CN116886619A - Load balancing method and device based on linear regression algorithm - Google Patents
Load balancing method and device based on linear regression algorithm Download PDFInfo
- Publication number
- CN116886619A CN116886619A CN202310989595.7A CN202310989595A CN116886619A CN 116886619 A CN116886619 A CN 116886619A CN 202310989595 A CN202310989595 A CN 202310989595A CN 116886619 A CN116886619 A CN 116886619A
- Authority
- CN
- China
- Prior art keywords
- data
- server
- weight
- load
- instance
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000012417 linear regression Methods 0.000 title claims abstract description 35
- 230000004044 response Effects 0.000 claims abstract description 19
- 238000013480 data collection Methods 0.000 claims abstract description 10
- 238000004891 communication Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 10
- 230000001419 dependent effect Effects 0.000 claims description 2
- 238000011478 gradient descent method Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims 1
- 230000003068 static effect Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003698 anagen phase Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
- H04L63/1425—Traffic logging, e.g. anomaly detection
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The application relates to a load balancing method and device based on a linear regression algorithm, comprising the following steps: a server load data acquisition module is deployed on each instance server, the data acquisition module acquires load data and service call log data of the server at regular time, and the load data is transmitted to a data acquisition module through a network communication means; the data collection module sorts the collected load data, stores the load data in a specified directory in the form of text logs, and each line of records respectively comprises a server number, a CPU (central processing unit) utilization rate, a memory utilization rate, a request starting time stamp and interface response time, wherein the data are separated by commas; the weight prediction module is responsible for carrying out statistics and arrangement on the collected data, carrying out load weight prediction by using a linear regression algorithm, and transmitting the calculated weight value of each instance server to the strategy forwarding module; and the strategy forwarding module forwards the service request to the back-end instance server according to the load weight.
Description
Technical Field
The application relates to the technical field of networked control, in particular to a load balancing method and device based on a linear regression algorithm.
Background
With the rapid development of the internet, network applications have involved aspects of daily life of people, network traffic increases exponentially, and a traditional single Web server architecture is difficult to cope with high concurrent requests in the current multi-network environment, so that server load pressure continuously increases. At the same time, with the development of JavaScript, CSS, JSON and other multimedia technologies, today's internet data transmission types are no longer simple text data, and picture and video data are in a fast growing phase, and the percentage of total network traffic increases rapidly year by year, resulting in a backbone network bandwidth strain and an increase in user access delay.
Load balancing refers to the fact that traffic is distributed to a plurality of operation units through a specific algorithm for execution, and provides a cheap, effective and transparent method for expanding the bandwidth of network equipment and servers, increasing throughput, enhancing network data processing capacity, improving flexibility and usability of networks, and with the rise of high-concurrency system architecture, especially WeChat applets, more and more application programs need to split services into a plurality of independent sub-services so as to improve the scalability and usability of the application programs.
The conventional load balancing algorithm is mainly divided into static and dynamic two types, however, when the load balancing under the high concurrency system architecture is processed, some problems exist, such as the dynamic allocation of resources according to actual demands, the inadaptation to the demands of the high concurrency system architecture, etc., the load balancing technology mainly depends on hardware resources of a virtual machine or a server, the dynamic allocation of resources according to actual demands is not possible, and thus the inadaptation to the demands of the high concurrency system architecture is not possible. Dynamic load balancing and static load balancing have the advantages and the disadvantages of the static algorithm, and the static algorithm is stable and quick, but is excessively dead, so that reasonable adjustment can not be performed according to the load condition of the server, and the maximum utilization of the server resources is not facilitated; although the dynamic load balancing algorithm makes up the defects of the static algorithm to a certain extent, only parameters such as the connection number are concerned, and the parameters cannot directly reflect the load condition of the server, so that the load response of the back-end server is ignored, the real performances of all servers cannot be known, and the load distribution is often unreasonable.
Disclosure of Invention
The application aims to overcome the defects of the prior art and provides a load balancing method and device based on a linear regression algorithm.
In order to achieve the above purpose, the technical scheme of the application is realized as follows:
in a first aspect, the present application provides a load balancing method based on a linear regression algorithm, including the steps of:
s1, deploying a server load data acquisition module on each instance server, and enabling the data acquisition module to acquire load data and service call log data of the server at fixed time and transmitting the load data to a data acquisition module through a network communication means.
S2, the data collection module sorts the collected load data, stores the load data in a specified directory in the form of text logs, and each row of records respectively comprises a server number, a CPU (Central processing Unit) utilization rate, a memory utilization rate, a request starting time stamp and an interface response time, and the data are separated by commas.
And S3, the weight prediction module is responsible for carrying out statistics and arrangement on the collected data, carrying out load weight prediction by using a linear regression algorithm, and transmitting the calculated weight value of each instance server to the strategy forwarding module.
And S4, the strategy forwarding module forwards the service request to the back-end instance server according to the load weight.
In a second aspect, the present application provides a load balancing device based on a linear regression algorithm, including:
and a data acquisition module: the system is responsible for regularly acquiring load data and service call log data of a server, and transmitting the load data to a data collection module through a network communication means;
and a data collection module: the method is in charge of sorting the collected server data, storing the collected server data in a log form, wherein each row of records respectively comprises a server number, CPU utilization rate, memory utilization rate, request starting time stamp and interface response time;
and a weight prediction module: the method is responsible for carrying out statistics arrangement on the collected data and carrying out load weight prediction by using a linear regression algorithm;
and a strategy forwarding module: is responsible for forwarding service requests to the backend server according to the load weights.
The application has the following beneficial effects: according to the intelligent dynamic load balancing method based on the linear regression algorithm, a traditional load balancing strategy is combined with a machine learning algorithm, and the load condition of a server is predicted, so that the load balancing parameters are reasonably and effectively automatically configured, the load balancing performance is improved, and the performance of an instance server is fully pressed; and the load balancing strategy is reconstructed by using a linear regression algorithm to predict and adjust the weight in quasi-real time. The configuration of load balancing greatly reduces human intervention, the parameters are more reasonable, the maximum performance of the back-end server is fully mined, and the throughput of the whole system is improved. The calculation process is mature and stable, is triggered at fixed time, and has low consumption on server resources.
Drawings
Fig. 1 is a flow chart of a load balancing method based on a linear regression algorithm of the present application.
Fig. 2 is a flowchart of step S3 in the load balancing method based on the linear regression algorithm of the present application.
Fig. 3 is a flowchart of step S4 in the load balancing method based on the linear regression algorithm of the present application.
Fig. 4 is a schematic diagram of data transfer in accordance with the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
It will be understood by those skilled in the art that all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs unless defined otherwise. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Figures 1-4 schematically illustrate a flow chart of a load balancing method based on a linear regression algorithm according to one embodiment of the application. Referring to fig. 1, the load balancing method based on the linear regression algorithm provided by the embodiment of the application comprises the following steps:
s1, deploying a server load data acquisition module on each instance server, wherein the data acquisition module acquires load data and service call log data of the server at regular time, and transmits the load data to the data acquisition module through a network communication means;
s2, the data collection module sorts the collected load data, the load data is stored in a specified directory in the form of text logs, each row of records respectively comprises a server number, CPU (Central processing Unit) utilization rate, memory utilization rate, request starting time stamp and interface response time, and the data are separated by commas, for example: "Sever1,0.82,0.70,1681880513,250" means that the response time of a request on the server1 host is 250ms, the CPU usage is 82% and the memory usage is 70%. Each data file record is smaller than 8000, and when the maximum value is reached, the current file is closed and a new file is created.
And S3, the weight prediction module is responsible for carrying out statistics and arrangement on the collected data, carrying out load weight prediction by using a linear regression algorithm, and transmitting the calculated weight value of each instance server to the strategy forwarding module. As shown in fig. 2, the method specifically comprises the following steps:
s31, initializing a timer, and triggering the timer in an initial sampling period T.
And S32, when the timer reaches the initial sampling period, scanning and analyzing log data of the instance server, and predicting response time by using a linear regression model.
S321, classifying the sample set data according to server numbers, and then calculating the time spent by each time t request.
S322, sorting sample data, calculating the response time of each server from t1 to tn at each time point, and predicting by adopting a typical linear regression model, wherein a training model y=wx+b, x epsilon t (1 is more than or equal to t is less than or equal to n) is a model independent variable, y is a model dependent variable, w is a weight value, and b is a bias value. The values of w and b need now be obtained through training of the sample data. To find the best straight line to fit the data, a cost function is typically introduced as a measure, whose mathematical expression is:
the Loss function Loss is the predicted value of n samples (wx i + b) and the actual value yi. Since the straight line with the smallest Euclidean distance is the best-fit straight line, the problem of obtaining the values of w and b is converted into the problem of minimizing the Loss function Loss when the values of w and b are taken. Since the loss function is a convex function, the gradient descent method is adopted for solving, and the solving steps are as follows:
s3221, loading a training data set and setting algorithm operation super parameters, wherein the setting is as follows:
training dataset x= { x1, x2, xi, …, xn };
the target value y= { y1, y2, yi, …, yn } corresponding to x;
weight w=random number (e.g. 0.1);
offset b=random number (e.g. 0.1);
learning rate η=0.00002 (empirical value);
iteration number is=80 (empirical value).
S3222, perform iteration: the gradient of the weight value w and the bias value b of the ith training data with respect to the loss function is calculated, referring to the following formula.
And calculating to obtain gradient, and updating the weight value and the offset value of each sample according to the following formula.
Substituting the obtained w and b into the following formula, and calculating the predicted value of each sample by using a set operation method.
pred=w*x+b
And calculating a loss function, recording w and b when the loss function value meets the precision requirement or reaches the maximum iteration number, and exiting the loop.
S3223, the algorithm is exited and w and b are returned.
S323, substituting the determined w and b into y=wx+b to calculate the response time of tn+1 to the instance server in the next sampling period.
S33, calculating the weight value of each server according to the response time predicted in the step S32.
The shorter the calculated response time, the more load-bearing the instance server is, and the greater the weight is. The calculation weight value of each instance server can be calculated by adopting a maximum common multiple method, the calculation mode refers to the following formula, and the result is reserved in a decimal.
S34, transmitting the calculated weight value of each instance server to a strategy forwarding module.
S35, calculating the next sampling trigger period.
Calculated weight jw i And the weight value jw of the previous period i The smaller the' absolute difference, the more accurate the prediction, the longer the predicted time from the next time, the duration of the next sampling trigger period is calculated using the following formula.
And S4, the strategy forwarding module forwards the service request to the back-end instance server according to the load weight. As shown in fig. 3, the method specifically comprises the following steps:
s41, initializing a strategy forwarding module variable, wherein the initial weight value can be simply configured according to the hardware configuration condition because the load capacity of each instance server at the back end cannot be accurately known, and the instance with the lowest hardware configuration is set as a reference standard to be 1.
There are N servers s= { S 1 ,S 2 ,S i ,...,S n Now set the following variables:
initial weight values of each server are configured according to hardware: w= { W 1 ,W 2 ,W i ,...,W n };
Effective weight: CW= { CW 1 ,CW 2 ,CW i ,...,CW n }。
In addition to the initial weight value Wi, the server Si has a current effective weight CWi, and the CWi is initialized to Wi, i.e., cwi=wi.
An indicator variable currentpos= -1, indicating that the currently selected instance id is initialized to-1.
Initial weights and: ws=w 1 +W 2 +W 3 +...+W n .
S42, selecting the instance server with the largest current effective weight, subtracting the initial weights and ws of all the instance servers from the current effective weight CWi, and pointing the variable currentPos to the position.
i=MAX(CW)
CW={CW i -ws}
currentPos=index(i)
S43, adding an initial weight Wi to the current effective weight CWi of each instance server Si;
CWi=CWi+Wi。
s44, forwarding the service request to the instance server pointed by the variable currentPos.
S45, judging whether a forwarding condition is met, namely whether the current effective weight CW is {0, …,0}, and whether (2 n-1) effective forwarding is carried out, wherein n is the number of back-end instance servers.
S451, if the forwarding condition is met, whether the weight value transmitted by the weight prediction module exists is further checked. If so, replacing the corresponding initial weight value with the calculated weight value of each instance server, and repeating the steps S41-S44, otherwise, continuing to execute;
s452, if the forwarding condition is not satisfied, the execution is continued.
S46, repeating the steps S42-S45 each time the timer reaches the sampling period.
Another embodiment discloses a device corresponding to the above disclosed load balancing method based on a linear regression algorithm, which is a virtual device structure corresponding to the method, and includes:
and a data acquisition module: the system is responsible for regularly acquiring load data and service call log data of a server, and transmitting the load data to a data collection module through a network communication means;
and a data collection module: the method is in charge of sorting the collected server data, storing the collected server data in a log form, wherein each row of records respectively comprises a server number, CPU utilization rate, memory utilization rate, request starting time stamp and interface response time;
and a weight prediction module: the method is responsible for carrying out statistics arrangement on the collected data and carrying out load weight prediction by using a linear regression algorithm;
and a strategy forwarding module: is responsible for forwarding service requests to the backend server according to the load weights.
From the above description of embodiments, it will be clear to a person skilled in the art that the present application may be implemented by means of software and necessary general purpose hardware, but of course also by means of hardware, although in many cases the former is a preferred embodiment. It is understood that the technical solution of the present application may be embodied in essence or a part contributing to the prior art in the form of a software product, and the computer software product may be stored in a computer readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a FLASH Memory (FLASH), a hard disk, or an optical disk of a computer, where the instructions include several instructions for causing an electronic device (which may be a mobile phone, a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments of the present application.
It should be noted that, in the embodiment of the load balancing device based on the linear regression algorithm, each unit and module included are only divided according to the functional logic, but not limited to the above division, so long as the corresponding function can be realized; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present application.
Those skilled in the art will appreciate that while some embodiments herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the application and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims (9)
1. The load balancing method based on the linear regression algorithm is characterized by comprising the following steps of:
s1, deploying a load data acquisition module on each instance server, wherein the data acquisition module acquires load data and service call log data of the servers at regular time, and transmits the load data to the data acquisition module through a network communication means;
s2, the data collection module sorts the collected load data, stores the load data in a specified directory in the form of text logs, and each row of records respectively comprises a server number, a CPU (central processing unit) utilization rate, a memory utilization rate, a request starting time stamp and interface response time, wherein the data are separated by commas;
s3, the weight prediction module is responsible for carrying out statistics and arrangement on the collected data, carrying out load weight prediction by using a linear regression algorithm, and transmitting the calculated weight value of each instance server to the strategy forwarding module;
and S4, the strategy forwarding module forwards the service request to the back-end instance server according to the load weight.
2. The load balancing method based on the linear regression algorithm according to claim 1, wherein: in step S3, the specific scheduling steps of the weight prediction module are as follows:
s31, initializing a timer, triggering the timer with a fixed sampling period, and assuming that the triggering period is T;
s32, when the timer reaches a sampling period, scanning and analyzing log data of the instance server, so as to predict response time of the instance server;
s33, adjusting the weight value of each server according to the response time predicted in the step S32;
s34, transmitting the calculated weight value of each instance server to a strategy forwarding module;
s35, calculating the next trigger time.
3. The load balancing method based on the linear regression algorithm according to claim 2, wherein: in step S32, the specific steps of predicting the response time are as follows:
s321, classifying sample set data according to server numbers, and then calculating the time spent by each time t request;
s322, sorting sample data, calculating the response time of each server from t1 to tn at each time point, and predicting by adopting a typical linear regression model, wherein a training model y=wx+b, x epsilon t (1 is more than or equal to t is less than or equal to n) x is a model independent variable, y is a model dependent variable, w is a weight, and b is a bias;
the values of w and b are obtained through training of sample data, a cost function is introduced as a measurement standard, and the mathematical expression is as follows:
the Loss function Loss is the predicted value of n samples (wx i +b) and the actual value y i A Euclidean distance between them; the straight line with the smallest Euclidean distance is the best fitting straight line, so when the Loss function Loss is smallest, the w value and the b value are obtained;
s323, substituting the determined w and b into y=wx+b to calculate the response time of tn+1 to the instance server in the next sampling period.
4. A load balancing method based on a linear regression algorithm according to claim 3, characterized in that: the loss function is a convex function and is solved by adopting a gradient descent method.
5. The load balancing method based on the linear regression algorithm according to claim 4, wherein: the specific solving steps are as follows:
s3221, loading a training data set and setting algorithm operation super parameters, wherein the setting is as follows:
training dataset x= { x1, x2, xi, …, xn };
the target value y= { y1, y2, yi, …, yn } corresponding to x;
weight w=random number;
bias value b = random number;
learning rate η=0.00002;
iteration number is=80;
s3222, perform iteration: the gradient of the weight w and bias b of the ith training data with respect to the loss function is calculated using the following formula:
the gradient was calculated, and the weight and bias values for each sample were updated using the following formula:
substituting the obtained w and b into the following formula, and calculating the predicted value of each sample by using a set operation method:
pred=w*x+b
calculating a loss function, recording w and b when the loss function value meets the precision requirement or reaches the maximum iteration number, and exiting the loop;
s3223, the algorithm is exited and w and b are returned.
6. The load balancing method based on the linear regression algorithm according to claim 2, wherein: in step S33, the weight of each instance server is calculated by the max common multiple method in the following manner, and the result retains a decimal fraction:
jw i to calculate the weights.
7. The load balancing method based on the linear regression algorithm according to claim 6, wherein: in step S35, the duration of the next trigger period is calculated using the following formula:
8. the load balancing method based on the linear regression algorithm according to claim 1, wherein: in step S4, the specific scheduling steps of the policy forwarding module are as follows:
s41, initializing a strategy forwarding module variable, configuring an initial weight value according to a hardware configuration condition, and setting 1 by taking an instance with the lowest hardware configuration as a reference standard;
there are N instance servers s= { S 1 ,S 2 ,S i ,...,S n Now set the following variables:
configuring each instance according to hardwareInitial weight value of server: w= { W 1 ,W 2 ,W i ,...,W n };
Effective weight: CW= { CW 1 ,CW 2 ,CW i ,...,CW n };
The above is that each instance server Si has a current valid weight CWi in addition to an initial weight value Wi, and let CWi be initialized to Wi, i.e. cwi=wi;
an indicator variable currentpos= -1, indicating that the currently selected instance id is initialized to-1;
initial weights and: ws=w 1 +W 2 +W 3 +…+W n ;
S42, selecting an instance server with the maximum current effective weight, subtracting initial weights and ws of all instance servers from the current effective weight CWi, and pointing the variable currentPos to the position;
i=MAX(CW)
CW={CW i -ws}
currentPos=index(i)
s43, adding an initial weight Wi to the current effective weight CWi of each instance server Si;
CW i =CW i +W i ;
s44, forwarding a service request to an instance server pointed by the variable currentPos;
s45, judging whether a forwarding condition is met, namely whether the current effective weight CW is {0, …,0}, and whether (2 n-1) effective forwarding is carried out, wherein n is the number of back-end instance servers;
s451, if the forwarding condition is met, further checking whether the calculated weight value transmitted by the weight prediction module exists, if so, replacing the corresponding initial weight value with the calculated weight value of each instance server, and repeating the steps S41-S44, otherwise, continuing;
s452, if the forwarding condition is not met, continuing to execute;
s46, repeating the steps S42-S45 each time the timer reaches the sampling period.
9. A load balancing device based on a linear regression algorithm, comprising:
and a data acquisition module: the system is responsible for regularly acquiring load data and service call log data of a server, and transmitting the load data to a data collection module through a network communication means;
and a data collection module: the method is in charge of sorting the collected server data, storing the collected server data in a log form, wherein each row of records respectively comprises a server number, CPU utilization rate, memory utilization rate, request starting time stamp and interface response time;
and a weight prediction module: the method is responsible for carrying out statistics arrangement on the collected data and carrying out load weight prediction by using a linear regression algorithm;
and a strategy forwarding module: is responsible for forwarding service requests to the backend server according to the load weights.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310989595.7A CN116886619A (en) | 2023-08-08 | 2023-08-08 | Load balancing method and device based on linear regression algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310989595.7A CN116886619A (en) | 2023-08-08 | 2023-08-08 | Load balancing method and device based on linear regression algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116886619A true CN116886619A (en) | 2023-10-13 |
Family
ID=88268174
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310989595.7A Pending CN116886619A (en) | 2023-08-08 | 2023-08-08 | Load balancing method and device based on linear regression algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116886619A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117130870A (en) * | 2023-10-26 | 2023-11-28 | 成都乐超人科技有限公司 | Transparent request tracking and sampling method and device for Java architecture micro-service system |
CN117149099A (en) * | 2023-10-31 | 2023-12-01 | 江苏华鲲振宇智能科技有限责任公司 | Calculation and storage split server system and control method |
-
2023
- 2023-08-08 CN CN202310989595.7A patent/CN116886619A/en active Pending
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117130870A (en) * | 2023-10-26 | 2023-11-28 | 成都乐超人科技有限公司 | Transparent request tracking and sampling method and device for Java architecture micro-service system |
CN117130870B (en) * | 2023-10-26 | 2024-01-26 | 成都乐超人科技有限公司 | Transparent request tracking and sampling method and device for Java architecture micro-service system |
CN117149099A (en) * | 2023-10-31 | 2023-12-01 | 江苏华鲲振宇智能科技有限责任公司 | Calculation and storage split server system and control method |
CN117149099B (en) * | 2023-10-31 | 2024-03-12 | 江苏华鲲振宇智能科技有限责任公司 | Calculation and storage split server system and control method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116886619A (en) | Load balancing method and device based on linear regression algorithm | |
CN107592345B (en) | Transaction current limiting device, method and transaction system | |
US11341097B2 (en) | Prefetching based on historical use and real-time signals | |
WO2017152734A1 (en) | Data processing method and relevant devices and systems | |
CN113938488A (en) | Load balancing method based on dynamic and static weighted polling | |
CN111629052B (en) | Content caching method, node, equipment and storage medium based on MEC | |
Li et al. | Method of resource estimation based on QoS in edge computing | |
CN115118602B (en) | Container resource dynamic scheduling method and system based on usage prediction | |
Lai et al. | Oort: Informed participant selection for scalable federated learning | |
CN108647329A (en) | Processing method, device and the computer readable storage medium of user behavior data | |
CN112398700B (en) | Service degradation method and device, storage medium and computer equipment | |
Xu et al. | Modeling buffer starvations of video streaming in cellular networks with large-scale measurement of user behavior | |
Yan et al. | RL-Bélády: A unified learning framework for content caching | |
CN113705981A (en) | Big data based anomaly monitoring method and device | |
CN113902116A (en) | Deep learning model-oriented reasoning batch processing optimization method and system | |
Zhang et al. | Optimizing video caching at the edge: A hybrid multi-point process approach | |
Amour et al. | Q2ABR: QoE‐aware adaptive video bit rate solution | |
US20220067569A1 (en) | Data Management Forecasting from Distributed Tracing | |
Zhou et al. | Cushion: A proactive resource provisioning method to mitigate SLO violations for containerized microservices | |
Aguilar-Armijo et al. | ECAS-ML: Edge computing assisted adaptation scheme with machine learning for HTTP adaptive streaming | |
Al Muktadir et al. | Prediction and dynamic adjustment of resources for latency-sensitive virtual network functions | |
CN112003900B (en) | Method and system for realizing high service availability under high-load scene in distributed system | |
CN114297099A (en) | Data cache optimization method and device, nonvolatile storage medium and electronic equipment | |
Cong et al. | Impact of prediction uncertainty of popularity distribution on proactive caching | |
Jehangiri et al. | Distributed predictive performance anomaly detection for virtualised platforms |
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 |