1, the service allocating apparatus 600 allocates the service request to the candidate processing modules.

In particular, in an implementable embodiment of the service allocation device 600 implemented based on the service allocation method shown in fig. 5, the service allocation control module 122 comprises a processor 1221 and a memory 1222. The data interface 121, the processor 1221, and the memory 1222 are connected by a communication bus. The service allocation control module 122 determines a service to be processed, determines a service processing module group for processing the service to be processed, and determines a candidate processing module, and if the number of the candidate processing modules is greater than 1, the process of allocating the service to be processed to the candidate processing module whose power consumption does not reach the upper limit by the service allocation unit may be implemented by the processor 1221 in the service allocation module 122.

In the service distribution apparatus 600 shown in fig. 6, the processor 1221 may be any computing device, and may be a general-purpose central processing unit 1221(CPU), a microprocessor 1221, a programmable controller, an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the above programs. The processor 1221 performs various functions of the service distribution apparatus 100 and processes data by running or executing software programs and/or application modules stored in the memory 1222 and calling data stored in the memory 1222, thereby performing overall monitoring of the service distribution apparatus 100. In particular implementations, processor 1221 may include one or more CPUs, as one embodiment.

A memory 1222 for temporarily storing the service request transmitted from the client. The memory 1222 also stores software programs and application modules. The processor 1221 executes various functional applications and data processing of the service distribution apparatus 100 by executing software programs and application modules stored in the memory 1222.

The memory 1222 mainly includes a program storage area and a data storage area. Wherein, the storage program area can store an operating system, an application program required by at least one function, such as sending a response message, etc.; the storage data area may store data to be processed by the application program. The Memory 1222 may include Volatile Memory (Volatile Memory), such as Random-Access Memory (RAM); the Memory 1222 may also include non-volatile Memory (non-volatile Memory), such as Read-Only Memory (ROM), Flash Memory (Flash Memory), a Hard Disk Drive (HDD), a Solid-State Drive (SSD), a Disk storage medium or other storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a network device, but is not limited to such.

The system bus 64 may be divided into an address bus, a data bus, a control bus, and the like. For clarity of illustration in the embodiments of the subject application, the various buses are illustrated in FIG. 6 as system bus 64.

Corresponding to the service allocation method shown in fig. 3, the service allocation apparatus provided in the embodiment of the present application is configured to perform the steps performed by the service allocation device 100 of the service allocation method shown in fig. 3. The service allocation apparatus provided in the embodiment of the present application may include modules corresponding to corresponding steps in the service allocation method shown in fig. 3. Each function module in the service distribution device can be realized according to each function in each step, or two or more functions can be integrated into one processing module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The division of the modules in the embodiment of the present application is schematic, and is only a logic function division, and there may be another division manner in actual implementation.

Corresponding to the service allocation method shown in fig. 5, another service allocation apparatus provided in the embodiment of the present application is configured to perform the steps performed by the service allocation device 100 of the service allocation method shown in fig. 5. The service allocation apparatus provided in the embodiment of the present application may include modules corresponding to corresponding steps in the service allocation method shown in fig. 5. Each function module in the service distribution device can be realized according to each function in each step, or two or more functions can be integrated into one processing module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The division of the modules in the embodiment of the present application is schematic, and is only a logic function division, and there may be another division manner in actual implementation.

Fig. 7 shows a schematic diagram of a possible structure of the service distribution apparatus 300 involved in the above embodiments. As shown in fig. 7, in the case of dividing each function module by corresponding each function, the service distribution apparatus 300 includes a service determination unit 301, a module group determination unit 302, and a service distribution unit 303.

In one implementation, the service determination unit 301 is used to support the service distribution apparatus 300 to perform S100 in the embodiment shown in fig. 3, and/or other processes for the technology described herein; the module group determination unit 302 is configured to support the service distribution apparatus 300 to perform S200 in the embodiment shown in fig. 3, and/or other processes for the technology described herein; the service distribution unit 303 is used to support the service distribution apparatus 300 to perform S300 in the embodiment shown in fig. 3, and/or other processes for the techniques described herein.

In a specific embodiment, as shown in fig. 8, the service determining unit 301 includes a service request obtaining subunit 3011, configured to obtain a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier; and the service analysis subunit 3012 is configured to obtain the service to be processed through analysis from the service request according to the service model identifier included in the service request.

When the service distribution apparatus 300 operates, the service distribution apparatus 300 performs the service distribution method of the embodiment shown in fig. 3. For the service allocation method, reference may be made to the related description of steps S100, S200, and S300 in the embodiment shown in fig. 3, which is not described herein again.

In another implementation, the service determination unit 301 is configured to support the service distribution apparatus 300 to perform S400 in the embodiment shown in fig. 5 and/or other processes for the technology described herein; the module group determining unit 302 is configured to support the service distribution apparatus 300 to perform S500 in the embodiment shown in fig. 5 and/or other processes for the technology described herein; the service distribution unit 303 is used to support the service distribution apparatus 300 to perform S600 and S700 in the embodiment shown in fig. 5, and/or other processes for the techniques described herein.

All relevant contents of each step related to the above method embodiment may be referred to the functional description of the corresponding functional module, and are not described herein again. Of course, the service distribution apparatus 300 provided in the embodiment of the present application includes, but is not limited to, the above modules, for example, the service distribution apparatus 300 may further include a receiving unit and a storage unit. The receiving unit is used for communicating with other devices, such as a storage device. The memory unit may be used for storing the program code of the service distribution apparatus 300.

In a specific implementation manner, as shown in fig. 8, the service determining unit 301 includes a service request obtaining subunit 3011, configured to obtain a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier; and the service analysis subunit 3012 is configured to obtain the service to be processed through analysis from the service request according to the service model identifier included in the service request.

When the service distribution apparatus 300 operates, the service distribution apparatus 300 performs the service distribution method of the embodiment shown in fig. 5. For the service allocation method, reference may be made to the related description of steps S100, S200, and S300 in the embodiment shown in fig. 5, which is not described herein again.

In the case of using integrated units, the service determination unit 301 and the module group determination unit 302 in this embodiment may be the processor 1221 in fig. 6, and the service distribution unit 303 may be the data interface 121 in fig. 6.

In other implementation manners, the service determination unit 301, the module group determination unit 302, and the service distribution unit 303 may be implemented in an integrated manner or in any combination, and the specific implementation manner is not limited by the embodiment of the present application.

The present embodiment further provides a server, which includes the foregoing service allocating device 100 and service processing device 200, or includes the foregoing service allocating device 600 and service processing device 200.

In an implementable manner, the service processing device 200 comprises at least one service processing module group, which comprises at least two service processing modules.

The present embodiments also provide a computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform the method as described above.

The present embodiments also provide a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method as described above.

The present application has been described in detail with reference to specific embodiments and illustrative examples, but the description is not intended to limit the application. Those skilled in the art will appreciate that various equivalent substitutions, modifications or improvements may be made to the presently disclosed embodiments and implementations thereof without departing from the spirit and scope of the present disclosure, and these fall within the scope of the present disclosure. The protection scope of this application is subject to the appended claims.

Claims

1. A service distribution method is characterized in that the service distribution method is used for a server, the server comprises service distribution equipment and service processing equipment, the service processing equipment comprises at least one service processing module group, and the service processing module group comprises at least two service processing modules; the method comprises the following steps:

the service distribution equipment determines a service to be processed;

the service distribution equipment determines a service processing module group for processing the service to be processed;

the service distribution equipment distributes the service to be processed to one service processing module in the service processing module group, wherein the service processing module group meets the following conditions;

the following conditions include: the power consumption is maximum, the power consumption does not reach the upper limit, the working period number is minimum, and the working duration parameter value is maximum;

the working period number represents the number of times of the accumulated working period of the service processing module, and the working duration parameter value represents the duration of the service processing module in the working state in the current working period.

2. The method of claim 1, wherein the determining of the pending traffic by the traffic distribution device comprises:

the service distribution equipment acquires a service request, wherein the service request comprises a service model identifier and a service to be processed corresponding to the service model identifier;

And the service distribution equipment analyzes and obtains the service to be processed from the service request according to the service model identifier included in the service request.

3. The method according to claim 1 or 2, characterized in that the method further comprises:

the service distribution equipment inquires the value of the working time length parameter of the service processing module;

if the value of the working duration parameter of the service processing module reaches the working duration preset by a working period, the service distribution equipment updates the working period number of the service processing module, the updated working period number is greater than the working period number before updating, and the updated working period number represents that the service processing module is in the next working period.

4. The method according to claim 1 or 2, wherein selecting one service processing module in the service processing module group which satisfies the following conditions successively comprises:

the service distribution equipment determines a candidate processing module in the service processing module group, wherein the candidate processing module meets the condition of minimum working cycle number;

determining a service processing module according to the following conditions from a plurality of candidate processing modules meeting the minimum condition of the working cycles, wherein the following conditions comprise: the power consumption is maximum, the power consumption does not reach the upper limit, and the working time parameter value is maximum.

5. An apparatus for traffic distribution, the apparatus comprising:

the service determining unit is used for determining the service to be processed;

a module group determining unit, configured to determine a service processing module group for processing the service to be processed;

the service distribution unit is used for distributing the service to be processed to one service processing module in the service processing module group, wherein the service processing module group meets the following conditions;

6. The apparatus of claim 5, wherein the traffic determination unit comprises:

a service request obtaining subunit, configured to obtain a service request, where the service request includes a service model identifier and a service to be processed corresponding to the service model identifier;

and the service analysis subunit is used for analyzing and obtaining the service to be processed from the service request according to the service model identifier included in the service request.

7. The apparatus according to claim 5 or 6, wherein the apparatus further comprises a duty cycle counting unit for querying a value of an operation duration parameter of the traffic processing module;

if the value of the working time length parameter of the service processing module reaches the working time length preset by a working period, the working period counting unit is further configured to update the working period number of the service processing module, where the updated working period number is greater than the working period number before updating, and the updated working period number indicates that the service processing module is in the next working period.

8. The apparatus of claim 5 or 6,

the service distribution unit is further configured to determine a candidate processing module in the service processing module group, where the candidate processing module meets a condition that a working cycle number is minimum; and determining a service processing module among the candidate processing modules which satisfy the condition of minimum number of working cycles according to the following conditions: the power consumption is maximum, the power consumption does not reach the upper limit, and the working time parameter value is maximum.

9. A service distribution device is characterized by comprising a data interface and a service distribution control module,

The data interface is used for receiving a service request sent by a client,

the service distribution control module is used for determining a service to be processed, determining a service processing module group for processing the service to be processed, and distributing the service to be processed to one service processing module in the service processing module group, wherein the service processing module group meets the following conditions;

the following conditions include: maximum power consumption, no upper limit of power consumption, minimum number of working cycles, and maximum working duration parameter value.

10. A server, characterized in that the server comprises a service distribution apparatus and a service processing apparatus according to claim 9, the service processing apparatus comprising at least one service processing module group, the service processing module group comprising at least one service processing module.

11. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the method of any of claims 1 to 4.