WO2015027935A1

WO2015027935A1 - Method and device for allocating computational resources

Info

Publication number: WO2015027935A1
Application number: PCT/CN2014/085392
Authority: WO
Inventors: Yansheng Jiang
Original assignee: Tencent Technology (Shenzhen) Company Limited
Priority date: 2013-08-28
Filing date: 2014-08-28
Publication date: 2015-03-05
Also published as: CN104426953A

Abstract

A method and a device for allocating computational resources are provided. At first, a virtual machine is selected according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit. Computational units are selected for the cloud service from unused computational units of the selected virtual machine. Then, the selected computational units are allocated to the cloud service. An average utilization rate of the allocated computational units is calculated according to respective utilization rates of the allocated computational units. At last, the allocated computational units are adjusted according to the average utilization rate of the allocated computational units. The device includes a selection module, an allocation module, a calculation module and an adjusting module. By the method and the device, the computational resources utilization efficiency of the cloud server is improved.

Description

METHOD AND DEVICE FOR ALLOCATING COMPUTATIONAL RESOURCES

CROSS-REFERENCE

This application claims priority to CN Patent Application No. CN 201310382078. X, filed on August 28, 2013, which are hereby incorporated herein by reference in its entirety.

FIELD OF THE TECHNOLOGY

The present disclosure relates to cloud computing field, and more particularly to a method and a device for allocating computational resources.

BACKGROUND OF THE DISCLOSURE

With rapid development of cloud service technology, many cloud platforms are built for various cloud services. Counts of page visits or page views of the cloud services fluctuate wildly depending on appearance or disappearance of hot spots. Appearance of the hot spots sharply increase the page visit counts or page view counts, and more computational resources are required for the cloud service； conversely, disappearance of the hot spots causes a big drop in the page visit/view counts and only less computational resources still be required for the cloud service. Therefore, cloud servers allocate the computational resources to different services according to their page visit/view counts.

A known method for allocating the computational resources is described therein. At first, a cloud server allocates at least one virtual machine to a cloud service according to a page visit count in response to a service request. During the process of the cloud service, the cloud server receives the page visit count in real time and compares the page visit count with a threshold number. If the real-time page visit count is greater than or equal to the threshold number, quantity of the allocated virtual machines for the cloud service is increased. Otherwise, if the real-time page visit count is less than the threshold number, the quantity of the allocated virtual machines for the cloud service is decreased.

One disadvantage of the above-described method is that required computational resources for the cloud service may be less than the real computational resources included in the allocated virtual machines for the cloud service. It wastes many unused computational resources included in the allocated virtual machines. Furthermore, low utilization efficiency of the allocated virtual machines is also unsatisfied.

SUMMARY

The present disclosure provides a method and a device for allocating computational resources to improve computational resources utilization efficiency.

An aspect of the present disclosure provides a method for allocating computational resources. At first, at least one virtual machine is selected according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit. Furthermore, a plurality of computational units are selected for the cloud service from unused computational units of the at least one selected virtual machine. The total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service. Then, the selected computational units are allocated to the cloud service. An average utilization rate of the allocated computational units is calculated according to respective utilization rates of the allocated computational units. At last, the allocated computational units for the cloud service are adjusted according to the average utilization rate of the allocated computational units.

Another aspect of the present disclosure provides a method for allocating computational resources. The device includes a selection module, an allocation module, a calculation module and an adjusting module. The selection module is configured to select at least one virtual machine according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit, and select a plurality of computational units for the cloud service from unused computational units of the at least one selected virtual machine. The total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service. The allocation module is configured to allocate the selected computational units to the cloud service. The calculation module is configured to calculate an average utilization rate of the allocated computational units according to respective utilization rates of the allocated computational units. The adjusting module is configured to adjust the allocated computational units for the cloud service according to the average utilization rate of the allocated computational units.

BRIEF DESCRIPTION OF THE DRAWINGS

The present disclosure will become more readily apparent to those ordinarily skilled in the art after reviewing the following detailed description and accompanying drawings, in which:

FIG. 1 is a flow chart illustrating an embodiment of a method for allocating computational resources according to the present application；

FIGs. 2A and 2B are flow charts illustrating another embodiment of a method for allocating computational resources according to the present application；

FIG. 3 is a schematic diagram illustrating an embodiment of a device for allocating computational resources according to the present application；

FIG. 4 is a schematic diagram illustrating a selection module of the device for allocating computational resources of FIG. 3；

FIG. 5 is a schematic diagram illustrating an adjusting module of the device for allocating computational resources of FIG. 3；

FIG. 6 is a schematic diagram illustrating an adjusting unit of the adjusting module of FIG. 5；

FIG. 7 is a schematic diagram illustrating another embodiment of a device for allocating computational resources according to the present application；

FIG. 8 is a schematic diagram illustrating an acquiring module of the device for allocating computational resources of FIG. 7；

FIG. 9 is a schematic diagram illustrating another acquiring module of the device for allocating computational resources of FIG. 7； and

FIG. 10 is a schematic diagram illustrating a further embodiment of a device for allocating computational resources according to the present application.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

The present disclosure will now be described more specifically with reference to the following embodiments. It is to be noted that the following descriptions of embodiments of this invention are presented herein for purpose of illustration and description only. It is not intended to be exhaustive or to be limited to the precise form disclosed.

Please refer to FIG. 1, a schematic diagram illustrating an embodiment of a method for allocating computational resources according to the present application. The method includes the following steps.

In step 101, at least one virtual machine is selected according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit. Then, a plurality of computational units are selected for the cloud service from unused computational units of the at least one selected virtual machine. The total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service.

In step 102, the selected computational units are allocated to the cloud service. In an optimum situation, all the allocated computational units are used during process of the cloud service. In other words, there is no idle computational unit during the process of the cloud service.

In step 103, an average utilization rate of the allocated computational units is calculated according to respective utilization rates of the allocated computational units.

In step 104, the allocated computational units for the cloud service are adjusted according to the average utilization rate of the allocated computational units.

In an embodiment, the step 101 further includes the following steps: calculating a required number of the computational units for the cloud service according to the total number of the computational resources required for the cloud service and the respective number of the computational resources per computational unit； selecting the at least one virtual machine according to the required number of the computational units for the cloud service； and selecting the computational units from the unused computational units of the selected virtual machine according to the required number of the computational units for the cloud service.

In an embodiment, the step 104 further includes the following steps: determining a utilization rate range covering the average utilization rate of the allocated computational units, and determining an adjusting parameter N corresponding to the determined utilization rate range according to a predetermined relation defining a plurality of utilization rate ranges and a plurality of adjusting parameters； and adjusting the allocated computational units for the cloud service according to the adjusting parameter N. For example, this step further includes sub-steps of: increasing the allocated computational units for the cloud service by N computational units if a lower boundary of the utilization rate range is greater than a first threshold； and decreasing the allocated computational units for the cloud service by N computational units if an upper boundary of the utilization rate range is smaller than a second threshold.

Furthermore, the step 103 further includes a step of acquiring the respective utilization rates of the allocated computational units. Optionally, the respective utilization rate-acquiring step includes steps of: acquiring a plurality of instant utilization rates of each of the allocated computational units from the each allocated computational unit； and calculating the respective utilization rates of the allocated computational units according to the instant utilization rates. Alternatively, the respective utilization rate-acquiring step includes a step of acquiring the respective utilization rates of the allocated computational units from the allocated computational units wherein the respective utilization rates of the allocated computational units are calculated according to a plurality of instant utilization rates of each of the allocated computational units.

Furthermore, the method for allocating computational resources includes a step of receiving a new utilization rate range and a new adjusting parameter, and updating the predetermined relation with the new utilization rate range and the new adjusting parameter.

In the embodiments of the present application, a virtual machine is not viewed as a basic unit to perform the cloud service. Therefore, the cloud server does not allocate an entire virtual machine to the cloud service. Instead, the cloud server selects at least one virtual machine and then allocates required computational units from unused computational units of the at least one virtual machine to the cloud service according to the required computational resources and the respective number of the computational resources included in a single computational unit. Thus, other computational units of the selected virtual machine (s) may be allocated to other cloud service to avoid waste. Hence, the utilization rates of the computational units allocated to the cloud service are improved. Furthermore, the computational resources are dynamically adjusted by adjusting the allocated computational units according to the average utilization rates. The granularity of adjusting the computational resources is finer than that of the conventional method so as to avoid waste.

Please refer to FIGs. 2A and 2B, flow charts illustrating another embodiment of a method for allocating computational resources according to the present application. The method includes the following steps.

In step 201, a required number of computational units for a cloud service is calculated according to a total number of computational resources required for the cloud service and a respective number of the computational resources per computational unit. The total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service.

Concretely, the total number of the computational resources divided by the respective number of the computational resources per computational unit makes a value. If the value is a positive integer, the required number of the computational units for the cloud service equals the value. If the value is not a positive integer, the required number of the computational units for the cloud service equals a nearest positive integer greater than the value.

For example, if 3,500 computational resources are required for the cloud service and a single computational unit includes 1,000 computational resources, the total number of the computational resources (3,500) divided by the respective number of the computational resources (1,000) per computational unit makes a value of 3. 5. Since the value is not a positive integer, a nearest positive integer greater than the value is 4. Therefore, four computational units are required for the cloud service.

In step 202, at least one virtual machine is selected, and a plurality of computational units are selected from unused computational units of the selected at least one virtual machine and allocated to the cloud service according to the required number of the computational units for the cloud service.

Concretely, the cloud server selects the virtual machine which includes the unused computational units more than or equal to the required computational units. Then, the cloud server randomly selects the computational units from the unused computational units according to the required number of the computational units for the cloud service and allocates the selected computational units to the cloud service.

For example, every virtual machine includes 10 computational units and the cloud server randomly selects a virtual machine according to the required number of the computational units (4) . Then, the cloud server randomly selects four unused computational units and allocates the four selected computational units to the cloud service.

In step 203, respective utilization rates of the allocated computational units are acquired.

Concretely, two schemes may be used. The first scheme includes steps of: acquiring a plurality of instant utilization rates of each of the allocated computational units from the each allocated computational unit； and calculating the respective utilization rates of the allocated computational units according to the instant utilization rates. Alternatively, the second scheme includes steps of: each of the allocated computational units acquiring a plurality of instant utilization rates, calculating the respective utilization rates of the allocated computational units according to the instant utilization rates, and reporting the respective utilization rates to the cloud server； and the cloud server receiving the respective utilization rates of the allocated computational units.

According to the first scheme, the respective utilization rates of the allocated computational units are calculated according to the instant utilization rates by equation (1) :

wherein S_m is the respective utilization rate of the mth allocated computational unit, k is a number of the instant utilization rates, and s_i is the ith instant utilization rate of the mth allocated computational unit.

For example, five instant utilization rates are acquired. The first allocated computational unit reports 80％, 85％, 90％, 78％and 82％as the five instant utilization rates. According to equation (1) , the respective utilization rate of the first allocated computational unit is 83％. The second allocated computational unit reports 90％, 80％, 60％, 75％and 85％as the five instant utilization rates. According to equation (1) , the respective utilization rate of the second allocated computational unit is 78％. The third allocated computational unit reports 80％, 95％, 90％, 95％and 90％as the five instant utilization rates. According to equation (1) , the respective utilization rate of the third allocated computational unit is 90％. The fourth allocated computational unit reports 90％, 80％, 70％, 85％and 95％as the five instant utilization rates. According to equation (1) , the respective utilization rate of the fourth allocated computational unit is 84％.

In step 204, an average utilization rate is calculated according to the acquired respective utilization rates of the allocated computational units.

The average utilization rate is calculated according to the acquired respective utilization rates of the allocated computational units by equation (2) :

wherein

is the average utilization rate and h is the number of the allocated computational units.

For example, the respective utilization rates of the four allocated computational units are 83％, 78％, 90％and 84％, respectively. According to equation (2) , the average utilization rate is 83.75％.

In step 205, a utilization rate range covering the average utilization rate is determined, and an adjusting parameter N corresponding to the determined utilization rate range is determined according to a predetermined relation defining a plurality of utilization rate ranges and a plurality of adjusting parameters.

Concretely, the predetermined relation defining the utilization rate ranges and the adjusting parameters are acquired. Then, the average utilization rate is compared with the utilization rate ranges to find one of the utilization rate ranges covering the average utilization rate. The adjusting parameter N is determined according to the determined utilization rate range and the predetermined relation defining the utilization rate ranges and the adjusting parameters.

For example, the predetermined relation defining the utilization rate ranges and the adjusting parameters is shown in Table 1. The utilization rate ranges include 90％-100％, 80％-90％and 0％-10％. The average utilization rate 83.75％is compared with the three utilization rate ranges 90％-100％, 80％-90％and 0％-10％, and the determined utilization rate range is 80-90％. From Table 1, the adjusting parameter N corresponding to the determined utilization rate range 80％-90％is 4.

Table 1

Utilization rate range	Adjusting parameter
Utilization rate range	Adjusting parameter	90％-100％	5
80％-90％	4	90％-100％	5
80％-90％	4	0％-10％	5

The predetermined relation defining the utilization rate ranges and the adjusting parameters is stored in the cloud server in advance. If the manager or user needs to change the predetermined relation, it can be performed by inputting a new utilization rate range and a new adjusting parameter through the cloud server or a client terminal.

Furthermore, the new utilization rate range and the new adjusting parameter are received and the predetermined relation is updated according to the new utilization rate range and the new adjusting parameter.

The updating step further includes steps of: searching the new utilization rate range in the determined relation； updating the predetermined relation with the new adjusting parameter if the new utilization rate range matches one of the utilization rate ranges defined in the predetermined relation； and inserting the new utilization rate range and the new adjusting parameter into the predetermined relation if the new utilization rate range does not match any of the utilization rate ranges defined in the predetermined relation.

For example, the cloud server receives a first new utilization rate range 70％-80％, a first new adjusting parameter 3, a second new utilization rate range 10％-20％and a second new adjusting parameter 3. Since each of the first new utilization rate range 70％-80％and the second new utilization rate range 10％-20％does not match any of the utilization rate ranges defined in the predetermined relation, the first new utilization rate range 70％-80％, the first new adjusting parameter 3, the second new utilization rate range 10％-20％and the second new adjusting parameter 3 are inserted into the predetermined relation. The updated predetermined relation is shown in Table 2.

Table 2

Utilization rate range	Adjusting parameter
Utilization rate range	Adjusting parameter	90％-100％	5
80％-90％	4	90％-100％	5
80％-90％	4	0％-10％	5
70％-80％	3	0％-10％	5
70％-80％	3	10％-20％	3

In step 206, the allocated computational units for the cloud service increases by N computational units if a lower boundary of the utilization rate range is greater than a first threshold.

Concretely, the lower boundary of the utilization rate range is compared with the first threshold. A virtual machine (may be the original virtual machine or a new virtual machine) is selected according to the adjusting parameter N if the lower boundary of the utilization rate range is greater than the first threshold. N computational units are randomly selected from unused computational units in the selected virtual machine, and the selected N computational units are additionally allocated to the cloud service. It is to be noted that a number of the unused computational units in the selected virtual machine is more than or equal to the adjusting parameter N.

For example, the first threshold is set 70％. Since the lower boundary 80％of the utilization rate range 80％-90％is greater than the first threshold 70％, a specific virtual machine is selected according to the adjusting parameter 4 corresponding to the utilization rate range 80％-90％. Further, four computational units are randomly selected from unused computational units in the selected virtual machine, and the selected four computational units are additionally allocated to the cloud service. Therefore, total eight computational units are allocated to the cloud service now.

In step 207, the allocated computational units for the cloud service decreases by N computational units if an upper boundary of the utilization rate range is smaller than a second threshold.

Concretely, the upper boundary of the utilization rate range is compared with the second threshold. N computational units are removed from the allocated computational units if the upper boundary of the utilization rate range is smaller than the second threshold.

Optionally, in the predetermined relation, the adjusting parameters may include positive integers and negative integers to indicate increase or decrease the allocated computational units, respectively. If the average utilization rate range matches one of the utilization rate ranges defined in the predetermined relation, the allocated computational units increases or decrease according to the corresponding adjusting parameter. For example, (+) 5 indicates adding five allocated computational units and -3 indicates removing three allocated computational units. Thus, the comparing steps of comparing the lower boundary/the upper boundary of the utilization rate range with the first threshold/second threshold are omitted so as to reduce the allocation time and improve the computational resource utilization efficiency.

In the embodiments of the present application, a virtual machine is not viewed as a basic unit to perform the cloud service.Therefore, the cloud server does not allocate an entire virtual machine to the cloud service. Instead, the cloud server selects at least one virtual machine and then allocates required computational units from unused computational units of the at least one virtual machine to the cloud service according to the required computational resources and the respective number of the computational resources included in a single computational unit. Thus, other computational units of the selected virtual machine (s) may be allocated to other cloud service to avoid waste. Hence, the utilization rates of the computational units allocated to the cloud service are improved. Furthermore, the computational resources are dynamically adjusted by adjusting the allocated computational units according to the average utilization rates. The granularity of adjusting the computational resources is finer than that of the conventional method so as to avoid waste.

Please refer to FIG. 3, a schematic diagram illustrating an embodiment of a device for allocating computational resources according to the present application. The device 3 includes a selection module 301, an allocation module 302, a calculation module 303 and an adjusting module 304. The selection module 301 is configured to select at least one virtual machine according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit, and select a plurality of computational units for the cloud service from unused computational units of the at least one selected virtual machine. The total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service. The allocation module 302 is configured to allocate the selected computational units to the cloud service. The calculation module 303 is configured to calculate an average utilization rate of the allocated computational units according to respective utilization rates of the allocated computational units. The adjusting module 304 is configured to adjust the allocated computational units for the cloud service according to the average utilization rate of the allocated computational units.

In an optimum situation, all the allocated computational units are used during process of the cloud service. In other words, there is no idle computational unit during the process of the cloud service.

In an embodiment, the selection module 301 includes a first calculation unit 311, a first selection unit 312 and a second selection unit 313 as shown in FIG. 4. The first calculation unit 311 is configured to calculate a required number of the computational units for the cloud service according to the total number of the computational resources required for the cloud service and the respective number of the computational resources per computational unit. The first selection unit 312 is configured to select the at least one virtual machine according to the required number of the computational units for the cloud service. The second selection unit 313 is configured to select the computational units from the unused computational units of the selected virtual machine according to the required number of the computational units for the cloud service.

In an embodiment, the adjusting module 304 includes a determination unit 341 and an adjusting unit 342, as shown in FIG. 5. The determination unit 341 is configured to determine a utilization rate range covering the average utilization rate of the allocated computational units, and determine an adjusting parameter N corresponding to the determined utilization rate range according to a predetermined relation defining a plurality of utilization rate ranges and a plurality of adjusting parameters. The adjusting unit 342 is configured to adjust the allocated computational units for the cloud service according to the adjusting parameter N.

In an embodiment, the adjusting unit 342 includes an increasing sub-unit 3421 and a decreasing sub-unit 3422, as shown in FIG. 6. The increasing sub-unit 3421 is configured to increase the allocated computational units for the cloud service by N computational units if a lower boundary of the utilization rate range is greater than a first threshold. The decreasing sub-unit 3422 is configured to decrease the allocated computational units for the cloud service by N computational units if an upper boundary of the utilization rate range is smaller than a second threshold.

Please refer to FIG. 7, a schematic diagram illustrating another embodiment of a device for allocating computational resources according to the present application. In addition to the modules 301-304 described with reference to FIG. 3, the device for allocating computational resources 3 further includes an acquiring module 305 configured to acquire the respective utilization rates of the allocated computational units.

The acquiring module 305 includes a first acquiring unit 351 and a second calculation unit 352, as shown in FIG. 8. The first acquiring unit 351 is configured to acquire a plurality of instant utilization rates of each of the allocated computational units from the each allocated computational unit. The second calculation unit 352 is configured to calculate the respective utilization rates of the allocated computational units according to the instant utilization rates.

Alternatively, the acquiring module 305 includes a second acquiring unit 353, as shown in FIG. 9, configured to acquire the respective utilization rates of the allocated computational units from the allocated computational units. The respective utilization rates of the allocated computational units are calculated according to a plurality of instant utilization rates of each of the allocated computational units.

In another embodiment, the device for allocating computational resources 3 further includes an updating module 306 (FIG. 10) . The updating module 306 is configured to receive a new utilization rate range and a new adjusting parameter, and update the predetermined relation with the new utilization rate range and the new adjusting parameter.

It should be noted that the device for allocating computational resources of the present application may be used with or perform the methods for allocating computational resources as described in the above embodiments, and the detailed operation or examples are not repeated here.

It is to be noted that the cloud server in the embodiments of the invention may be a computer or other electronic device with communication function. As used in this specification and any claims of this application, the terms “device” and “server” refer to electronic devices. These terms exclude people or groups of people.

For those having ordinary skill in the art, it is understood that all or part of the steps in the various embodiments described above can be executed by hardware or programs, and the corresponding program may be stored in a computer-readable medium. The computer-readable medium carries one or more sequences of computer-executable instructions for causing one or more processing units to perform the methods in the embodiments. The so-called computer-readable medium can be, for example, a ROM/RAM, disk or optical disk, etc.

While the invention has been described in terms of what is presently considered to be the most practical and preferred embodiments, it is to be understood that the invention needs not be limited to the disclosed embodiment. On the contrary, it is intended to cover various modifications and similar arrangements included within the spirit and scope of the appended claims which are to be accorded with the broadest interpretation so as to encompass all such modifications and similar structures.

Claims

A method for allocating computational resources comprising steps of:

selecting at least one virtual machine according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit, and selecting a plurality of computational units for the cloud service from unused computational units of the at least one selected virtual machine, wherein the total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service；

allocating the selected computational units to the cloud service；

calculating an average utilization rate of the allocated computational units according to respective utilization rates of the allocated computational units； and

adjusting the allocated computational units for the cloud service according to the average utilization rate of the allocated computational units.
The method according to claim 1, wherein the selecting step comprises steps of:

calculating a required number of the computational units for the cloud service according to the total number of the computational resources required for the cloud service and the respective number of the computational resources per computational unit；

selecting the at least one virtual machine according to the required number of the computational units for the cloud service； and

selecting the computational units from the unused computational units of the selected virtual machine according to the required number of the computational units for the cloud service.
The method according to claim 1 or 2, wherein the calculating step comprises steps of:

determining a utilization rate range covering the average utilization rate of the allocated computational units, and determining an adjusting parameter N corresponding to the determined utilization rate range according to a predetermined relation defining a plurality of utilization rate ranges and a plurality of adjusting parameters； and

adjusting the allocated computational units for the cloud service according to the adjusting parameter N.
The method according to claim 3, wherein the step of adjusting the allocated computational units for the cloud service according to the adjusting parameter N comprises sub-steps of:

increasing the allocated computational units for the cloud service by N computational units if a lower boundary of the utilization rate range is greater than a first threshold； and

decreasing the allocated computational units for the cloud service by N computational units if an upper boundary of the utilization rate range is smaller than a second threshold.
The method according to claim 3, the method further comprising a step of receiving a new utilization rate range and a new adjusting parameter, and updating the predetermined relation with the new utilization rate range and the new adjusting parameter.
The method according to claim 1, wherein the method, before the calculating step, further comprises a step of acquiring the respective utilization rates of the allocated computational units.
The method according to any one of claim 6, wherein the acquiring step comprises steps of:

acquiring a plurality of instant utilization rates of each of the allocated computational units from the each allocated computational unit； and

calculating the respective utilization rates of the allocated computational units according to the instant utilization rates.
The method according to claim 6, wherein the acquiring step comprises a step of acquiring the respective utilization rates of the allocated computational units from the allocated computational units wherein the respective utilization rates of the allocated computational units are calculated according to a plurality of instant utilization rates of each of the allocated computational units.
A device for allocating computational resources comprising:

a selection module configured to select at least one virtual machine according to a total number of computational resources required for a cloud service and a respective number of the computational resources per computational unit, and select a plurality of computational units for the cloud service from unused computational units of the at least one selected virtual machine, wherein the total number of the computational resources required for the cloud service is determined by a page visit count of the cloud service and processing resources required for a cloud server to process the cloud service；

an allocation module configured to allocate the selected computational units to the cloud service；

a calculation module configured to calculate an average utilization rate of the allocated computational units according to respective utilization rates of the allocated computational units； and

an adjusting module configured to adjust the allocated computational units for the cloud service according to the average utilization rate of the allocated computational units.
The device according to claim 9, wherein the selection module comprises:

a first calculation unit configured to calculate a required number of the computational units for the cloud service according to the total number of the computational resources required for the cloud service and the respective number of the computational resources per computational unit；

a first selection unit configured to select the at least one virtual machine according to the required number of the computational units for the cloud service； and

a second selection unit configured to select the computational units from the unused computational units of the selected virtual machine according to the required number of the computational units for the cloud service.
The device according to claim 9 or 10, wherein the adjusting module comprises:

a determination unit configured to determine a utilization rate range covering the average utilization rate of the allocated computational units, and determine an adjusting parameter N corresponding to the determined utilization rate range according to a predetermined relation defining a plurality of utilization rate ranges and a plurality of adjusting parameters； and

an adjusting unit configured to adjust the allocated computational units for the cloud service according to the adjusting parameter N.
The device according to claims 11, wherein the adjusting unit comprises:

an increasing sub-unit configured to increase the allocated computational units for the cloud service by N computational units if a lower boundary of the utilization rate range is greater than a first threshold； and

a decreasing sub-unit configured to decrease the allocated computational units for the cloud service by N computational units if an upper boundary of the utilization rate range is smaller than a second threshold.
The device according to claim 11, wherein the device further comprises an updating module configured to receive a new utilization rate range and a new adjusting parameter, and update the predetermined relation with the new utilization rate range and the new adjusting parameter.
The device according to claim 9, wherein the device further comprises an acquiring module configured to acquire the respective utilization rates of the allocated computational units.
The device according to claim 14, wherein the acquiring module comprises:

a first acquiring unit configured to acquire a plurality of instant utilization rates of each of the allocated computational units from the each allocated computational unit； and

a second calculation unit configured to calculate the respective utilization rates of the allocated computational units according to the instant utilization rates.
The device according to claim 14, wherein the acquiring module comprises a second acquiring unit configured to acquire the respective utilization rates of the allocated computational units from the allocated computational units, wherein the respective utilization rates of the allocated computational units are calculated according to a plurality of instant utilization rates of each of the allocated computational units.
A computer-readable medium carrying one or more sequences of computer-executable instructions for causing one or more processing units to perform the method according to any one of claims 1-8.