CN111930511A - Identifier resolution node load balancing device based on machine learning - Google Patents
Identifier resolution node load balancing device based on machine learning Download PDFInfo
- Publication number
- CN111930511A CN111930511A CN202010854118.6A CN202010854118A CN111930511A CN 111930511 A CN111930511 A CN 111930511A CN 202010854118 A CN202010854118 A CN 202010854118A CN 111930511 A CN111930511 A CN 111930511A
- Authority
- CN
- China
- Prior art keywords
- server
- load
- weight
- load balancing
- node
- 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
- 238000010801 machine learning Methods 0.000 title claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 230000015654 memory Effects 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000000034 method Methods 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 2
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 claims description 2
- 125000004122 cyclic group Chemical group 0.000 claims description 2
- 230000000694 effects Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 12
- 238000013461 design Methods 0.000 description 7
- 238000002474 experimental method Methods 0.000 description 4
- 238000007726 management method Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000007620 mathematical function Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- BULVZWIRKLYCBC-UHFFFAOYSA-N phorate Chemical compound CCOP(=S)(OCC)SCSCC BULVZWIRKLYCBC-UHFFFAOYSA-N 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000013068 supply chain management Methods 0.000 description 1
Images
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]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses an identification analysis node load balancing device based on machine learning, which comprises: the data acquisition module is used for acquiring the performance index and the load information of the server; the load balancing module is used for calculating the weight of the server by using the request connection data obtained by the prediction model and the data obtained by the data acquisition module and selecting the optimal node; the traffic scheduling module is used for performing traffic scheduling according to the optimal node obtained by prediction; and the cluster management module is used for managing the servers in the cluster and providing an information acquisition interface for the data acquisition module. By adopting the technical scheme of the invention, the pressure of mass requests on a secondary node system is solved, and the time for processing the requests by the load balancer is reduced.
Description
Technical Field
The invention belongs to the technical field of machine learning, and particularly relates to an identification analysis node load balancing device based on machine learning.
Background
The industrial internet identification analysis system is an important component of an industrial internet network system and is a neural hub for supporting interconnection and intercommunication of the industrial internet. The industrial internet identification analysis system structure comprises an international root node, a national top level node, a secondary node and an enterprise node from top to bottom, the nodes are interconnected and intercommunicated, the data interaction barrier in the industry and between the industries is broken, and a huge industrial ecosystem is formed. With the rapid development of the identification analysis technology, typical applications such as supply chain management, life cycle management and the like are developed for products by many industry enterprises through the connection of secondary nodes and the utilization of the identification analysis technology, and the registration amount and the analysis amount of the identification also reach a massive level, so that the effective processing of high-concurrency identification service requests by a secondary node system becomes a key problem to be solved urgently.
In the past research, the mainly adopted load balancing scheme can be divided into static load balancing and dynamic load balancing, the static load balancing algorithm is to allocate requests in a certain proportion by only utilizing some statistical values or mathematical functions without considering the real-time load state of a server, and mainly comprises a polling scheduling algorithm, a weighted polling scheduling algorithm, an address hash scheduling algorithm and a consistent hash algorithm. The dynamic load balancing algorithm mainly means that after a cluster runs for a long time, the load states of the cluster generate large differences, the load states of the server need to be evaluated by using real-time running information of the server while request distribution is carried out, and the request distribution is carried out according to the load states. At present, the existing load balancing mechanism for label resolution secondary nodes is mainly based on a dynamic load balancing algorithm to regularly monitor the CPU utilization rate, the network card bandwidth utilization rate and the memory utilization rate of a server, calculate the server weight by using the above information and the server performance indexes (CPU, network card bandwidth and memory size), and forward the request to the server with the largest weight.
With the long-time operation of the server, the load state of the server can generate large difference. The current load balancing mechanism of the identification analysis secondary node calculates corresponding real-time weight by using the running state of the server and forwards the request to the most appropriate server. But the ability of the server to process requests is not only related to its performance and load status, but also to the request traffic. In the above mechanism, if a large request traffic exists in the request processing cache queue of one server or a large-scale request burst traffic occurs at the next time, load imbalance of the cluster still occurs and a large concurrent request pressure is brought to the cluster.
Disclosure of Invention
In view of the above problems, the present invention provides a perceptual dynamic load balancing apparatus, which first establishes a Long-Term prediction model of server connection number based on a Long Short-Term Memory network (LSTM), and can predict and obtain the requested connection number of the server at the next time and reflect the traffic at the next time through the model. And then, on the basis, based on a weighted minimum connection scheduling algorithm, a server with the maximum weight is obtained by utilizing the performance index and real-time load information of the cluster for processing the request.
A device for balancing load of an identification resolution node based on machine learning comprises:
the data acquisition module is used for acquiring the performance index and the load information of the server;
the load balancing module is used for calculating the weight of the server by using the request connection data obtained by the prediction model and the data obtained by the data acquisition module and selecting the optimal node;
the traffic scheduling module is used for performing traffic scheduling according to the optimal node obtained by prediction;
and the cluster management module is used for managing the servers in the cluster and providing an information acquisition interface for the data acquisition module.
Drawings
FIG. 1 illustrates a two-level node network topology;
FIG. 2 identifies a parsing secondary node load balancing device framework;
FIG. 3 illustrates a timing diagram of a system for analyzing the load balancing of secondary nodes;
FIG. 4 identifies a parsing secondary node environment deployment diagram.
Detailed Description
The industrial internet secondary nodes are public nodes for providing identification services for specific industries or multiple industries, as shown in fig. 1, the secondary node service deployment adopts a distributed deployment mode, different nodes of the same industry can be deployed in different regions, the secondary nodes of different industries are interconnected with each other through the internet, and resources between the secondary nodes are shared to form an identification information network. Inside the industry secondary node, nodes in different areas deploy a plurality of identification analysis proxy services, data among the proxy services are mutually copied, and resources are shared. In the face of mass identification requests, the invention provides a load balancing system device, which is used for uniformly distributing access flow to a plurality of servers to execute tasks, so that the response time of the system to the service provided by a user can be shortened, the load among the servers can be balanced, and the aim of improving the performance of the whole server cluster can be fulfilled.
According to the network topology, a load balancing system is deployed in each secondary node in a plurality of industries. The load balancing device can be divided into two parts, as shown in fig. 2, a data acquisition module of the first part is responsible for acquiring performance indexes and load information of the server, and a load balancing module is responsible for calculating the weight of the server and selecting an optimal node by using the request connection data obtained by the prediction model and the data obtained by the data acquisition module. The traffic scheduling module of the second part is mainly used for performing traffic scheduling according to the predicted optimal node to achieve the purpose of load balancing, and the cluster management module is mainly used for managing the servers in the cluster and providing an information acquisition interface for the data acquisition module.
The dynamic load balancing algorithm provided by the invention is mainly integrated in a load balancing algorithm module, the algorithm calculates the weight of the server by using the performance index, the load state and the request connection number of the server, and then the server with the largest weight is selected as the best node, and the specific design thought is as follows.
Order SiRepresents the ith server in the cluster, and Cd, Md and Bd respectively representThe CPU, the memory capacity and the network bandwidth of the server, Cu, Mu and Bu respectively represent the CPU utilization rate, the memory utilization rate and the network bandwidth utilization rate of the server, and Conn represents the request connection number of the server. Let the weight of the ith server be W (S)i) The greater the weight, the greater the likelihood of allocating a request. The weight is in direct proportion to the performance of the server, the higher the obtained performance index is, the more requests can be processed, the performance index of the server, namely the performance index P (S) of the ith server is calculated by adopting the CPU capable of reflecting the performance of the server, the bandwidth of a network card and the size of a memoryi) Comprises the following steps:
as the server runs for a long time, the residual performance of the server is less and less, the index is limited, and therefore, the load index L (S) of the server is introducedi) The invention adopts the CPU utilization rate, the network card bandwidth utilization rate and the internal utilization rate which can reflect the performance of the server to calculate the load state of the server, namely the load state L (S) of the ith serveri) Comprises the following steps:
L(Si)＝αCui+βMui+γBui
the long-term and short-term memory network is a time cycle network and is specially designed for solving the long-term dependence problem of the general cyclic neural network. The invention adopts a long-short term memory network to construct a request connection number prediction model, the input dimension of which is N and represents the historical request connection number of a cluster server, the output dimension of which is 1 and represents the request connection number q at the next momentt+1. The task data are normalized and then input into a neural network, an Adaptive Moment Estimation (Adam) optimizer is used for optimization training, a Mean Square Error (MSE) function is used as a loss function, and the training effect of the model can be expressed through the function. Reduce the loss through multiple trainingTo a certain extent, a suitable prediction model is obtained. Through the model, the connection number q of the server at the next moment can be predictedt+1The present invention uses this index to predict the best node. In addition, the weight of the server is not only connected with the number q of the next timet+1Related to the number of connections q at the current timetThere is a great relationship, and therefore the requested connection number Rank (S) of the ith serveri) Comprises the following steps:
Rank(Si)＝αqt(Si)+βqt+1(Si)
different servers have different capabilities and processing the same number of requests may occupy only a small portion of the resources for a high-capability server, but may already occupy a large portion of the resources for a low-capability server, which is also a load imbalance, thus making the weight inversely proportional to the number of requested connections, the higher the number of requested connections, the lower the weight. In summary, the weight W (S) of the ith serveri) Comprises the following steps:
according to the above description of the load balancing apparatus, the present invention designs the execution timing sequence between the modules of the system, as shown in fig. 3.
The specific process is as follows,
(1) the user initiates an identification service request.
(2) The information collection module regularly collects the CPU utilization rate Cd, the memory utilization rate Md and the network card bandwidth utilization rate Bd of the servers in the cluster.
(3) The information collection module stores the collected information into a database.
(4) And the load balancing algorithm module reads the information in the database and calculates the weight of each server in the cluster.
(5) And the load balancing algorithm module predicts the optimal node according to the weight and stores the information of the optimal node into the database.
(6) The load balancer reads the information of the best node.
(7) The load balancer forwards the request to the best node for processing.
(8) The node returns the request result of the user.
The invention designs a simple deployment mode aiming at a load balancing system of an identification analysis secondary node, as shown in figure 4. Nginx is a high-performance Web proxy server, and because the code of the Nginx is open-source, the Nginx is convenient for the independent development and use of the experiment, and is used as a load balancing server by the experiment. The Redis database is a key-value storage system, the storage speed is high, the performance is high, the running efficiency of the experiment can be greatly improved by using the Redis database, and therefore the Redis database is selected for the database of the experiment.
The invention has the beneficial effects that:
1. the invention designs a dynamic load balancing method for an identification analysis secondary node system, and solves the problem of pressure brought to the secondary node system by mass requests.
2. The invention designs a deployment mode of a dynamic load balancer in a cluster for an identification resolution secondary node system.
3. The invention designs a time sequence flow for processing the request by the load balancer facing to the identification analysis secondary node system, and the flow enables the load balancing algorithm module and the flow forwarding module to run concurrently, takes the database as an intermediate cache, and reduces the time for processing the request by the load balancer.
4. The invention designs a weight calculation method based on a weighted minimum connection scheduling algorithm, takes the performance index, the real-time load state and the number of requested connections of the server as evaluation parameters of the weight, and considers the operation state of the server and the burstiness of the flow.
5. The number of the requested connections of the server can be divided into two parts, wherein the first part is the number of the requested connections at the current moment and is acquired by the information acquisition module, and the second part is the number of the requested connections at the next moment and is obtained by prediction of an LSTM prediction model.
Claims (3)
1. An identifier resolution node load balancing device based on machine learning is characterized by comprising:
the data acquisition module is used for acquiring the performance index and the load information of the server;
the load balancing module is used for calculating the weight of the server by using the request connection data obtained by the prediction model and the data obtained by the data acquisition module and selecting the optimal node;
the traffic scheduling module is used for performing traffic scheduling according to the optimal node obtained by prediction;
and the cluster management module is used for managing the servers in the cluster and providing an information acquisition interface for the data acquisition module.
2. The machine learning based identity resolution node load balancing apparatus of claim 1,
in the load balancing module, the weight of the server is obtained by calculation according to the performance index, the load state and the request connection number of the server, and then the server with the largest weight is selected as the best node, and the specific process is as follows:
order SiRepresenting the ith server in the cluster, wherein Cd, Md and Bd respectively represent the CPU, memory capacity and network bandwidth of the server, Cu, Mu and Bu respectively represent the CPU utilization rate, memory utilization rate and network bandwidth utilization rate of the server, and Conn represents the request connection number of the server; let the weight of the ith server be W (S)i) The performance index of the server, the performance index P (S) of the ith server is calculated by adopting the CPU capable of reflecting the performance of the server, the bandwidth of the network card and the size of the memoryi) Comprises the following steps:
introducing the load index L (S) of the serveri) The load of the server is inversely proportional to the weight, and the utilization rate of the CPU which can reflect the performance of the server is adoptedCalculating the load state of the server by the bandwidth utilization rate and the internal utilization rate of the network card, and calculating the load state L (S) of the ith serveri) Comprises the following steps:
L(Si)＝αCui+βMui+γBui
the long-short term memory network is a time cycle network and is specially designed for solving the long-term dependence problem of a general cyclic neural network, the long-short term memory network is adopted to construct a request connection number prediction model, the input dimensionality of the model is N, the input dimensionality represents the historical request connection number of a cluster server, the output dimensionality is 1, and the output dimensionality represents the request connection number q at the next momentt+1(ii) a Inputting normalized task data into a neural network, performing optimization training by using an Adaptive Moment Estimation (Adam) optimizer, and expressing the training effect of a model by using a Mean Square Error (MSE) function as a loss function; loss is reduced to a certain degree through multiple times of training, and a proper prediction model is obtained; through the model, the connection number q of the server at the next moment can be predictedt+1Using this index to predict the best node; because the weight of the server is not only connected with the number q of the next momentt+1Related to the number of connections q at the current timetThere is a great relationship, and therefore the requested connection number Rank (S) of the ith serveri) Comprises the following steps:
Rank(Si)＝αqt(Si)+βqt+1(Si)
different servers have different performances, and processing the same number of requests only occupies a small part of resources for the high-performance servers, but occupies most of resources for the low-performance servers, so that the load is unbalanced, so that the weight is inversely proportional to the number of request connections, and the higher the number of request connections is, the lower the weight is; in summary, the weight W (S) of the ith serveri) Comprises the following steps:
3. the load balancing apparatus for label resolution nodes based on machine learning according to claim 2, wherein the execution sequence among the modules of the load balancing apparatus is as follows,
(1) a user initiates an identification service request;
(2) the information collection module regularly collects the CPU utilization rate Cd, the memory utilization rate Md and the network card bandwidth utilization rate Bd of the servers in the cluster;
(3) the information collection module stores the collected information into a database;
(4) the load balancing algorithm module reads information in the database and calculates the weight of each server in the cluster;
(5) the load balancing algorithm module predicts the optimal node according to the weight and stores the information of the optimal node into a database;
(6) the load balancer reads the information of the optimal node;
(7) the load balancer transmits the request to the optimal node for processing;
(8) the node returns the request result of the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010854118.6A CN111930511A (en) | 2020-08-24 | 2020-08-24 | Identifier resolution node load balancing device based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010854118.6A CN111930511A (en) | 2020-08-24 | 2020-08-24 | Identifier resolution node load balancing device based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111930511A true CN111930511A (en) | 2020-11-13 |
Family
ID=73304976
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010854118.6A Pending CN111930511A (en) | 2020-08-24 | 2020-08-24 | Identifier resolution node load balancing device based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111930511A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112565399A (en) * | 2020-12-02 | 2021-03-26 | 天翼电子商务有限公司 | Adaptive traffic load balancing method for online learning |
| CN112866132A (en) * | 2020-12-31 | 2021-05-28 | 网络通信与安全紫金山实验室 | Dynamic load balancer and method for massive identification |
| CN113778683A (en) * | 2021-09-14 | 2021-12-10 | 码客工场工业科技(北京)有限公司 | Handle identification system analysis load balancing method based on neural network |
| CN114430398A (en) * | 2022-03-11 | 2022-05-03 | 北京邮电大学 | Bandwidth efficiency optimization method and device for aggregation compression of identification analysis requests |
| CN114866480A (en) * | 2022-05-31 | 2022-08-05 | 北京天融信网络安全技术有限公司 | NAT load balancing implementation method, system, electronic equipment and storage medium |
| CN114945024A (en) * | 2022-05-19 | 2022-08-26 | 东北林业大学 | Server load balancing optimization method based on long-term and short-term memory network |
| CN115208889A (en) * | 2022-05-12 | 2022-10-18 | 国家信息中心 | High-concurrency high-flow video safety isolation transmission method and system |
| CN115225733A (en) * | 2022-02-22 | 2022-10-21 | 北京邮电大学 | Identification analysis method and device based on direct routing and dynamic quantitative analysis load |
| CN116112493A (en) * | 2023-02-09 | 2023-05-12 | 网易（杭州）网络有限公司 | Communication method, device, electronic equipment and storage medium |
| CN116192858A (en) * | 2023-02-17 | 2023-05-30 | 通明智云(北京)科技有限公司 | Load balancing method and device based on weighted traffic |
| CN116418749A (en) * | 2023-02-17 | 2023-07-11 | 通明智云(北京)科技有限公司 | Load balancing method and device for dynamically adjusting weights |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108667878A (en) * | 2017-03-31 | 2018-10-16 | 北京京东尚科信息技术有限公司 | Server load balancing method and device, storage medium, electronic equipment |
| CN109104500A (en) * | 2018-09-29 | 2018-12-28 | 广东省信息工程有限公司 | A kind of server load balancing method and device of dynamic adjustment |
| CN110933139A (en) * | 2019-11-05 | 2020-03-27 | 浙江工业大学 | System and method for solving high concurrency of Web server |
| US20200125545A1 (en) * | 2018-10-18 | 2020-04-23 | Oracle International Corporation | Automated configuration parameter tuning for database performance |
| CN111277648A (en) * | 2020-01-19 | 2020-06-12 | 北京工业大学 | Nginx-based dynamic weight load balancing system and method |
-
2020
- 2020-08-24 CN CN202010854118.6A patent/CN111930511A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108667878A (en) * | 2017-03-31 | 2018-10-16 | 北京京东尚科信息技术有限公司 | Server load balancing method and device, storage medium, electronic equipment |
| CN109104500A (en) * | 2018-09-29 | 2018-12-28 | 广东省信息工程有限公司 | A kind of server load balancing method and device of dynamic adjustment |
| US20200125545A1 (en) * | 2018-10-18 | 2020-04-23 | Oracle International Corporation | Automated configuration parameter tuning for database performance |
| CN110933139A (en) * | 2019-11-05 | 2020-03-27 | 浙江工业大学 | System and method for solving high concurrency of Web server |
| CN111277648A (en) * | 2020-01-19 | 2020-06-12 | 北京工业大学 | Nginx-based dynamic weight load balancing system and method |
Non-Patent Citations (1)
| Title |
|---|
| 乔国娟等: "一种基于预测模型的负载均衡算法", 《计算机与现代化》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112565399A (en) * | 2020-12-02 | 2021-03-26 | 天翼电子商务有限公司 | Adaptive traffic load balancing method for online learning |
| CN112565399B (en) * | 2020-12-02 | 2022-12-09 | 天翼电子商务有限公司 | Adaptive traffic load balancing method for online learning |
| CN112866132A (en) * | 2020-12-31 | 2021-05-28 | 网络通信与安全紫金山实验室 | Dynamic load balancer and method for massive identification |
| CN113778683A (en) * | 2021-09-14 | 2021-12-10 | 码客工场工业科技(北京)有限公司 | Handle identification system analysis load balancing method based on neural network |
| CN115225733A (en) * | 2022-02-22 | 2022-10-21 | 北京邮电大学 | Identification analysis method and device based on direct routing and dynamic quantitative analysis load |
| CN115225733B (en) * | 2022-02-22 | 2024-04-05 | 北京邮电大学 | Identification analysis method and device based on direct routing and dynamic quantization analysis load |
| CN114430398B (en) * | 2022-03-11 | 2022-06-24 | 北京邮电大学 | Bandwidth efficiency optimization method and device for aggregation compression of identifier resolution request |
| CN114430398A (en) * | 2022-03-11 | 2022-05-03 | 北京邮电大学 | Bandwidth efficiency optimization method and device for aggregation compression of identification analysis requests |
| CN115208889A (en) * | 2022-05-12 | 2022-10-18 | 国家信息中心 | High-concurrency high-flow video safety isolation transmission method and system |
| CN115208889B (en) * | 2022-05-12 | 2023-11-28 | 国家信息中心 | High-concurrency large-flow video safety isolation transmission method and system |
| CN114945024A (en) * | 2022-05-19 | 2022-08-26 | 东北林业大学 | Server load balancing optimization method based on long-term and short-term memory network |
| CN114866480A (en) * | 2022-05-31 | 2022-08-05 | 北京天融信网络安全技术有限公司 | NAT load balancing implementation method, system, electronic equipment and storage medium |
| CN116112493A (en) * | 2023-02-09 | 2023-05-12 | 网易（杭州）网络有限公司 | Communication method, device, electronic equipment and storage medium |
| CN116192858A (en) * | 2023-02-17 | 2023-05-30 | 通明智云(北京)科技有限公司 | Load balancing method and device based on weighted traffic |
| CN116418749A (en) * | 2023-02-17 | 2023-07-11 | 通明智云(北京)科技有限公司 | Load balancing method and device for dynamically adjusting weights |
| CN116418749B (en) * | 2023-02-17 | 2023-11-17 | 通明智云(北京)科技有限公司 | Load balancing method and device for dynamically adjusting weights |
| CN116192858B (en) * | 2023-02-17 | 2024-01-09 | 通明智云(北京)科技有限公司 | Load balancing method and device based on weighted traffic |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111930511A (en) | Identifier resolution node load balancing device based on machine learning | |
| TWI725744B (en) | Method for establishing system resource prediction and resource management model through multi-layer correlations | |
| CN103152389A (en) | Method and system of responding peak access in cloud computer system | |
| CN104375897A (en) | Cloud computing resource scheduling method based on minimum relative load imbalance degree | |
| Ding et al. | Kubernetes-oriented microservice placement with dynamic resource allocation | |
| Paul et al. | I/o load balancing for big data hpc applications | |
| Zhu et al. | DR2: Dynamic request routing for tolerating latency variability in online cloud applications | |
| Cheng et al. | Proscale: Proactive autoscaling for microservice with time-varying workload at the edge | |
| Kumar et al. | Novel Dynamic Scaling Algorithm for Energy Efficient Cloud Computing. | |
| El-Zoghdy | A hierarchical load balancing policy for grid computing environment | |
| CN113014649B (en) | Cloud Internet of things load balancing method, device and equipment based on deep learning | |
| Khelifa et al. | SLA-aware task scheduling and data replication for enhancing provider profit in clouds | |
| Liao et al. | Efficient time-evolving stream processing at scale | |
| KR20160044623A (en) | Load Balancing Method for a Linux Virtual Server | |
| Zhang et al. | NADE: nodes performance awareness and accurate distance evaluation for degraded read in heterogeneous distributed erasure code-based storage | |
| Sujaudeen et al. | TARNN: Task‐aware autonomic resource management using neural networks in cloud environment | |
| Lin et al. | A workload-driven approach to dynamic data balancing in MongoDB | |
| Dharmapriya et al. | Smart platform for cloud service providers | |
| Subbiah et al. | Energy efficient big data infrastructure management in geo-federated cloud data centers | |
| Muchori et al. | Machine Learning Load Balancing Techniques in Cloud Computing: A Review | |
| CN112698944A (en) | Distributed cloud computing system and method based on human brain simulation | |
| Abdel Raouf et al. | A predictive replication for multi‐tenant databases using deep learning | |
| Zhao et al. | vSpec: Workload-adaptive operating system specialization for virtual machines in cloud computing | |
| Cavallo et al. | Fragmenting Big Data to boost the performance of MapReduce in geographical computing contexts | |
| Duan et al. | Simulation of Cloud Computing Resource Allocation Optimization Model Based on Graph Neural Network |
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 |