CN114490012B - Memory scheduling method, device and readable storage medium - Google Patents
Memory scheduling method, device and readable storage medium Download PDFInfo
- Publication number
- CN114490012B CN114490012B CN202011267888.7A CN202011267888A CN114490012B CN 114490012 B CN114490012 B CN 114490012B CN 202011267888 A CN202011267888 A CN 202011267888A CN 114490012 B CN114490012 B CN 114490012B
- Authority
- CN
- China
- Prior art keywords
- data
- accessed
- main cause
- memory
- order
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 230000015556 catabolic process Effects 0.000 claims abstract description 64
- 238000006731 degradation reaction Methods 0.000 claims abstract description 64
- 238000012937 correction Methods 0.000 claims description 22
- 238000004590 computer program Methods 0.000 claims description 7
- 230000000694 effects Effects 0.000 claims description 3
- 230000000875 corresponding effect Effects 0.000 description 178
- 238000005457 optimization Methods 0.000 description 30
- 238000004458 analytical method Methods 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 239000011159 matrix material Substances 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012886 linear function Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 2
- 230000006866 deterioration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 208000018910 keratinopathic ichthyosis Diseases 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000006467 substitution reaction 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
- G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
- G06F9/5016—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals the resource being the memory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F5/00—Methods or arrangements for data conversion without changing the order or content of the data handled
- G06F5/06—Methods or arrangements for data conversion without changing the order or content of the data handled for changing the speed of data flow, i.e. speed regularising or timing, e.g. delay lines, FIFO buffers; over- or underrun control therefor
- G06F5/065—Partitioned buffers, e.g. allowing multiple independent queues, bidirectional FIFO's
-
- 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
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4812—Task transfer initiation or dispatching by interrupt, e.g. masked
- G06F9/4831—Task transfer initiation or dispatching by interrupt, e.g. masked with variable priority
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明实施例提供一种内存调度方法、装置及可读存储介质，本发明方法实施例中，基于劣化小区工单对应的主因，确定各主因对应的待访问数据的全局优先调度次序，其中，所述劣化小区工单对应的主因指导致所述劣化小区对应的网络性能指标劣化的主要原因，基于所述各主因对应的待访问数据的全局优先调度次序，调度内存数据，通过基于劣化小区工单对应的主因，确定各主因对应的待访问数据的全局优先调度次序，基于所述各主因对应的待访问数据的全局优先调度次序，调度内存数据，不仅可以减少缓存无关数据的内存空间，而且提高了内存数据命中率。
An embodiment of the present invention provides a memory scheduling method, device and readable storage medium. In the method embodiment of the present invention, based on the main cause corresponding to the degraded cell work order, the global priority scheduling order of the data to be accessed corresponding to each main cause is determined, wherein the main cause corresponding to the degraded cell work order refers to the main reason causing the degradation of the network performance indicator corresponding to the degraded cell. Based on the global priority scheduling order of the data to be accessed corresponding to each main cause, the memory data is scheduled. By determining the global priority scheduling order of the data to be accessed corresponding to each main cause based on the main cause corresponding to the degraded cell work order, the memory data is scheduled based on the global priority scheduling order of the data to be accessed corresponding to each main cause. This can not only reduce the memory space for caching irrelevant data, but also improve the hit rate of memory data.
Description
Technical Field
The present invention relates to the field of IT applications, and in particular, to a memory scheduling method, apparatus, and readable storage medium.
Background
The intelligent centralized optimization platform provides support for centralized optimization work in province and city, processes problems of the work orders of the degraded cells, improves network quality, optimizes staff to analyze wireless network problems of the cells, completes partial scheme execution work in daily network optimization and special optimization, and completes processing of the work orders of the degraded cells.
The intelligent centralized optimization platform is a big data analysis platform, because of the large number of related data sources and large data volume, the original analysis data is usually stored on a Hadoop/Hbase (Hadoop database) big data platform, and the analysis result data is usually stored in an MPP (MASSIVELY PARALLEL Processing, massive parallel Processing) database, such as a popular GreenPlum database or Gbase database. When a user queries a system analysis result, the analysis data is generally extracted from an MPP database through a lightweight J2EE system frame such as SSM (SPRING MVC, spring, mybatis) and the like, and then the data is rendered into a multimedia page such as a table or a map by utilizing a front end frame for the user to query.
In order to assist network optimization personnel in diagnosing network problems, the system needs to extract analysis data from a database for a plurality of times to carry out page display. Because the analysis result data is large, the database is time-consuming to search, the network transmission speed and other factors affect, and the speed of page query data is usually slower. Single page reaction times are sometimes above 30 seconds, severely affecting the diagnostic efficiency of network problems.
The existing memory scheduling method basically stores recent data in a memory, and the common memory management policies at present include FIFO (First Input First Output, first-in first-out method), LRU (LEAST RECENTLY Used ), random policy, etc., where FIFO means that when the memory usage reaches a certain threshold, the data that enters the memory first is removed, and LRU means that the data that is least recently Used is removed. The random strategy randomly selects a part of data to remove the memory, thereby freeing the memory and calling the data needed by the system into the memory.
Although the above memory management strategy can accelerate the system response speed to a certain extent, a large data analysis system like an intelligent centralized optimization platform usually needs to cache the whole network analysis data for 7 days, if a traditional memory scheduling method such as FIFO or LRU, random scheduling and the like is adopted, the cache quantity is required to be huge, and because some analysis dimension data needs the whole display of 7 days index, the FIFO is used to remove the memory of the data of the first day, and the memory needs to be reloaded back to the memory soon, so that the cache data jitter is caused, therefore, the memory scheduling method needs a quite large physical memory, and the data cannot be efficiently transferred in and out in the centralized network optimization system.
Disclosure of Invention
Aiming at the technical problems existing in the prior art, the embodiment of the invention provides a memory scheduling method, a memory scheduling device and a readable storage medium.
In a first aspect, an embodiment of the present invention provides a memory scheduling method, including:
Determining a global priority scheduling sequence of data to be accessed corresponding to each main factor based on the main factor corresponding to the degraded cell work order, wherein the main factor corresponding to the degraded cell work order leads to the main reason of the degradation of the network performance index corresponding to the degraded cell;
and scheduling the memory data based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor.
Optionally, according to an embodiment of the present invention, the determining, based on the main factors corresponding to the degraded cell worksheets, a global priority scheduling order of data to be accessed corresponding to each main factor specifically includes:
Determining data to be accessed corresponding to each main factor based on the main factor corresponding to the degraded cell work order;
Determining a global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor in a preset time period;
the main factor corresponding to the degraded cell work order and the priority of the main factor are determined based on each sub-factor corresponding to the degraded cell work order and the influence weight thereof, wherein the influence weight is used for representing the influence degree of the sub-factor corresponding to the degraded cell work order on the cell network performance index.
Optionally, according to an embodiment of the present invention, the determining, based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor within a preset duration, the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
And determining the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority and the weight corresponding to the priority of the main factor corresponding to the degraded cell work order, the historical access frequency of the data to be accessed in the preset time length and the weight corresponding to the access frequency.
Optionally, according to an embodiment of the present invention, the memory scheduling method performs scheduling of memory data based on a global priority scheduling order of data to be accessed corresponding to each main factor, and specifically includes:
And loading the data to be accessed into the memory in sequence according to the priority from high to low based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, and/or sequentially removing the data to be accessed from the memory according to the priority from low to high.
Optionally, according to the memory scheduling method of an embodiment of the present invention, the impact weight is determined based on a preset correction coefficient and probability association information, heartbeat association information and pearson association information of a child factor corresponding to the degraded cell work order and a cell network performance index, and the preset correction coefficient is determined based on a conditional probability that a main factor corresponding to the historical degraded cell work order leads to degradation of the cell network performance index.
Optionally, according to an embodiment of the present invention, after determining a global priority scheduling order of data to be accessed corresponding to each main factor, the memory scheduling method further includes:
and updating the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the access frequency of the data to be accessed corresponding to each main factor in a preset time period after the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined.
Optionally, according to the memory scheduling method of an embodiment of the present invention, the conditional probability that the main factor corresponding to the history degraded cell worksheet causes degradation of the cell network performance index is calculated by a bayesian formula based on the ratio of degradation of the cell network performance index caused by the main factor in all the history degraded cell worksheets, the ratio of degradation of the cell worksheets caused by the main factor in all the history degraded cell worksheets, and the ratio of degradation of the cell network performance index in all the history degraded cell worksheets.
In a second aspect, an embodiment of the present invention further provides a memory scheduling device, including:
The priority scheduling sequence determining module is used for determining the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the main factor corresponding to the work order of the degraded cell, wherein the main factor corresponding to the work order of the degraded cell refers to the main reason for the degradation of the network performance index corresponding to the degraded cell;
and the memory scheduling module is used for scheduling the memory data based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method as provided in the first aspect above when the program is executed.
In a fourth aspect, embodiments of the present invention also provide a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method as provided in the first aspect described above.
According to the memory scheduling method, the memory scheduling device and the readable storage medium, the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined based on the main factor corresponding to the degraded cell work order, and the memory data is scheduled based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, so that the memory space of the cache irrelevant data can be reduced, and the hit rate of the memory data is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a memory scheduling method according to an embodiment of the present invention;
fig. 2 is a detailed flow chart of a memory scheduling method according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a memory scheduling device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The existing memory scheduling method not only needs a quite large physical memory, but also cannot efficiently perform data input and output. Fig. 1 is a flow chart of a memory scheduling method according to an embodiment of the present invention, as shown in fig. 1, where the method includes:
Step 110, determining a global priority scheduling order of data to be accessed corresponding to each main factor based on the main factor corresponding to the degraded cell work order, wherein the main factor corresponding to the degraded cell work order is the main cause of degradation of the network performance index corresponding to the degraded cell.
Specifically, in the centralized network optimization process, the degraded cell worksheet corresponds to more than one of the main factors, i.e., the main factors causing the degradation of the network performance index. The main factor corresponding to the degraded cell work order is determined from a plurality of sub factors, wherein the sub factors are all possible reasons which can cause the degradation of the cell network performance index according to the prior experience summary, and comprise the dimensions of fault alarming, coverage, interference, resource use, parameter configuration and modification, neighbor cells and the like.
The network optimization engineer generally queries corresponding analysis data, i.e. data to be accessed, according to the main factor output by the centralized network optimization system as a main basis for diagnosing network problems. Based on the method, the global priority scheduling sequence of the data to be accessed corresponding to each main factor can be determined according to the main factor corresponding to the degraded cell work order, so that the network optimization engineer can access the data.
And step 120, scheduling the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor.
Specifically, the memory scheduling device schedules the memory data based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, stores or moves the data to be accessed corresponding to each main factor into or out of the memory according to the priority scheduling sequence, avoids the caching of irrelevant data, and enables a network optimization engineer to quickly access the data related to the wireless network problem. The data to be accessed corresponding to each main factor may be stored in a data table form, so as to be convenient for calling, and of course, it may also be stored in other forms according to actual needs, and the embodiment of the invention is not limited in detail.
According to the method provided by the embodiment of the invention, the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined based on the main factor corresponding to the degraded cell work order, and the memory data is scheduled based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, so that the memory space of the cache irrelevant data can be reduced, and the hit rate of the memory data is improved.
Based on the above embodiment, the determining, based on the main factors corresponding to the degraded cell worksheets, the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
Determining data to be accessed corresponding to each main factor based on the main factor corresponding to the degraded cell work order;
Determining a global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor in a preset time period;
the main factor corresponding to the degraded cell work order and the priority of the main factor are determined based on each sub-factor corresponding to the degraded cell work order and the influence weight thereof, wherein the influence weight is used for representing the influence degree of the sub-factor corresponding to the degraded cell work order on the cell network performance index.
Specifically, the data to be accessed corresponding to each child is stored in the database in advance in the form of a data table. As shown in table 1, three-level data constructed according to the page-data extraction URL-data table related to each URL, the table lists the data to be accessed corresponding to the coverage, neighbor, resource and parameter of the 4 sub-factors, but it is understood that the actual database is not limited to the data to be accessed corresponding to the 4 sub-factors, but includes the data to be accessed corresponding to all the sub-factors.
Table 1 example of data to be accessed for child factor
The memory scheduling device can determine the data to be accessed corresponding to each main factor on the basis of determining the main factor corresponding to the work order of the degraded cell. The main factor and the priority of the main factor corresponding to the degraded cell work order are determined based on each sub factor and the influence weight thereof, wherein the larger the influence weight is, the higher the influence degree of the degradation influence of the corresponding sub factor on the cell network performance index is, the main factor and the priority of the main factor corresponding to the degraded cell work order can be determined based on the magnitude sequence of the influence weight corresponding to each sub factor, and the larger the influence weight is, the higher the priority is. As for the number of main factors, it may be empirically determined in advance, that is, the first few sub-factors with the highest priority are taken as main factors.
In order to ensure that the data to be accessed which is transferred into the memory is the data which is actually accessed by a network optimization engineer, the global priority scheduling sequence of the data to be accessed which is corresponding to each main factor is determined based on the priority of the main factor which is corresponding to the degraded cell work order and the historical access frequency of the data to be accessed which is corresponding to each main factor in the preset time length. According to experience, the higher the priority of the main factor corresponding to the degraded cell work order is, the higher the probability that the corresponding data to be accessed is, and the priority of each data to be accessed corresponding to a certain main factor can be obtained according to the historical access frequency, based on this, the global priority scheduling sequence of each data to be accessed corresponding to the main factor can be determined by combining the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of each data to be accessed corresponding to each main factor in the preset time. As for the time length corresponding to the preset time length, the setting may be performed according to experience of a network optimization engineer, which is not particularly limited in the embodiment of the present invention.
According to the method provided by the embodiment of the invention, the main factors corresponding to the degraded cell worksheet are used for determining the data to be accessed corresponding to each main factor, and the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined based on the priority of the main factor corresponding to the degraded cell worksheet and the historical access frequency of the data to be accessed corresponding to each main factor in the preset time length, so that the hit rate of the memory data can be ensured on the basis of reducing the memory space of the cache irrelevant data.
Based on the above embodiment, the determining, based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor within the preset duration, the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
And determining the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority and the weight corresponding to the priority of the main factor corresponding to the degraded cell work order, the historical access frequency of the data to be accessed in the preset time length and the weight corresponding to the access frequency.
Specifically, corresponding weights are distributed for the main factor priority and the historical access frequency of the data to be accessed according to the historical experience, so that the accuracy of the global priority scheduling sequence of the data to be accessed corresponding to each main factor is further ensured, and the hit rate of the memory data is further ensured. Assuming that the main factors corresponding to the degraded cell worksheet obtained by the memory scheduling device are 4, the frequency of access of the corresponding data table to be accessed by the network optimization engineer in a period of time can be determined according to the main factors, and the global priority scheduling sequence of the data to be accessed corresponding to the main factors is determined according to the priority of the main factors corresponding to the degraded cell worksheet, the weight corresponding to the priority, the historical access frequency of the data to be accessed in a preset duration and the weight corresponding to the access frequency, wherein the global priority scheduling sequence of the data to be accessed is determined according to the 4 main factors of which the priority is sequentially from high to low and which are the coverage-neighbor-resource-parameter, as shown in table 2.
Table 2 global priority scheduling order representation example
The priority of each data table to be accessed is equal to the priority of C (main factor priority) and the priority of D (access frequency weight) the statistics of the access frequency of the table (C, D is adjustable, in the above table, for the purpose of illustrating the calculation process, C is 100, D is 1, and because of 4 main factors, the priority of the main factor with the highest priority is 4, and the priorities of the other main factors are sequentially reduced by 1 as span), in actual calculation, C takes the total data table number corresponding to all the sub factors, and D can be set as the total data table number divided by the total data table access frequency so as to ensure sufficient priority distinction. The larger the value corresponding to the scheduling ordering in table 2, the higher the global priority of the corresponding data to be accessed.
According to the method provided by the embodiment of the invention, the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined based on the priority of the main factor corresponding to the degraded cell work order, the weight corresponding to the priority, the historical access frequency of the data to be accessed in the preset time period and the weight corresponding to the access frequency, so that the accuracy of the global priority scheduling sequence of the data to be accessed corresponding to each main factor can be ensured, and the hit rate of the memory data is further ensured.
Based on the above embodiment, the scheduling of the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
And loading the data to be accessed into the memory in sequence according to the priority from high to low based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, and/or sequentially removing the data to be accessed from the memory according to the priority from low to high.
Specifically, assuming that the total number of sub-factors is 6, on the basis of presetting that the main factors corresponding to the degraded cell worksheet are 4, the memory scheduling device may generate the global priority scheduling sequence table of the data to be accessed corresponding to each main factor ordering scene according to the ordering condition statistics of the 4 main factors in advance, and since the global priority scheduling sequence table of the 4 main factors has 6-dimensional sub-factors, the global data priority scheduling sequence table of the 4 main factors has 6×5×4×3=360 scene directories, and each directory is identified by the main factor scene corresponding to R1-R2-R3-R4, for example, the coverage-resource-neighbor-parameter. An example of a global priority schedule sequence table corresponding to a main cause scenario is shown in table 3.
Table 3 global priority scheduling order representation example for main cause scenario
When the memory scheduling device is started, the memory management program scans all worksheets, searches a global data priority scheduling sequence table according to 4 dimensions related to each worksheet main factor and R1-R2-R3-R4 main factor scene identifiers to obtain a global data priority scheduling sequence table corresponding to the main factor scenes, and then loads the tables required by the worksheets into the memory in sequence according to the priority in the table until the memory utilization rate reaches a use threshold.
When the data accessed by the network optimization engineer is not in the memory, the data form related to the distributed work order is moved out of the memory, then the data form with lower priority and without the distributed work order is moved out of the memory, for example, the form with lower priority is moved out of the memory, or the form with lower priority is moved out of the memory, and the like, so that enough data in the memory loading hard disk is left out to enter the memory.
According to the method provided by the embodiment of the invention, the data to be accessed is sequentially loaded into the memory according to the order of the priority from high to low based on the global priority scheduling order of the data to be accessed corresponding to each main factor, and/or the data to be accessed is sequentially moved out of the memory according to the order of the priority from low to high, so that the hit rate of the memory data can be improved on the basis of reducing the memory space of the cache irrelevant data.
Based on the above embodiment, the impact weight is determined based on a preset correction coefficient and probability association information, heartbeat association information and pearson association information of a factor corresponding to the degraded cell work order and the cell network performance index, and the preset correction coefficient is determined based on a conditional probability that the factor corresponding to the history degraded cell work order causes degradation of the cell network performance index.
Specifically, the sub-factor corresponding to the degraded cell work order is denoted as R, and in view of different degradation effects of different sub-factors R on different cell network performance indexes, in an actual operation and maintenance environment, the cell network performance index serving as an evaluation reference may be predetermined, the cell network performance index may be denoted as K, and the impact weight may be used to characterize the impact degree of the degradation impact. The influence weight is determined based on a preset correction coefficient and probability association information, heartbeat association information and pearson association information of a sub-factor corresponding to the degraded cell work order and the cell network performance index, and the preset correction coefficient is determined based on a conditional probability that the main factor corresponding to the historical degraded cell work order causes the degradation of the cell network performance index. The main factor can be determined based on the sub-factor R and the corresponding influence weight. Fig. 2 is a detailed flow chart of a memory scheduling method according to an embodiment of the present invention. The determination of the main cause is described in detail below in conjunction with fig. 2:
1) Quantifying and normalizing the degradation degree of the index K
Network performance KPIs are typically used to reflect network quality, and if network quality deteriorates, the system samples the metrics in time series (e.g., at an hour granularity during degradation), and a sampled time series of each network performance metric can be obtained, and then normalized according to the following formula, and assigned a degradation level according to the degradation range after normalization:
wherein K jn is an index value in an hour period, K is a problem threshold value, K 0 is a completely problem-free value, and after normalization of a plurality of index sampling sequences, an assignment matrix of K can be obtained as
2) Quantization of the factor R:
According to an empirical formula, calculating and quantifying a sub-factor R which causes network index degradation, carrying out hour grade assignment according to severity, obtaining a degradation time assignment table of the sub-factor R, and marking whether a problem exists or not;
Elements of the quality difference dimension are classified into three categories. The first quality difference dimension presents a switching characteristic in one heartbeat period, namely, the dimension has two states of problem and no problem, and the elements in the quality difference dimension comprise alarms, parameters and neighbor cells. The second poor quality dimension exhibits tri-state characteristics including severe poor quality, weak poor quality, and no problems. The third dimension of quality difference presents one-key judging feature, namely the dimension is single feature in the degradation period, and has or has no problem.
Carrying out R dimension hour grade assignment (R dimension tristate assignment rule is that serious quality difference assignment is 4, weak quality difference assignment is 2 and no problem assignment is 0) aiming at the R dimension characteristic state;
Calculating and counting to obtain an assignment matrix of the N-hour granularity in the degradation period T;
3) And (3) screening an incidence matrix C:
by counting a large amount of work order data and expert experience, whether a statistics factor is associated with an index, namely whether the factor can cause index degradation, the association mark is 1, and the unassociated mark is 0. A matrix list C is obtained if the index R affects K as follows:
and (3) correlating the sub-factor R with the correlation coefficient matrix C, screening whether the sub-factor R is correlated with the index degradation, and eliminating the irrelevant reason R to obtain the relevant wireless reason R.
4) And (5) calculating an influence weight:
the influence weight of the ranking of the main cause is calculated according to the following formula:
the 3 items in the brackets described above correspond to probability association information, heartbeat association information, and pearson association information, respectively. The probability association information corresponds to a probability association, the heartbeat association information corresponds to a heartbeat association, and the pearson association information corresponds to a pearson association.
In the probability association method, the number of influence times of the sub-factor R corresponding to the degraded cell work order on the cell network performance index K can be counted, wherein the number of influence times is probability association information and can be recorded as x. In the actual calculation process, x is a probability constant.
For the heartbeat correlation method, the sub-factor R and the cell index K may be combined first to obtain the correlation direction.
The stronger the R element, the more serious the degradation of the K element, when the correlation direction is forward, indicating that the stronger the influence of the R element on the K element. When the correlation direction is negative, the R element is strong but the K element is not affected by it, i.e., the less likely that degradation of R results in degradation of K.
Wherein, the heartbeat association calculation formula in each hour can be expressed as follows,
Ri*Kj
Wherein K j represents a normalized index value under the hour granularity, R i represents the quality difference degree of the cell network performance index, and the values can be 0, 1, 2;i and j represent serial numbers.
Then, comprehensive weighting calculation can be performed on the time domain to obtain heartbeat association information.
Wherein, the heartbeat-related information can be recorded as,
Wherein f (K, R) represents heartbeat-related information, K represents a cell network performance index, R represents a factor, and n represents the number of hours of degradation time. The K, R associations that can be weighted can be filtered from the association matrix C.
Regarding the pearson correlation method, in view of the fact that the cell network performance index K can show different changes for the change of the sub-factor R, by using pearson correlation coefficients, the variance correlation of the cell network performance index change condition of the same time period when different sub-factors change can be explored, meanwhile, positive correlation and negative correlation are introduced, and the influence of the sub-factor change on the cell network performance index can be described, which is shown as follows:
Wherein T represents the statistics of each hour during the time degradation period, Mean values of R and K at different times are shown, respectively.
D is a correction coefficient reflecting the possibility of K index deterioration due to this factor, and the real main factor R that causes K index deterioration, which is fed back by the net optimization engineer, is counted, and the linear function f (P (r|k)) of the conditional probability calculated by the bayesian equation is used as the correction coefficient D.
5) Correction factor D adjustment
And adjusting the correction coefficient by adopting a Bayesian statistical theory. And manually feeding back the main factor deduction correctness generated by the memory scheduling device by using network optimization personnel, wherein the data of each work order participated in the adjustment of the correction coefficient D comprises a work order degradation index set K [ K 1,K2,K3..,Kn ], and the main factor sequencing R X of the manual feedback. Firstly, counting the conditional probability of a work order degradation index set K [ K 1,K2,K3..,Kn ] caused by a historical work order main factor Rx:
P (K|Rx) counts the rate of degradation of the index K caused by the main factor Rx and fed back by all historical work order network optimization engineers;
p (Rx) counts the rate of degraded worksheets of all historical worksheets, which are mainly caused by Rx;
p (K) counts the degradation rate of index K in all worksheets;
where Rx refers to the main influence factor of Rx fed back by the network optimization engineer on the degradation index set K.
The conditional probability is then calculated using a bayesian formula:
P(Rx|K)=P(K|Rx)*P(Rx)/P(K)
And obtaining Bayesian probability estimation of the work order K caused by Rx main factors. The adjusted linear function of P (Rx|K) is used as a correction coefficient D for influencing the weight value in the next calculation.
6) Main factor output:
The main factor is output according to the priority order of the influence weight (for example, the main factor of the factor 4 before the priority is taken as the main factor to be output), and it is worth noting that the interference and the alarm of the cell are directly output no matter the priority, because the interference and the alarm are the direct factors causing the degradation, and the purpose is to avoid the interference and the alarm which are caused by the back priority order and cannot be normally output.
According to the method provided by the embodiment of the invention, the influence weight is determined based on the preset correction coefficient and the probability association information, the heartbeat association information and the Pearson association information of the sub-factor corresponding to the degraded cell work order and the cell network performance index, the preset correction coefficient is determined based on the conditional probability that the main factor corresponding to the history degraded cell work order causes the degradation of the cell network performance index, the accuracy of the influence weight is ensured, the accuracy of the priority of each main factor is ensured, and the accuracy of the global priority scheduling sequence of the data to be accessed is further ensured.
Based on the foregoing embodiment, after determining the global priority scheduling order of the data to be accessed corresponding to each main factor, the method further includes:
and updating the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the access frequency of the data to be accessed corresponding to each main factor in a preset time period after the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined.
Specifically, as shown in fig. 2, each table access priority of the global data priority scheduling sequence table is counted and dynamically maintained according to the number of library table times actually accessed by the network optimization engineer during use. The memory scheduling device counts the number of data forms to be accessed actually accessed by a network optimization engineer during use, counts the frequency of the data forms actually accessed by a user in each R1-R2-R3-R4 main factor scene, updates a global data priority scheduling sequence table for use in a memory scheduling process, and can continuously improve the hit rate of memory data.
According to the method provided by the embodiment of the invention, the global priority scheduling sequence of the data to be accessed corresponding to each main factor is updated based on the access frequency of the data to be accessed corresponding to each main factor in the preset time period after the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined, so that the method can be continuously improved, and the hit rate of the memory data is improved.
Based on the above embodiment, the conditional probability that the main factor corresponding to the history degraded cell worksheet causes degradation of the cell network performance index is calculated by a bayesian formula based on the ratio of degradation of the cell network performance index caused by the main factor in all the history degraded cell worksheets, the ratio of degradation of the cell worksheets caused by the main factor in all the history degraded cell worksheets, and the ratio of degradation of the cell network performance index in all the history degraded cell worksheets.
Specifically, in the main factor calculating method, the correction coefficient D is dynamically adjusted according to feedback of a net engineer on the accuracy of a main factor judging algorithm of the memory scheduling device. The memory scheduling device records the association relation between each work order degradation index and the deduced main factor and the main factor fed back by the net optimization engineer in the operation process, records the times of the net optimization engineer accessing the database table in each R1-R2-R3-R4 main factor scene, and updates the correction coefficient D by using the information:
The probability distribution of P (K) in all historical worksheets is reckoned with daily worksheet feedback data, P (K|Rx), P (Rx). Wherein Rx refers to the influence main factor of Rx fed back by a network optimization engineer on a degradation index set K, K is the degradation index set of each work order, and then a Bayesian formula is utilized to calculate the conditional probability:
P(Rx|K)=P(K|Rx)*P(Rx)/P(K)
the adjusted P (Rx|K) linear function is used as a correction coefficient D for influencing the weight value in the next calculation.
According to the method provided by the embodiment of the invention, the correction coefficient D is dynamically adjusted by calculating through a Bayesian formula based on the ratio of degradation of the cell network performance index caused by the main factor in all the historical degradation cell worksheets, the ratio of degradation cell worksheets caused by the main factor in all the historical degradation cell worksheets and the ratio of degradation of the cell network performance index in all the historical degradation cell worksheets, so that the accuracy of main factor judgment can be continuously improved, and the memory data hit rate is further improved.
Based on any of the above embodiments, fig. 3 is a schematic diagram of a memory scheduling device according to an embodiment of the present invention, as shown in fig. 3, where the device includes:
The priority scheduling order determining module 310 is configured to determine, based on the main factors corresponding to the degraded cell worksheets, a global priority scheduling order of the data to be accessed corresponding to the main factors, where the main factors corresponding to the degraded cell worksheets refer to main causes of degradation of network performance indexes corresponding to the degraded cells.
Specifically, the network optimization engineer generally queries corresponding analysis data, i.e. data to be accessed, according to the main factor output by the centralized network optimization system as a main basis for diagnosing network problems. Based on this, the priority scheduling order determining module 310 may determine, according to the main factors corresponding to the degraded cell worksheets, a global priority scheduling order of the data to be accessed corresponding to each main factor, so as to enable the network optimization engineer to access.
The memory scheduling module 320 is configured to schedule the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor.
Specifically, the memory scheduling module 320 schedules the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor, and stores or moves the data to be accessed corresponding to each main factor into or out of the memory according to the priority scheduling order, thereby avoiding the buffering of irrelevant data and enabling the network optimization engineer to quickly access the data related to the wireless network problem.
According to the device provided by the embodiment of the invention, the global priority scheduling order of the data to be accessed corresponding to each main factor is determined based on the main factor corresponding to the degraded cell work order by the priority scheduling order determining module, and the memory scheduling module schedules the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor, so that the memory space for caching irrelevant data can be reduced, and the hit rate of the memory data is improved.
Based on the above embodiment, the determining, based on the main factors corresponding to the degraded cell worksheets, the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
Determining data to be accessed corresponding to each main factor based on the main factor corresponding to the degraded cell work order;
Determining a global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor in a preset time period;
the main factor corresponding to the degraded cell work order and the priority of the main factor are determined based on each sub-factor corresponding to the degraded cell work order and the influence weight thereof, wherein the influence weight is used for representing the influence degree of the sub-factor corresponding to the degraded cell work order on the cell network performance index.
Based on the above embodiment, the determining, based on the priority of the main factor corresponding to the degraded cell work order and the historical access frequency of the data to be accessed corresponding to each main factor within the preset duration, the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
And determining the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the priority and the weight corresponding to the priority of the main factor corresponding to the degraded cell work order, the historical access frequency of the data to be accessed in the preset time length and the weight corresponding to the access frequency.
Based on the above embodiment, the scheduling of the memory data based on the global priority scheduling order of the data to be accessed corresponding to each main factor specifically includes:
and loading the data to be accessed into the memory in sequence from high priority to low priority based on the global priority scheduling sequence of the data to be accessed corresponding to each main factor, and/or sequentially removing the data to be accessed from the memory in sequence from low priority to high priority.
Based on the above embodiment, the impact weight is determined based on a preset correction coefficient and probability association information, heartbeat association information and pearson association information of a factor corresponding to the degraded cell work order and the cell network performance index, and the preset correction coefficient is determined based on a conditional probability that the factor corresponding to the history degraded cell work order causes degradation of the cell network performance index.
Based on the foregoing embodiment, after determining the global priority scheduling order of the data to be accessed corresponding to each main factor, the method further includes:
and updating the global priority scheduling sequence of the data to be accessed corresponding to each main factor based on the access frequency of the data to be accessed corresponding to each main factor in a preset time period after the global priority scheduling sequence of the data to be accessed corresponding to each main factor is determined.
Based on the above embodiment, the conditional probability that the main factor corresponding to the history degraded cell worksheet causes degradation of the cell network performance index is calculated by a bayesian formula based on the ratio of degradation of the cell network performance index caused by the main factor in all the history degraded cell worksheets, the ratio of degradation of the cell worksheets caused by the main factor in all the history degraded cell worksheets, and the ratio of degradation of the cell network performance index in all the history degraded cell worksheets.
The memory scheduling device provided by the embodiment of the present invention can execute the above memory scheduling method, and its specific working principle and corresponding technical effects are the same as those of the above method embodiment, and are not described herein again.
Fig. 4 illustrates a physical schematic diagram of an electronic device, as shown in fig. 4, which may include a processor (processor) 410, a communication interface (Communications Interface) 420, a memory (memory) 430, and a communication bus 440, where the processor 410, the communication interface 420, and the memory 430 perform communication with each other through the communication bus 440. Processor 410 may invoke logic instructions in memory 430 to perform the flow of steps provided by the method embodiments described above.
Further, the logic instructions in the memory 430 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. The storage medium includes a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention further provides a non-transitory computer readable storage medium, on which a computer program is stored, where the computer program when executed by a processor implements the step flow provided by the above method embodiment.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
It should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention, and not for limiting the same, and although the present invention has been described in detail with reference to the above-mentioned embodiments, it should be understood by those skilled in the art that the technical solution described in the above-mentioned embodiments may be modified or some technical features may be equivalently replaced, and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solution of the embodiments of the present invention.
Claims (9)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011267888.7A CN114490012B (en) | 2020-11-13 | 2020-11-13 | Memory scheduling method, device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011267888.7A CN114490012B (en) | 2020-11-13 | 2020-11-13 | Memory scheduling method, device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114490012A CN114490012A (en) | 2022-05-13 |
| CN114490012B true CN114490012B (en) | 2025-03-21 |
Family
ID=81490182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011267888.7A Active CN114490012B (en) | 2020-11-13 | 2020-11-13 | Memory scheduling method, device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114490012B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118567869B (en) * | 2024-08-01 | 2024-10-01 | 芯方舟(上海)集成电路有限公司 | Memory access method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102098737A (en) * | 2009-12-11 | 2011-06-15 | 中兴通讯股份有限公司 | Cell priority level based cooperation scheduling method and system |
| CN108463973A (en) * | 2016-01-08 | 2018-08-28 | 华为技术有限公司 | Fingerprint recognition basic reason is analyzed in cellular system |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5408600B2 (en) * | 2008-05-28 | 2014-02-05 | インターナショナル・ビジネス・マシーンズ・コーポレーション | Determination apparatus, database apparatus, program, and determination method |
| CN109963293A (en) * | 2017-12-25 | 2019-07-02 | 中国移动通信集团上海有限公司 | A kind of performance index optimization method and device |
| CN111104169B (en) * | 2017-12-29 | 2021-01-12 | 上海寒武纪信息科技有限公司 | Instruction list scheduling method, device, computer equipment and storage medium |
| CN109993388B (en) * | 2017-12-30 | 2021-07-06 | 中国移动通信集团湖北有限公司 | Method, device, equipment and medium for judging main cause of degraded cell |
-
2020
- 2020-11-13 CN CN202011267888.7A patent/CN114490012B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102098737A (en) * | 2009-12-11 | 2011-06-15 | 中兴通讯股份有限公司 | Cell priority level based cooperation scheduling method and system |
| CN108463973A (en) * | 2016-01-08 | 2018-08-28 | 华为技术有限公司 | Fingerprint recognition basic reason is analyzed in cellular system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114490012A (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105701216B (en) | A kind of information-pushing method and device | |
| CN114219097A (en) | Federal learning training and prediction method and system based on heterogeneous resources | |
| CN112398700B (en) | Service degradation method and device, storage medium and computer equipment | |
| CN113010310B (en) | Method, device and server for processing job data | |
| CN117539648A (en) | Service quality management method and device for electronic government cloud platform | |
| CN114490012B (en) | Memory scheduling method, device and readable storage medium | |
| CN117370058A (en) | Service processing method, device, electronic equipment and computer readable medium | |
| CN118467186A (en) | Multi-tenant server-free platform resource management method and system | |
| CN116860789A (en) | Data distribution optimization method and distributed database system | |
| CN111737233B (en) | Data monitoring method and device | |
| CN119645618A (en) | Resource adjustment method, device, equipment and medium for edge computing nodes | |
| JP2008158748A (en) | Variable selection device and method, and program | |
| CN119441320A (en) | A method, system and device for quickly querying application topology | |
| CN118394497A (en) | Data processing method, device and system | |
| CN110399095A (en) | A kind of statistical method and device of memory space | |
| CN111245649A (en) | News click rate prediction algorithm | |
| CN112800089B (en) | Intermediate data storage level adjusting method, storage medium and computer equipment | |
| CN116149848A (en) | Load prediction method and device, electronic equipment and storage medium | |
| CN113204430B (en) | Resource scheduling method, scheduling management platform, equipment and medium | |
| CN119094603B (en) | A low-latency data push system and method for edge terminals | |
| CN118410189B (en) | Method, device, equipment and storage medium for managing multimedia materials | |
| CN111178610A (en) | Multi-layer convolution and GRU (generalized regression Unit) -based capacity prediction method and device | |
| CN113626491B (en) | Data query method, device and distributed data query system | |
| CN115879744B (en) | Limited production scheduling method, device, storage medium and electronic equipment | |
| CN116382925B (en) | Dynamic adjustment method and device for task queue and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |