CN110609747A - Information processing method and electronic equipment - Google Patents
Information processing method and electronic equipment Download PDFInfo
- Publication number
- CN110609747A CN110609747A CN201910810668.5A CN201910810668A CN110609747A CN 110609747 A CN110609747 A CN 110609747A CN 201910810668 A CN201910810668 A CN 201910810668A CN 110609747 A CN110609747 A CN 110609747A
- Authority
- CN
- China
- Prior art keywords
- resource
- load
- condition information
- load condition
- period
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 15
- 238000003672 processing method Methods 0.000 title claims abstract description 15
- 238000012544 monitoring process Methods 0.000 claims description 109
- 238000000034 method Methods 0.000 claims description 27
- 230000008859 change Effects 0.000 claims description 21
- 230000001174 ascending effect Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 description 25
- 238000004458 analytical method Methods 0.000 description 16
- 238000003860 storage Methods 0.000 description 8
- 238000007726 management method Methods 0.000 description 7
- 230000009467 reduction Effects 0.000 description 7
- 230000005012 migration Effects 0.000 description 6
- 238000013508 migration Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000012806 monitoring device Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000003491 array Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005265 energy consumption Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 238000013468 resource allocation Methods 0.000 description 3
- 238000012952 Resampling Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 108010028984 3-isopropylmalate dehydratase Proteins 0.000 description 1
- 241001122767 Theaceae Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application discloses an information processing method and electronic equipment. The information processing method comprises the following steps: predicting load condition information of a prediction object in a resource platform according to a service scene of the resource platform; and according to the load condition information, executing a predetermined operation matched with the load condition information on the resource platform in a predicted time period.
Description
Technical Field
The present application relates to the field of information technologies, and in particular, to an information processing method and an electronic device.
Background
In order to realize optimization of resource configuration, a plurality of individual resources are integrated together to realize resource platformization, thereby forming a resource platform. In order to achieve good operation in the resource platform, operations such as monitoring the resource platform and/or resource optimal configuration are required.
However, in the related art, resource optimization configuration is performed through real-time monitoring, but this method may have a plurality of problems, and on one hand, if the real-time monitoring is abnormal, the resource optimization configuration cannot be executed; on the other hand, the real-time monitoring configuration is realized, and the monitoring consumption is large; yet another aspect is that dynamic adjustment of resources cannot be achieved.
Disclosure of Invention
In view of the above, embodiments of the present application are intended to provide an information processing method and an electronic device.
The technical scheme of the application is realized as follows:
an embodiment of the present application provides an information processing method, including:
predicting load condition information of a prediction object in a resource platform according to a service scene of the resource platform;
and according to the load condition information, executing a predetermined operation matched with the load condition information on the resource platform in a predicted time period.
Based on the above solution, the performing, according to the load condition information, the predetermined operation adapted to the load condition information on the resource platform in the predicted time period includes at least one of:
according to the load condition information, resource scheduling in a time period for predicting the resource platform is carried out;
and adjusting the monitoring configuration of the resource platform in a predicted time period according to the load condition information.
Based on the above scheme, the scheduling the resource in the time period in which the resource platform is predicted according to the load condition information includes:
and determining the load change time of the prediction object according to the load condition information, and adjusting the running state of the resource object in the resource platform before the load change time in a prediction time period, wherein the prediction object is configured on the resource object.
Based on the above scheme, the adjusting the running state of the resource object in the resource platform includes at least one of:
adjusting the over-selling ratio of the resource object in the resource platform according to the descending or ascending condition of the load;
when the load capacity is reduced, the prediction object is transferred from the resource object with low over-sale ratio to the resource object with high over-sale ratio;
when the load capacity is reduced, closing the idle resource object;
when the load capacity is increased, transferring the predicted object from the resource object with high over-selling ratio to the resource object with low over-selling ratio, and reducing the over-selling ratio of the resource object from which the predicted object is transferred;
when the load capacity rises, the closed resource objects are started.
Based on the above scheme, the method further comprises:
when the load condition information indicates that: when at least one specific sub-period in the predicted period is within the period, scheduling reserved resources in the resource platform to share the load capacity, wherein the specific sub-period is as follows: a sub-period is predicted which contains the moment when the load rate is above the first threshold.
Based on the above scheme, the scheduling of the reserved resources in the resource platform to share the load capacity includes at least one of:
in the specific sub-period, reserving resource sharing load in resource objects with the enabled over-selling ratio exceeding the over-selling ratio threshold;
and enabling the reserved resource object to share the load in the specific sub-period.
Based on the above scheme, the adjusting the monitoring configuration of the resource platform according to the load condition information includes:
and adjusting the monitoring frequency of the resource platform according to the load condition information.
A second aspect of the embodiments of the present application provides an electronic device, including:
the prediction module is used for predicting the load condition information of a prediction object in the resource platform according to the service scene of the resource platform;
and the execution module is used for executing the preset operation matched with the load condition information to the resource platform in a predicted time period according to the load condition information.
Based on the foregoing solution, the prediction module is specifically configured to perform at least one of the following:
according to the load condition information, resource scheduling in a time period for predicting the resource platform is carried out;
and adjusting the monitoring configuration of the resource platform in a predicted time period according to the load condition information.
Based on the above scheme, the execution module is specifically configured to determine load change time of the predicted object according to the load condition information, and adjust an operating state of the resource object in the resource platform before the load change time in a predicted time period; and/or adjusting the monitoring frequency of the resource platform according to the load condition information.
The technical solution provided in the embodiment of the present application is to predict load condition information according to a service scenario, so that load condition information in a predicted time period is known before the predicted time period comes, and execution of a predetermined operation with load configuration equal to that of the load condition information can be performed in advance according to the load condition information, on one hand, compared with real-time monitoring, a phenomenon that the predetermined operation cannot be performed due to failure of obtaining a monitoring result when the real-time monitoring is abnormal is reduced, and on the other hand, compared with real-time monitoring, a prediction manner can reduce resource consumption, for example, at least reduce consumption of electric energy; on the other hand, compared with real-time monitoring, the predetermined operation required to be executed in the predicted time period can be predicted in advance, so that the determination of the predetermined operation is realized in advance, and the efficiency of the predetermined operation can be improved; meanwhile, the prediction method has the characteristic of low calculation complexity, saves calculation power and saves running power consumption generated by calculation.
Drawings
Fig. 1 is a schematic flowchart of a first information processing method according to an embodiment of the present application;
fig. 2 is a flowchart illustrating a second information processing method according to an embodiment of the present application;
fig. 3 is a flowchart illustrating a third information processing method according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The technical solution of the present application is further described in detail with reference to the drawings and specific embodiments of the specification.
As shown in fig. 1, the present embodiment provides an information processing method including:
step S110: predicting load condition information of a prediction object in a resource platform according to a service scene of the resource platform;
step S120: and according to the load condition information, executing a predetermined operation matched with the load condition information on the resource platform in a predicted time period.
The information processing method provided by the embodiment can be applied to resource scheduling equipment or monitoring equipment of a resource platform. The resource platform includes, but is not limited to, various types of cloud platforms. The cloud platform includes, but is not limited to, a private cloud platform.
In this embodiment, load condition information of a prediction object in the resource platform is predicted in advance according to a service scenario of the resource platform, and thus, by predicting in advance, load condition information at a certain moment or a plurality of moments in the future is known, so that a predetermined operation adapted to the load condition information can be performed, and the predetermined operation can be used for load balancing or a monitoring operation adapted to current load condition information.
Different service scenes and different service tasks executed by the resource platform.
Different load fluctuation rules or load rate fluctuation rules can be presented in different service scenes.
For example, the service scenario is: an office scene. In an office scenario, the load is office load generated during the office process. The office load is high in load in working days, and low in load in non-working days; the load amount in the working period of the working day is high, and the load amount in the non-working period is low. For example, the amount of load in the noon break period is lower than the amount of load in the work period in the morning or afternoon.
The non-workdays include, but are not limited to: weekends and legal holidays.
The non-operating period includes: the meal times and intermediate rest periods, e.g., the lunch period and the afternoon tea period.
For another example, the service scenario is: a multimedia scene. In a multimedia scenario, the load is a multimedia load generated during multimedia generation and/or playing. The multimedia load is high during the leisure entertainment period and low during the non-entertainment period.
Therefore, under different service scenes, the load condition information can be determined according to various information representing the service scenes.
For example, the business scenarios are characterized by working calendars under different business scenarios.
The load condition information includes at least one of:
the load capacity;
the load factor.
In some embodiments, the predicted objects include, but are not limited to: a virtual machine configured within the resource platform.
In other embodiments, the predictive objects include, but are not limited to: a host of the resource platform. The host machine can be configured with a virtual machine.
In still other embodiments, the predicted object may further include: a resource pool within a resource platform. The resource platform may include a resource pool that may include: a computing resource pool and a storage resource pool. The computing resource pool includes a plurality of computing resources, such as a Central Processing Unit (CPU) or a Graphic Processing Unit (GPU), and the like. The storage resource pool includes, but is not limited to: memory and/or hard disk.
In the present embodiment, in order to reduce load imbalance between individual prediction objects or to generate unnecessary monitoring, a predetermined operation adapted to load condition information is performed within a predicted period; therefore, by means of prediction in advance, on one hand, the phenomenon that reference basis of preset operations such as resource allocation and the like is not carried out on a monitoring result under the condition of real-time prediction and real-time monitoring abnormity is reduced. On the other hand, real-time monitoring is not needed in advance prediction, and consumption of various expenses required by monitoring can be reduced. Various consumptions herein include, but are not limited to: electrical energy and/or software and hardware resources.
In the application of the Internet of things or large-scale cloud computing, the prediction method can complete computing under low algorithm time complexity, saves the computing power of resource scheduling modules of resource platforms such as cloud platforms and the like, and saves the running power consumption of embedded equipment of the Internet of things.
In other embodiments, as shown in fig. 2 and 3, the step S120 may include at least one of the following steps:
step S121: according to the load condition information, resource scheduling in a time period for predicting the resource platform is carried out;
step S122: and adjusting the monitoring configuration of the resource platform in a predicted time period according to the load condition information.
Thus, according to the load condition information, the resource scheduling in the predicted period can be performed in advance, and the resource scheduling includes but is not limited to at least one of the following:
for example, adjusting the resources contained by different prediction objects;
closing one or more predicted objects;
opening one or more predicted objects;
the number and/or type of resources in the one or more prediction objects that are in an active state is adjusted.
In some embodiments, the resource platform monitoring configuration in the prediction time period can be further performed according to the load condition information, so that, compared with the adoption of the uniform static monitoring configuration, the monitoring high configuration can be adjusted in advance, on one hand, unnecessary monitoring is reduced, and on the other hand, the phenomenon of monitoring omission or insufficient monitoring strength is reduced.
The monitoring configuration includes at least one of:
monitoring the frequency;
a monitored object;
a monitoring mode;
a monitor, etc.
In some embodiments, the step S120 may include: and determining the load change time of the prediction object according to the load condition information, and adjusting the running state of the resource object in the resource platform before the load change time in a prediction time period, wherein the prediction object is configured on the resource object.
In some embodiments, the adjusting the running state of the resource object in the resource platform comprises at least one of:
adjusting the over-selling ratio of the resource object in the resource platform according to the descending or ascending condition of the load;
when the load capacity is reduced, the prediction object is transferred from the resource object with low over-sale ratio to the resource object with high over-sale ratio;
when the load capacity is reduced, closing the idle resource object;
when the load capacity is increased, transferring the predicted object from the resource object with high over-selling ratio to the resource object with low over-selling ratio, and reducing the over-selling ratio of the resource object from which the predicted object is transferred; in some embodiments, if the resource objects with low over-selling ratio are in a closed state, the resource objects with low over-selling ratio need to be started first, and then the predicted object is migrated;
when the load capacity rises, the closed resource objects are started.
The super-selling ratio is as follows: the ratio of the number of allocated resources to the number of resources actually included in the resource object is a measure of the resource over-sale. For example, resource over-selling refers to the sum of the amount of resources allocated to a task by a resource sharing platform on a single resource object (e.g., a single host) being greater than the capacity of the single resource object. Since the resources occupied by the running of the tasks are far less than the resource allocation amount in part of time, even if the sum of the allocation amounts of the tasks is greater than the capacity of the server, the server can meet the resource requirements of the tasks during the task execution. With resource reselling, the number of tasks performed simultaneously on a single resource is greater than when not reselling. Different tasks have different resource requirements, and different resources (such as different server brands or series) have a great influence on the same task, and of course, even if the same task has different resource requirements at different times.
The load condition information may be used to determine a trend of the load to increase or decrease by comparing the current time period or a time period before the predicted time period. Then the resource is configured according to the change trend.
For example, a portion of the resources may be turned on or off based on the amount of load rising and falling.
In this embodiment, load balancing may be achieved and/or unnecessary resource objects started may be reduced by adjusting the over-sell ratio of resource objects in the resource platform.
The adjusting the over-selling ratio of the resource object in the resource platform according to the decreasing or increasing condition of the load amount may include one of:
when the load capacity is reduced, reducing the over-selling ratio of a single resource object in the resource platform, wherein the load balance is better realized through the reduction of the over-selling ratio of the single resource object, and the response rate of a single task is improved through the load balance;
when the load capacity is increased, the over-selling ratio of a single resource object of a preset type capable of supporting the high over-selling ratio is increased, wherein the resource platform can support more tasks and the task amount of the whole resource platform is increased through the increase of the over-selling ratio of the single resource object;
when the load capacity is reduced, the over-selling ratio of a single resource object of a preset type capable of supporting the high over-selling ratio is improved, and the predicted object is transferred to the resource object with the high over-selling ratio from the resource object with the low over-selling ratio, so that one or more resource objects with the low over-selling ratio are not configured with the predicted object, the starting state does not need to be kept, one or more idle resource objects with the low over-selling ratio can be closed, and the electric energy expenditure and the like of the whole resource platform are reduced.
In some embodiments, the scheduling of resources for the period of time during which the resource platform is predicted based on the load condition information includes:
when the load condition information indicates that: when at least one specific sub-period in the predicted period is within the period, scheduling reserved resources in the resource platform to share the load capacity, wherein the specific sub-period is as follows: a sub-period is predicted which contains the moment when the load rate is above the first threshold.
The predicted period may be divided into a plurality of sub-periods, e.g., the predicted period is equally divided into a plurality of sub-periods. And if the moment when the load rate is higher than the first threshold value occurs in at least one sub-period, scheduling the reserved resources in the resource platform to share the load capacity. The sub-period containing the moment when the load rate is higher than the first threshold is the specific sub-period.
In some embodiments, according to the currently monitored monitoring data, if it is predicted that the overtime time exists in the predicted time period, the load change trend is increased in the predicted time period, and if the increase rate reaches a preset threshold, the specific sub-time period may exist very greatly. For example, sudden overtime during non-working hours may cause a dramatic increase in office load; thus, even if the load of the burst is increased sharply, the resource allocation can be performed quickly.
In some embodiments, the scheduling reserved resources in the resource platform to share a load amount includes at least one of:
in the specific sub-period, reserving resource sharing load in resource objects with the enabled over-selling ratio exceeding the over-selling ratio threshold;
and enabling the reserved resource object to share the load in the specific sub-period.
The reserved resources can be reserved under the condition of dealing with high load capacity, so that the resource platform can bear the load in different time periods as required.
In the non-enabled condition, the reserved resources are in an off state, thereby reducing the power consumption resulting from the enablement of these reserved resources.
The reserved resources include, but are not limited to, at least one of:
computing resources, e.g., a CPU and/or GPU;
power supply resources, such as generators and/or transformers;
network resources, such as wired bandwidth and/or bandwidth of wireless traffic.
In some embodiments, one or more resource objects may be caused to be free when the amount of load decreases. A resource object herein includes, but is not limited to, one or more hosts, which may be worker nodes or physical machines. If the host is in an on state, power consumption still exists even under the condition of no load. If the host is turned off, the power consumption of the host is minimized.
Here, the free resource object includes an empty resource object.
In some embodiments, the closing of the free resource object may include: closing the resource object through a physical key may further include: and remotely closing the resource object in a remote control mode. For example, the corresponding resource object is turned off by the control device, for example, by powering down, or one or more idle resource objects are turned off without powering down by a soft switch within the resource object.
In some embodiments, the adjusting the monitoring configuration of the resource platform according to the load condition information includes:
and adjusting the monitoring frequency of the resource platform according to the load condition information.
In this embodiment, the monitoring frequency includes at least one of the following frequencies:
monitoring the data using frequency;
monitoring the frequency of analysis of the data;
frequency of warehousing of data.
The higher the acquisition frequency is, the higher the working frequency of the acquisition equipment for acquiring the monitoring data is, the higher the acquisition load of the acquisition equipment is, and the higher the energy consumption is,
the higher the analysis frequency is, the higher the working frequency of the analysis equipment for analyzing the monitoring data is, the larger the analysis load is, and the higher the energy consumption is.
The higher the warehousing frequency of the data is, the higher the frequency of writing the monitoring data and/or the analysis data into the database is, the larger the load is, and the higher the energy consumption is.
Analyzing the data may include: various analysis results of the data are monitored.
In some embodiments, the adjusting the monitoring frequency of the resource platform according to the load condition information includes at least one of:
when the load condition information indicates that the duration of the load rate of the resource platform lower than a second threshold reaches a first duration, determining to acquire monitoring data of the target condition at a first frequency;
and when the load condition information indicates that the load change rate of the resource platform is equal to or greater than a second threshold value and reaches a second time length, determining to acquire monitoring data of the target condition at a second frequency, wherein the second frequency is higher than the first frequency.
It can be seen that the load amount is positively correlated with the monitoring frequency, i.e. the monitoring frequency is higher if the load amount is larger, and the monitoring frequency is lower if the load amount is smaller. In this embodiment, the monitoring frequency includes at least an acquisition frequency. In some embodiments, the acquisition frequency and the analysis frequency are coincident.
In this embodiment, by dynamically adjusting the sampling frequency according to the load change rate, a suitable sampling frequency can be selected for sampling according to the load amount and/or the load change trend, so that on one hand, under the condition of a high load rate, it is ensured that under the condition that high-frequency sampling is required to obtain detailed monitoring data, a high sampling frequency is adopted, and under the condition of a low load rate, a lower sampling frequency is adopted to reduce unnecessary data sampling, and the overhead generated by unnecessary sampling data is reduced.
In some embodiments, the adjusting the monitoring frequency of the resource platform according to the load condition information includes:
determining the load rates of different types of resources according to the load condition information;
and respectively setting the monitoring frequency of the corresponding type resources according to the load rate of the corresponding type resources.
The types of resources may include at least:
storage resources including, but not limited to: a memory and/or hard disk;
computing resources, including but not limited to: a CPU and/or GPU;
network resources including, but not limited to: a bandwidth of wired and/or wireless traffic;
power supply resources including, but not limited to: an Uninterruptible Power Supply (UPS) and/or a transformer;
heat dissipation resources including, but not limited to: a fan in an air conditioner or a host, and the like.
The load rates of the different types of resources are different, and the monitoring frequency of the corresponding type can be dynamically adjusted. Therefore, the monitoring frequency of the single type resource is adjusted, the adjustment is refined, and the accuracy of the monitoring result and/or the analysis result is improved due to the refined adjustment.
In some embodiments, the method further comprises:
the monitoring data is stored to a first monitoring database separate from the monitoring device.
In some cases, the method further comprises: and querying a second monitoring database of the monitoring equipment according to the received query operation. And the monitoring equipment receives the monitoring data and then preferentially stores the monitoring data in the second monitoring database. On the one hand, the first monitoring database has a limited storage capacity, and on the other hand, if all the query operations are submitted to the first monitoring database of the monitoring device, this may result in an overload phenomenon of the monitoring device. In this embodiment, in order to reduce the problems in the two aspects, the monitoring data is transferred from the second monitoring database to the first monitoring database which is independent from the second monitoring database.
Therefore, some query operations are directly responded by the first monitoring database, load transfer of the monitoring data query operations is achieved, and overload phenomena of the monitoring data caused by responding to the query operations are reduced.
For example, monitoring data will be transferred from the second monitoring database to the first monitoring database on a regular basis.
The first monitoring database includes, but is not limited to: and if the time sequence database is used, according to the time parameters carried by the query operation, the rapid query based on the time sequence can be realized, and the query efficiency of monitoring data and/or analyzing data is improved. The monitoring data includes at least one of:
load data for one or more resources at different points in time, the load data including but not limited to: load rate and/or load amount;
status data for one or more resources at different points in time, the status data being usable to characterize the health of the resource and/or the amount of available resource, e.g., if there is a resource anomaly, then the resource cannot be used; if some resources are overloaded, the resources are in a sub-health state, and if some resources are normal, the resources are in a monitoring state;
the performance data of one or more resources at different time points, some resources are good in performance when being put into use newly, aging can occur along with increase of the use times, and further performance is reduced, and the performance data can well represent the performance of the resources in the corresponding time points.
As shown in fig. 4, the present embodiment provides an electronic device, including:
the prediction module 110 is configured to predict load condition information of a prediction object in a resource platform according to a service scenario of the resource platform;
an executing module 120, configured to execute, according to the load condition information, a predetermined operation adapted to the load condition information on the resource platform in a predicted time period.
In some embodiments, the electronic device may be at least one of a monitoring device, a resource scheduling device, and/or a control device of a resource platform.
In some embodiments, the prediction module 110 and the execution module 120 may be program modules; the program module is capable of predicting the load condition information and executing the predetermined operation after being executed by the processor.
In other embodiments, the prediction module 110 and the execution module 120 may be a combination of hardware and software modules; the soft and hard combining module includes but is not limited to various programmable arrays. The programmable array includes, but is not limited to: complex programmable arrays and/or field programmable arrays.
In still other embodiments, the prediction module 110 and the execution module 120 may be purely hardware modules; the pure hardware module includes, but is not limited to, a complex programmable array.
In some embodiments, the execution module 120 is specifically configured to determine a load change time of the predicted object according to the load condition information, and adjust an operating state of a resource object in the resource platform before the load change time in a predicted time period; and/or the execution module 120 is specifically configured to adjust the monitoring frequency of the resource platform according to the load condition information.
In some embodiments, the execution module 120 is specifically configured to determine a load change time of the predicted object according to the load condition information, and adjust an operating state of a resource object in the resource platform before the load change time in a predicted time period, where the predicted object is configured on the resource object.
In some embodiments, the executing module 120 is specifically configured to execute at least one of the following: adjusting the over-selling ratio of the resource object in the resource platform according to the descending or ascending condition of the load; when the load capacity is reduced, transferring the predicted object from the resource object with low over-selling ratio to the resource object with high over-selling ratio, and reducing the over-selling ratio of the resource object from which the predicted object is transferred; when the load capacity is reduced, closing the idle resource object; when the load is increased, the prediction object is transferred from the resource object with high over-sale ratio to the resource object with low over-sale ratio; when the load capacity rises, the closed resource objects are started.
In some embodiments, the executing module 120 is specifically configured to, when the load condition information indicates: when at least one specific sub-period in the predicted period is within the period, scheduling reserved resources in the resource platform to share the load capacity, wherein the specific sub-period is as follows: a sub-period is predicted which contains the moment when the load rate is above the first threshold.
In some embodiments, the executing module 120 is specifically configured to, in the specific sub-period, reserve resource sharing load in a resource object whose enabled over-sell ratio exceeds the over-sell ratio threshold; and enabling the reserved resource object to share the load in the specific sub-period.
In some embodiments, the execution module 120 is specifically configured to adjust the monitoring frequency of the resource platform according to the load condition information.
Several specific examples are provided below in connection with any of the embodiments described above:
example 1:
the present example provides an information processing method, which may include:
through correlation analysis with the working calendar, the correlation between the load of the virtual machine and the working calendar is obtained, the load of the virtual machine is calculated based on the working calendar time of administrative organization, and the load prediction accuracy of the virtual machine can be greatly improved.
Based on the analysis result, the load capacity drop and the back-up time of the virtual machine can be predicted.
And dynamically adjusting the over-selling ratio of the host machine based on the load, and migrating the virtual machine to a server with a high over-selling ratio after the load is reduced.
When all the virtual machines of one host machine are migrated, the idle host machine is moved out of the cloud platform, namely the idle host machine computing resources are remotely shut down, so that the power consumption of the data center is reduced.
And after the load capacity rises back, remotely starting the host machine. And executing the health check of the newly started host machine, and accessing the checked host machine to the private cloud computing resource pool.
And migrating the virtual machine with the increased load from the host with the high overfall ratio to the host with the normal overfall ratio (the host comprises but is not limited to a server), and setting the overfall ratio of the host back to the normal overfall ratio.
When the predicted load of the virtual machine is reduced, but the load of the virtual machine is increased due to burst traffic, such as overtime, the relevant virtual machine is migrated back to the host machine by reserving resources (such as reserved cache), starting the host machine, and the like, so as to ensure the running efficiency of the application on the virtual machine.
The foregoing is described in the form of a private cloud virtual machine, and is also applicable to other cloud platforms, such as containers, big data platform computing services, and the like.
Through a resampling mode and a time series algorithm, missing data points in the middle of low-frequency sampling data can be complemented back. Therefore, a method such as a normal time series analysis can still be performed based on such time series data resampling.
The above is described in the form of a private cloud virtual machine, and is also applicable to other cloud resources, such as a container, a big data platform computing service, and the like.
For another example, the acquisition frequency of the acquisition platform is determined according to the activity degree of the acquisition object. The acquisition objects can be divided into: animate and inanimate objects. For example, if the change is small after the animal enters the sleep, the sleep and activity of the animal can be predicted according to factors such as solar terms, calendars, illumination, weather forecast and the like. The sampling frequency is decreased during sleep of the animal and increased during activity. This will likely reduce the power consumption of the internet of things sampling device. Although the existing patent has an analysis method, a method for reducing the power consumption of the internet of things equipment and reducing the storage pressure of the back-end time series data by changing the sampling frequency is not provided.
Example 2:
and the load of the computing resources is computed according to the historical load, so that the prediction accuracy is poor. By combining with a business scene, such as a working calendar, prediction accuracy can be greatly improved.
The over-selling ratio of the host machine is dynamically adjusted, and remote shutdown and/or remote startup operation of the host machine is carried out, so that the power consumption of the data center is favorably reduced. The power consumption of the data center is one of the most important cost expenses, and the design is favorable for improving the cost of the data center, reducing the power consumption and realizing economic value and social value.
Sudden load changes will likely cause the traffic to run slower. By reserving resources (e.g., reserving cache) and triggering based on load size, emergency response in case of emergency or change of service environment is facilitated by the host's startup and/or migration of virtual machines.
The whole cloud platform resource can be conveniently regulated, the computing resource is provided for the service with high priority, and the normal operation of the service is ensured.
Example 3:
data preparation and offline analysis phases
The data preparation may include:
acquiring resource load data from the monitoring system, for example, CPU load, memory allocation, disk Input Output (IO), network Input Output (IO), and the like. Therefore, it is desirable to be able to acquire metrology data from a monitoring system. The measurement data can be data which can evaluate loads and/or resources at different latitudes or represent objective phenomena in the monitoring data; e.g., load rate and/or number of free resources, etc.
Since the monitoring system cannot generally carry a large amount of data query operations, the database of the monitoring system may be over-stressed. It may be necessary to periodically import the data into a separate database, such as a time series database, for subsequent off-line data analysis.
The offline analysis may include:
and (3) periodically calculating the resource load characteristics of the cloud platform, and analyzing whether the load of the cloud platform is in positive correlation with the working calendar or not, such as predicting the load characteristics of the object. If the prediction object exhibits a property that is positively correlated with the working calendar, it is recorded to a time frame that can be over-sold.
And acquiring the attribute of the cloud platform load according to the configuration management database or the similar function database, wherein the attribute belongs to production level application and the like, for example, through connection physical hardware exists, for example, a GPU or an FPGA card is connected.
And generating a server range capable of executing resource migration according to the working calendar according to the resource state capable of executing the periodical increase of the over-sale ratio and by combining the resource information positively correlated with the working calendar.
The work calendar may come from a state-issued legal vacation, an in-company vacation or activity, or may come from a departmental vacation or on-duty schedule. The attributes of the work calendar may be marked according to the administrative department or project team associated with the business scenario to which it belongs as a boundary.
The cloud computing resources and the working calendar mostly have relevance in units of days, but there are also business systems having time relevance in units of natural months or financial years. The matching relationship between the common cloud computing resource and the working calendar may be:
the working time is long, and the load of the server is large; the rest time server load is reduced, for example, by a user self-service administration tool.
The working time is short, and the load of the server is small; rest time, server load is large, e.g., some ETL systems.
And presenting the service characteristics related to the natural month or the financial period, such as monthly report generation, financial year data analysis and the like.
During the large-scale period, the dynamic allocation of resources, such as 618, twenty-one, etc. electric quotient. The working calendar associated with such a business scenario is also marked as a high load state.
Unlike monitoring where sampling is performed at a fixed frequency, devices that monitor variable frequency sampling in the calendar mode typically require time checking by a timing program, such as a time server. Only accurate timing can each time point of the frequency conversion process be accurately determined.
As an option, time zone conversion is required for frequency conversion operation to ensure that equipment in different time zones around the world can work normally. The time zone can use the time zone of the equipment, and the UTC time zone is generally used for unified management.
And migrating the load of the virtual machine to a server with a high over-sale ratio, and remotely shutting down the host machine.
Offline host resource analysis and host preparation may include:
according to the range of the virtual machines obtained in the last step, the number of the virtual machines with dynamically increased over-sell ratio, virtual CPUs (central processing units) of different virtual machines, the occupied size of a memory, CPU instruction set information and the like can be calculated. And obtaining the number of hosts which can be dynamically migrated in different types based on the number.
And adjusting the over-sale ratio of the host machine resources for load compression by using the cloud platform API. The virtual machine can be migrated further, usually only if the resell ratio setting is increased.
The adjusted high over-sell ratio host range is recorded in a database, such as a configuration management database or a cloud platform management database.
Virtual machine migration and load reduction may include:
after the host range is determined and the over-sell ratio is set, the virtual machine starts to be migrated to the host with the high over-sell ratio when the virtual machine starts to be migrated is reached.
After migration of each virtual machine is completed, it may happen that many hosts run the virtual machines simultaneously, but the migration of the virtual machines is not completed completely. At this time, calculation is needed, and a host range capable of completely migrating the virtual machine is selected. And migrating all the virtual machines in the hosts to other hosts, and then shutting down the hosts by using a remote management tool.
Restarting and recovering the shutdown host, wherein the recovering the over-sale ratio of the host can comprise:
and according to the working calendar correlation marked by the virtual machine, before the working time begins, the virtual machine begins to be migrated back to the host with the normal over-selling ratio.
The off state host is powered on using remote management software or an interface some time before the working calendar flags the start time, e.g., 1 hour in advance. Such remote management software may be IPMI based or may be management software based on different vendors, such as lenova xClarity.
And after the host machine is started, executing an automatic checking program to ensure that the running state of each host machine is normal. And if the host is normal, accessing the host to a cloud computing platform, for example, to an OpenStack platform. After access, the relevant database, such as the configuration management database or the cloud platform management database, is updated.
And calculating the over-selling ratio of the host after the host migrates the virtual machine, namely the over-selling ratio of the host entering a high-load operation stage.
And migrating the running host with the high over-selling ratio to restore the running virtual machine to the running host started in the previous step. And after the virtual machine is migrated out, namely when the load of the host machine is reduced to be within the range of the over-selling ratio calculated previously, modifying the over-selling ratio configuration of the private cloud about the host machine. If not, then in the subsequent load balancing process, several virtual machines may be migrated back to the host.
The high over-sell ratio host machine burst load handling can comprise:
when the virtual machine is predicted to enter low-load operation according to the working calendar and is migrated to the host with high over-selling ratio, a temporary job occurs to cause a short-time load increase, such as collective overtime or temporary job.
In the first scheme, certain machines are reserved outside standard load and high-load servers according to a certain quantity or proportion, and when burst load occurs, the servers can quickly bear the burst load.
And a certain cache is reserved in the process of carrying out the over-selling ratio calculation, so that the load is suddenly increased, and a certain capacity is reserved for receiving the migration of the high-load virtual machine.
And thirdly, starting the host machine in the shutdown state, and migrating the virtual machine to the newly started host machine for running.
And step four, fusing mixed execution of the three schemes.
Example 4:
the present example provides a method of performing a monitor sampling frequency adjustment based on predicted load condition information, which may include:
according to the analysis process, the correlation between the resource load and the working calendar is identified, the resource load is known to be reduced according to the working calendar, and the low-load operation state is entered. The sampling frequency of the monitoring sampling program is down-converted.
The configuration of the updated sampling frequency downconversion implementation may include:
issuing configuration and operation instructions to a sampling terminal through a central monitoring server;
subscribing configuration information of a monitoring server through a monitoring sampling terminal;
the remotely running performance sampling program is orchestrated by an automated orchestration tool to adjust the configuration, e.g., modify and validate Telegraf configuration information using the Angle/SaltStack.
The monitoring frequency of different applications within the same cloud computing resource may be different. For example, the monitoring items such as the CPU and the memory are subjected to frequency reduction operation from once sampling per second to once sampling 15 seconds, and the execution frequency of the database-holding monitoring sampling script is 10 minutes. Therefore, the frequency reduction setting may be global, and the frequency reduction operation may be performed on different monitoring items or monitoring indexes according to the service scenario.
According to the analysis process, for the cloud computing resources entering the low-load operation, through the correlation analysis with the working calendar, if the time to enter the normal-load operation is less than the set time threshold value delta, the monitoring data acquisition frequency returns to normal. The configuration of monitoring frequency recovery and the instruction issuing mode are consistent with the frequency reduction.
The burst load handling may include:
when the virtual machine is predicted to enter low load operation from the working calendar, but a temporary job occurs with an increase in load, for example, a collective overtime or temporary job. Then one or more of the following policies are selected according to the rules:
strategy one: and if the load does not obviously drop for a plurality of continuous sampling time periods, the frequency modulation operation of the cloud computing resource is canceled in the period, and the normal frequency is used for monitoring and sampling.
And (2) strategy two: monitoring the load of the cloud computing resources, and reducing the monitoring sampling frequency when the load is reduced to a certain threshold value within a time window range needing frequency reduction for a continuous period of time.
Example 5:
the present example provides an information processing method, which may include:
and determining a service scene through working calendar analysis and the like, and comprehensively predicting the load condition information in the resource platform by combining auxiliary scenes except the service scene and combining the service scene and the auxiliary scenes.
For example, the service scenario run by the resource platform includes: an office scene; the auxiliary scene comprises: in the parking scene of the parking lot of the office building, if the number of vehicle inputs in the parking lot is rapidly increased, the load capacity is predicted to be increased after 15 minutes or half an hour by combining the service scene and the parking scene.
For another example, the service scenario executed by the resource platform includes: a mobile communication service scenario; the auxiliary scene may be a human flow monitoring scene; if the people flow detection scene detects that the people flow changes rapidly, load condition information in a prediction time end after the specified time is predicted by combining the mobile communication service scene and the people flow monitoring scene.
And scheduling the resources in the resource platform and/or changing the monitoring configuration of the monitoring system of the resource platform according to the predicted load condition information. In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, all functional units in the embodiments of the present application may be integrated into one processing module, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
Technical features disclosed in any embodiment of the present application may be combined arbitrarily to form a new method embodiment or an apparatus embodiment without conflict.
The method embodiments disclosed in any embodiment of the present application can be combined arbitrarily to form a new method embodiment without conflict.
The device embodiments disclosed in any embodiment of the present application can be combined arbitrarily to form a new device embodiment without conflict.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. An information processing method comprising:
predicting load condition information of a prediction object in a resource platform according to a service scene of the resource platform;
and according to the load condition information, executing a predetermined operation matched with the load condition information on the resource platform in a predicted time period.
2. The method of claim 1, wherein the performing, based on the load condition information, a load condition information-adapted predetermined operation on the resource platform for a predicted period of time comprises at least one of:
according to the load condition information, resource scheduling in a time period for predicting the resource platform is carried out;
and adjusting the monitoring configuration of the resource platform in a predicted time period according to the load condition information.
3. The method of claim 2, wherein the scheduling of resources for the period of time during which the resource platform is predicted from the load condition information comprises:
and determining the load change time of the prediction object according to the load condition information, and adjusting the running state of the resource object in the resource platform before the load change time in a prediction time period, wherein the prediction object is configured on the resource object.
4. The method of claim 3, wherein the adjusting the running state of the resource object in the resource platform comprises at least one of:
adjusting the over-selling ratio of the resource object in the resource platform according to the descending or ascending condition of the load;
when the load capacity is reduced, the prediction object is transferred from the resource object with low over-sale ratio to the resource object with high over-sale ratio;
when the load capacity is reduced, closing the idle resource object;
when the load capacity is increased, transferring the predicted object from the resource object with high over-selling ratio to the resource object with low over-selling ratio, and reducing the over-selling ratio of the resource object from which the predicted object is transferred;
when the load capacity rises, the closed resource objects are started.
5. The method of claim 2, wherein the scheduling of resources for the period of time during which the resource platform is predicted from the load condition information comprises:
when the load condition information indicates that: when at least one specific sub-period in the predicted period is within the period, scheduling reserved resources in the resource platform to share the load capacity, wherein the specific sub-period is as follows: a sub-period is predicted which contains the moment when the load rate is above the first threshold.
6. The method of claim 5, wherein the scheduling reserved resources in the resource platform to share a load amount comprises at least one of:
in the specific sub-period, reserving resource sharing load in resource objects with the enabled over-selling ratio exceeding the over-selling ratio threshold;
and enabling the reserved resource object to share the load in the specific sub-period.
7. The method of claim 2, wherein said adjusting the monitoring configuration for the resource platform based on the load condition information comprises:
and adjusting the monitoring frequency of the resource platform according to the load condition information.
8. An electronic device, comprising:
the prediction module is used for predicting the load condition information of a prediction object in the resource platform according to the service scene of the resource platform;
and the execution module is used for executing the preset operation matched with the load condition information to the resource platform in a predicted time period according to the load condition information.
9. The electronic device of claim 8, wherein the prediction module is specifically configured to perform at least one of:
according to the load condition information, resource scheduling in a time period for predicting the resource platform is carried out;
and adjusting the monitoring configuration of the resource platform in a predicted time period according to the load condition information.
10. The electronic device according to claim 9, wherein the execution module is specifically configured to determine a load change time of the prediction object according to the load condition information, and adjust an operating state of a resource object in the resource platform before the load change time in a predicted time period; and/or adjusting the monitoring frequency of the resource platform according to the load condition information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910810668.5A CN110609747A (en) | 2019-08-29 | 2019-08-29 | Information processing method and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910810668.5A CN110609747A (en) | 2019-08-29 | 2019-08-29 | Information processing method and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110609747A true CN110609747A (en) | 2019-12-24 |
Family
ID=68890686
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910810668.5A Pending CN110609747A (en) | 2019-08-29 | 2019-08-29 | Information processing method and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110609747A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112631784A (en) * | 2020-12-31 | 2021-04-09 | 联想未来通信科技(重庆)有限公司 | Resource-based allocation/processing method and electronic equipment |
| CN112866131A (en) * | 2020-12-30 | 2021-05-28 | 神州绿盟成都科技有限公司 | Traffic load balancing method, device, equipment and medium |
| CN113377539A (en) * | 2021-06-10 | 2021-09-10 | 中国工商银行股份有限公司 | Processing method and device for realizing load balance |
| CN113448943A (en) * | 2021-06-21 | 2021-09-28 | 北京优炫软件股份有限公司 | Database configuration method and device, electronic equipment and readable storage medium |
| CN113778663A (en) * | 2021-07-28 | 2021-12-10 | 荣耀终端有限公司 | Scheduling method of multi-core processor and electronic equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102711177A (en) * | 2012-04-26 | 2012-10-03 | 北京邮电大学 | Service prediction based load balancing method |
| CN103220337A (en) * | 2013-03-22 | 2013-07-24 | 合肥工业大学 | Cloud computing resource optimizing collocation method based on self-adaptation elastic control |
| CN103778474A (en) * | 2012-10-18 | 2014-05-07 | 华为技术有限公司 | Resource load capacity prediction method, analysis prediction system and service operation monitoring system |
| CN103970587A (en) * | 2014-04-21 | 2014-08-06 | 华为技术有限公司 | Resource scheduling method, device and system |
| CN105302632A (en) * | 2015-11-19 | 2016-02-03 | 国家电网公司 | Cloud computing working load dynamic integration method |
| CN105868004A (en) * | 2015-01-23 | 2016-08-17 | 中兴通讯股份有限公司 | Cloud computing based business system scheduling method and apparatus |
| US20170331705A1 (en) * | 2015-01-30 | 2017-11-16 | Huawei Technologies Co., Ltd. | Resource Scaling Method on Cloud Platform and Cloud Platform |
| CN108241526A (en) * | 2016-12-26 | 2018-07-03 | 摩根士丹利服务集团有限公司 | ADAPTIVE CONTROL based on model prediction |
| CN109936473A (en) * | 2017-12-19 | 2019-06-25 | 华耀(中国)科技有限公司 | A distributed computing system based on deep learning prediction and its operation method |
-
2019
- 2019-08-29 CN CN201910810668.5A patent/CN110609747A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102711177A (en) * | 2012-04-26 | 2012-10-03 | 北京邮电大学 | Service prediction based load balancing method |
| CN103778474A (en) * | 2012-10-18 | 2014-05-07 | 华为技术有限公司 | Resource load capacity prediction method, analysis prediction system and service operation monitoring system |
| CN103220337A (en) * | 2013-03-22 | 2013-07-24 | 合肥工业大学 | Cloud computing resource optimizing collocation method based on self-adaptation elastic control |
| CN103970587A (en) * | 2014-04-21 | 2014-08-06 | 华为技术有限公司 | Resource scheduling method, device and system |
| CN105868004A (en) * | 2015-01-23 | 2016-08-17 | 中兴通讯股份有限公司 | Cloud computing based business system scheduling method and apparatus |
| US20170331705A1 (en) * | 2015-01-30 | 2017-11-16 | Huawei Technologies Co., Ltd. | Resource Scaling Method on Cloud Platform and Cloud Platform |
| CN105302632A (en) * | 2015-11-19 | 2016-02-03 | 国家电网公司 | Cloud computing working load dynamic integration method |
| CN108241526A (en) * | 2016-12-26 | 2018-07-03 | 摩根士丹利服务集团有限公司 | ADAPTIVE CONTROL based on model prediction |
| CN109936473A (en) * | 2017-12-19 | 2019-06-25 | 华耀(中国)科技有限公司 | A distributed computing system based on deep learning prediction and its operation method |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112866131A (en) * | 2020-12-30 | 2021-05-28 | 神州绿盟成都科技有限公司 | Traffic load balancing method, device, equipment and medium |
| CN112631784A (en) * | 2020-12-31 | 2021-04-09 | 联想未来通信科技(重庆)有限公司 | Resource-based allocation/processing method and electronic equipment |
| CN113377539A (en) * | 2021-06-10 | 2021-09-10 | 中国工商银行股份有限公司 | Processing method and device for realizing load balance |
| CN113448943A (en) * | 2021-06-21 | 2021-09-28 | 北京优炫软件股份有限公司 | Database configuration method and device, electronic equipment and readable storage medium |
| CN113778663A (en) * | 2021-07-28 | 2021-12-10 | 荣耀终端有限公司 | Scheduling method of multi-core processor and electronic equipment |
| CN113778663B (en) * | 2021-07-28 | 2023-05-12 | 荣耀终端有限公司 | Scheduling method of multi-core processor and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110609747A (en) | Information processing method and electronic equipment | |
| Radovanović et al. | Carbon-aware computing for datacenters | |
| Brown | Report to congress on server and data center energy efficiency: Public law 109-431 | |
| CN102326134B (en) | Energy-aware server admin | |
| US9003211B2 (en) | Method and apparatus for holistic power management to dynamically and automatically turn servers, network equipment and facility components on and off inside and across multiple data centers based on a variety of parameters without violating existing service levels | |
| US9153965B2 (en) | System and method for energy storage management | |
| US20120144219A1 (en) | Method of Making Power Saving Recommendations in a Server Pool | |
| US20110289329A1 (en) | Leveraging smart-meters for initiating application migration across clouds for performance and power-expenditure trade-offs | |
| US8327170B2 (en) | Computing load management method and system | |
| US20110191773A1 (en) | System and Method for Datacenter Power Management | |
| Cheng et al. | Heterogeneity-aware workload placement and migration in distributed sustainable datacenters | |
| Bates et al. | Electrical grid and supercomputing centers: An investigative analysis of emerging opportunities and challenges | |
| US20100318827A1 (en) | Energy use profiling for workload transfer | |
| US11960255B2 (en) | Compute load shaping using virtual capacity and preferential location real time scheduling | |
| Cao et al. | Energy-efficient resource management for scientific workflows in clouds | |
| Li et al. | Opportunistic scheduling in clouds partially powered by green energy | |
| US11334135B1 (en) | Power supply optimization using backup battery power supplementation | |
| Bajracharyay et al. | Economic analysis of a data center virtual power plant participating in demand response | |
| KR102350298B1 (en) | Energy management system and management method therefor | |
| CN106095531A (en) | A kind of dispatching method of virtual machine loaded based on grade and physical machine in cloud platform | |
| Chang et al. | Optimizing energy consumption for a performance-aware cloud data center in the public sector | |
| Wolski et al. | QPRED: Using quantile predictions to improve power usage for private clouds | |
| CN110806918A (en) | Virtual machine operation method and device based on deep learning neural network | |
| US9195514B2 (en) | System and method for managing P-states and C-states of a system | |
| CN110033144A (en) | A kind of the dynamic self-adapting appraisal procedure and system of translation project risk |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191224 |
|
| RJ01 | Rejection of invention patent application after publication |