CN113312151A - Load balancing method of IPSecVPN cluster - Google Patents
Load balancing method of IPSecVPN cluster Download PDFInfo
- Publication number
- CN113312151A CN113312151A CN202110697486.9A CN202110697486A CN113312151A CN 113312151 A CN113312151 A CN 113312151A CN 202110697486 A CN202110697486 A CN 202110697486A CN 113312151 A CN113312151 A CN 113312151A
- Authority
- CN
- China
- Prior art keywords
- load
- service node
- server
- node
- load balancing
- 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 47
- 239000002131 composite material Substances 0.000 claims description 9
- 230000004044 response Effects 0.000 claims description 9
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 claims description 6
- 239000003345 natural gas Substances 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 14
- 238000004364 calculation method Methods 0.000 description 22
- 230000003068 static effect Effects 0.000 description 12
- 230000008569 process Effects 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 5
- 238000013459 approach Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012546 transfer 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
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- 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
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
-
- 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
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45595—Network integration; Enabling network access in virtual machine instances
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer And Data Communications (AREA)
Abstract
The invention belongs to the technical field of load balancing, and particularly relates to a load balancing method of an IPSec VPN cluster. Setting initial time step length, collecting state information of each service node, and storing the state information into corresponding attribute values of the nodes; calculating the load degree of each service node, and selecting an optimal node; predicting the next feedback step length according to the feedback step length adjusting formula and the load degree of each service node; and the next feedback step length load balancer sends a request for acquiring the load attribute state to the server. The invention provides a load balancing method of an IPSec VPN cluster aiming at the problem of load balancing in an IPSec VPN cluster environment. The invention can better promote the load balance of the IPSec VPN cluster and effectively improve the processing efficiency.
Description
Technical Field
The invention belongs to the technical field of load balancing, and particularly relates to a load balancing method of an IPSec VPN cluster.
Background
In the era of high-speed connection to the internet, demands for personal information security and business cases, such as cloud computing, large data throughput, and stable connection, are becoming increasingly important. When the server processes high concurrent access and a large amount of data, the server needs to analyze traffic and return results in time, and the process is very obvious in experience feeling of a user, so that great challenges are brought to the performance of the server. Poor processing performance of the server may result in the user failing to meet the demand, and further, the business of the company is greatly affected. When the server crashes due to resource exhaustion, the whole server service is greatly affected, so that it is important to accelerate the processing performance of the server.
In addition to the increased performance of a single server, it is also contemplated to use multiple servers instead of the previous task of a single server. Server clustering and load balancing techniques have evolved and have optimized server performance in terms of hardware and software.
Load balancing is the distribution of workload to several nodes in a workspace to ensure that no node in the system is idle or overloaded. Load balancing may be a study of server clusters, multi-core, disk drives, or other devices, whereas the present invention only deals with the load balancing problem for IPSec VPN gateways implemented in a cluster.
An efficient load balancing method determines which node to handle the current task by comparing the possible workload of each node in the system. Common load balancing methods include static load balancing and dynamic load balancing.
The static balancing method is suitable for a system with small load change in the task execution process. The dynamic balancing method is more adaptive and efficient in both homogeneous and heterogeneous environments. In the static load balancing approach, traffic is evenly distributed among the servers. This approach requires a priori knowledge of the system resources. The performance of the processor is specified at the beginning of execution, and therefore the decision to transfer load is not dependent on the current state of the system. Static load balancing applications have fewer overhead procedures than dynamic load balancing. Load balancing helps to allocate resources fairly to improve the stability of the system. The static load balancing approach has some drawbacks. For example, a task is assigned to a processor or computer only after it is created, and cannot be transferred to any other computer for load balancing during execution.
Unlike static load balancing, in dynamic load balancing, the load decision is affected by the current state of the system. Many attributes are taken into account. The load may also be assigned to a particular node and then reassigned to another node based on runtime load attribute information collected while performing the task. Despite the increased system overhead due to runtime communication between nodes, dynamic load balancing techniques have many advantages over static balancing techniques. In the case where one node is turned off, the system is not stopped although the performance of the system is degraded. When some nodes are overloaded, tasks can be moved to underutilized nodes, and the method is more universal than a static method and provides better results under heterogeneous and dynamic conditions.
In order to improve the accuracy and the efficiency of decision making, the best load balancing method can dynamically consider the load condition of each service node in real time and select a proper service node through load calculation so as to achieve load balancing. To study a load balancing method for the IPSec VPN, it is necessary to study a load balancing method suitable for this specific environment, starting from characteristics of the IPSec VPN.
Disclosure of Invention
The invention aims to provide a load balancing method of an IPSecVPN cluster.
The purpose of the invention is realized by the following technical scheme: the method comprises the following steps:
step 1: setting an initial time step, starting to collect state information of each service node, and storing the state information into corresponding attribute values of each service node;
step 2: calculating the load degree alpha of each service node iiSelecting the service node with the lowest corresponding load value as the optimal service node;
step 2.1: computing composite load CTL for service node ii;
CTLi=βdNsa+βaRtime+βbNalloc+βcNconn
Wherein N issaEstablishing numbers for IKE SA; rtimeIs the response time; n is a radical ofallocNumber of assigned IPSec requests; n is a radical ofconnThe number of established IPSec SA; beta is aa、βb、βc、βdIs a weight coefficient, betaa+βb+βc+βd=1;
Step 2.2: computing the carrying capacity PE of a service node ii;
Wherein, CsiThe CPU frequency of the service node i; msiThe memory capacity of the service node i; ns (natural gas)iA back-end server network bandwidth for service node i; ac is the average frequency of the CPUs of all the service nodes in the server; am is the average memory capacity of all service nodes in the server; an is the average network bandwidth of the back-end servers of all the service nodes in the server; alpha is alphac、αm、αnIs a weight coefficient, αc+αm+αn=1;
Step 2.3: calculating the weight WE of a service node ii;
Step 2.4: calculating the load degree alpha of the service node ii;
And step 3: degree of load α through each service node iiPredicting the next feedback step length tau;
wherein τ' represents a previous feedback step;andrespectively representing the average load degree of all service nodes in the server in the first two feedback time step lengths;
and 4, step 4: and (3) the load balancer sends a request for obtaining the load attribute state to the server when the next feedback step is long, and the step (2) is returned.
The invention has the beneficial effects that:
the invention provides a load balancing method of an IPSec VPN cluster aiming at the problem of load balancing in an IPSec VPN cluster environment. The invention can better promote the load balance of the IPSec VPN cluster and effectively improve the processing efficiency.
Drawings
Fig. 1 is a diagram of an IPSec VPN cluster environment.
Fig. 2 is a comparison effect diagram of the load balancing method of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The invention provides a load balancing method of an IPSec VPN cluster aiming at the problem of load balancing in an IPSec VPN cluster environment. The invention can better promote the load balance of the IPSec VPN cluster and effectively improve the processing efficiency.
A load balancing method of IPSec VPN cluster comprises the following steps:
step 1: setting an initial time step, starting to collect state information of each service node, and storing the state information into corresponding attribute values of each service node;
step 2: calculating the load degree alpha of each service node iiSelecting the service node with the lowest corresponding load value as the optimal service node;
step 2.1: computingCompound load CTL of service node ii;
CTLi=βdNsa+βaRtime+βbNalloc+βcNconn
Wherein N issaEstablishing numbers for IKE SA; rtimeIs the response time; n is a radical ofallocNumber of assigned IPSec requests; n is a radical ofconnThe number of established IPSec SA; beta is aa、βb、βc、βdIs a weight coefficient, betaa+βb+βc+βd=1;
Step 2.2: computing the carrying capacity PE of a service node ii;
Wherein, CsiThe CPU frequency of the service node i; msiThe memory capacity of the service node i; ns (natural gas)iA back-end server network bandwidth for service node i; ac is the average frequency of the CPUs of all the service nodes in the server; am is the average memory capacity of all service nodes in the server; an is the average network bandwidth of the back-end servers of all the service nodes in the server; alpha is alphac、αm、αnIs a weight coefficient, αc+αm+αn=1;
Step 2.3: calculating the weight WE of a service node ii;
Step 2.4: calculating the load degree alpha of the service node ii;
And step 3: degree of load α through each service node iiPrediction ofThe next feedback step τ;
wherein τ' represents a previous feedback step;andrespectively representing the average load degree of all service nodes in the server in the first two feedback time step lengths;
and 4, step 4: and (3) the load balancer sends a request for obtaining the load attribute state to the server when the next feedback step is long, and the step (2) is returned.
Example 1:
the cluster environment deployment mode adopted by the invention is shown in fig. 1, the IPSec VPN gateway is replaced by a load balancer in the environment, the load balancer is only responsible for distributing tasks, and the specific IPSec data packet processing process is completed in the service node of the cluster. The cluster also shares one IP address, but another set of IP address is arranged in the cluster for communicating with the load balancer to feed back real-time load information. The service node receives the response information of the real server, and the response information is directly returned to the client after being encrypted without passing through the load balancer, so that the task load of the load balancer is reduced, and the phenomenon of rapid performance reduction can not occur even if the request quantity is large, namely, the single-point performance bottleneck effect does not exist.
To achieve better security during longer VPN sessions, IPSec provides a mechanism to periodically update IKE and IPSec keys. Updating the IKE key requires running two IKE phases, but updating the IPSec key only requires running the second phase again (fast mode). IKE typically does not run only once during the establishment of a virtual private network connection. It runs periodically during longer virtual private network sessions to enhance security. Furthermore, frequent disconnections in a wireless environment may cause a client to run IKE multiple times.
There are studies conducted and concluded with respect to various overheads of IPSec:
1) the overhead of the IKE protocol is much higher than that generated by the ESP processing the data packet;
2) the encryption operation accounts for 32-60% of the IKE overhead and 34-55% of the ESP overhead;
3) digital signature generation and Diffie-Hellman computation are the largest sources of overhead in the IKE process, and only a small amount of overhead can be attributed to symmetric key encryption and hashing;
4) symmetric key encryption is the most expensive operation in ESP processing.
In conclusion, the present invention uses IKE negotiation as an important basis in load calculation. In addition to the factors of IKE negotiation, the load attributes also take into account other factors:
1) considering the processing power, i.e. the performance, of the server, this property represents how much the server can handle the amount of tasks. Attributes determining the processing capacity of the server include CPU frequency, memory capacity, network bandwidth and hard disk capacity, and the hard disk capacity is not considered to be placed in load calculation of the server, because IPSec traffic processing generally occupies little hard disk space, and attribute weights with small changes do not change greatly, which is not very useful for load calculation.
2) The parameters of various attributes of the server are dynamically changed in the running process, and the attributes reflect the real-time load condition of the server. The dynamic parameters include CPU utilization, memory utilization, and network bandwidth utilization. These parameters can all be obtained by obtaining the corresponding file of the server.
3) The dynamic load balancing method based on the minimum number of connections is to give priority to the server with the minimum number of connections in consideration that the performance of the server with the minimum number of connections may be better. For the processing procedure of the IPSec VPN gateway, various SA numbers generated in the IPSec connection procedure may be used as a load basis. The IPSec connection process includes performing IKE negotiation first to generate IKE SA, and then generating IPSec SA using the negotiated information, so that the established IKE SA and IPSec SA numbers can be obtained as load attributes.
4) After receiving the IPSec connection request, the load balancer selects the best serving node to distribute the IPSec requests, which updates the number of IPSec requests distributed by each server, and when the number of IPSec requests distributed by a certain server increases, the overall performance of the server is affected, and at this time, the load of the server should increase, so this attribute is summarized in load calculation.
In summary, the load attributes include: static parameters (CPU frequency, memory capacity, and network bandwidth) and dynamic parameters (CPU utilization, memory utilization, network bandwidth utilization, response time, number of assigned IPSec requests, number of established IPSec SA establishments, and number of IKE SA establishments). The load balancer uses these attribute values to compute the load state of the service node. After each load calculation is completed, the number of distributed IPSec requests of all service nodes is cleared, so that the number of IPSec requests is a variable value in a time step in the next load calculation.
The static load parameters (CPU frequency, memory capacity and network bandwidth) reflect the load-bearing capacity of the server, i.e. the performance of the server. The CPU utilization, memory utilization and network bandwidth utilization in the dynamic load parameters reflect the load on the server. And the response time, the number of IPSec vpn requests, the number of IPSec SA setups, and the number of IKE SA setups indirectly reflect the load situation.
And calculating the load capacity and load of the server node according to the collected load information, and further calculating the weight of the server. Finally, the composite load will be calculated from the response time, the number of assigned IPSec security channel establishment requests, the number of established IPSec security channels, and the number of SA establishments. The greater the weight of the server, the less the composite load, and the greater the likelihood of assigning it to a request.
1) Calculation of composite load
Considering the composite load (CTL) combining the IKE SA establishment number, response time, allocated IPSec request number, and established IPSec SA number, and using CTLi to represent the composite load of node n, the calculation method is shown in equation (1).
CTLi=βdNsa+βaRtime+βbNalloc+βcNconn (1)
Wherein, betaa+βb+βc+βd1 is used to reflect the importance of each factor. When CTL is usediWhen smaller, the probability of assigning a request to node n is greater.
2) Calculation of node bearing capacity
The static load parameter reflects the load-bearing capacity of the server, i.e. the performance of the server. Ac, Am and An respectively represent the average frequency and the average memory capacity of the CPU and the average network bandwidth of the back-end server, are respectively calculated by using a formula (2), a formula (3) and a formula (4), then the bearing capacity of the ith node is calculated by using a formula (5), and the PE is usediTo indicate. Wherein alpha isc+αm+αnThis is to reflect the importance factor of each.
3) Calculation of node load conditions
The dynamic load factor reflects the load on the server. The load of the ith node is calculated using the following equation (6), LSiAnd (4) showing.
LSi=αcCsj+αmMsj+αnNsj (6)
When LSiIf the load is larger than a certain value, the load of the node i can be considered to be too largeThe load distribution to it should be stopped.
4) Calculation of node weights
Let the weight of a node be WEi. The greater the weight, the greater the likelihood of being assigned to a request. WEiBearer capability PE with serveriPositively correlated with LSiA negative correlation. WEiThe calculation formula of (2) is shown in formula (7).
5) Calculating load degree of service node
From the foregoing analysis, it is possible to know the probability that node n is assigned to a request and its composite load CTLnNegatively correlated with its weight WEnAnd (4) positively correlating. The load degree of the service node i is shown in formula (8).
The node with the lowest load degree is selected as the best serving node as shown in equation (9).
In the load calculation, the node load capacity, the node load condition and the composite load all relate to the weight value of the attribute, and the calculation of the attribute weight is discussed next. In many dynamic load balancing methods, a static weight is used to calculate a load, and when an attribute index of a server reaches the load, the calculation method still uses an originally determined weight to calculate, which may cause a situation that the server is crashed due to a certain attribute because of small load calculation, and the method cannot dynamically adjust, and the distributed tasks cannot be processed. Therefore, after the attribute weight is calculated, the attribute weight is adjusted, and then the server load is calculated according to the collected server attribute information.
1) Data normalization
Since the metrics of the indices are not consistent, the data must first be normalized. It maps the result value to [0, 1] using Min-max normalization. The normalized calculation formula is shown in formula (11).
YijIs a pre-processed value of server j with respect to attribute i.
2) Determining a weight coefficient
The proportion of the server j in the attribute i under the attribute is pijThe calculation method is shown in formula (12).
The entropy of information for each attribute is denoted as E1,E2,...,EnAs shown in equation (13).
The weight W of each attribute index is calculated using equation (14)i。
The method takes attribute data obtained in an experiment as experimental data, normalization is carried out through a formula (11), and then a weight coefficient of the attribute is obtained through information entropy calculation.
3) Weight adjustment
Equation (15) represents the current of all the attributes of each serverRatio of properties Fi. The server node weights may be adjusted as shown in equation (16).
W′i=Fi*Wi (16)
The condition of the load imbalance node can be improved by adjusting the weight. However, the imbalance of the weighting is usually caused by high concurrent accesses in a short time. Therefore, when there are no high load nodes, the server node weight will gradually adjust to the initial weight.
The dynamic feedback load balancing method can be divided into the following strategies: information collection, feedback, and task reassignment. Collecting load information and using it to determine task migration or reallocation to improve the degree of load balancing
After the task is completed at a given time step, load information is collected and the task is dynamically redistributed based on historical load information. The feedback mechanism aims to realize dynamic load balancing and make the method adaptive. Load information feedback may result in resource overhead. However, for large scale computation problems, the feedback overhead does not significantly affect the computation efficiency if the feedback time step is properly controlled. Dynamic load balancing is achieved by using a load prediction method in conjunction with a dynamic programming method for task reallocation. The feedback time step is dynamically varied to adapt to the dynamic environment. When the load changes frequently, the method can achieve better adaptability.
1) Information gathering policy
The main procedure is set as a time step loop. In the process of one step length, the load balancer sends a request to the service node to acquire each load attribute value, and the load degree of each server is calculated through a load calculation method. And then adjusting the feedback time step according to the average load degree of the service node.
2) Feedback strategy and feedback time step prediction
To reduce resource overhead, the processing power of each node is predicted and the task is reallocated once per feedback time step. To avoid reducing the throughput of the system, for a given feedback time step, allThe same allocation scheme is used in the steps. The exponential smoothing method is used for predicting the processing capacity of each node at the next time step. The feedback time step is based on the average load balance of all servers in the first two feedback time steps (And). The next feedback time step long time average load method is shown in equation (17).
When the average load balance degree in the next time step is smaller, the load of some current servers can be predicted to be smaller, the feedback step can be properly increased, and when the average load degree is larger, the load of some current servers is predicted to be larger, and the feedback step needs to be reduced. When the average load degree is basically unchanged, the load change of all the current servers is predicted to be small, and the feedback step length does not need to be changed.
Therefore, the average load difference between the server in the next feedback step and the server in the previous feedback step is used as the basis for the change of the feedback step, and the calculation method is shown in formula (18).
τ' is the last feedback step.Since a system is considered to be stable when the average load change degree is relatively small, the last feedback step is used as it is.This is the case when the load becomes large, thus reducing the step size.This is the case when the load becomes small and the step size should be increased.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (1)
1. A load balancing method of IPSecVPN cluster is characterized by comprising the following steps:
step 1: setting an initial time step, starting to collect state information of each service node, and storing the state information into corresponding attribute values of each service node;
step 2: calculating the load degree alpha of each service node iiSelecting the service node with the lowest corresponding load value as the optimal service node;
step 2.1: computing composite load CTL for service node ii;
CTLi=βdNsaAnd betaaRtimeAnd betabNallocAnd betacNconn
Wherein N issaEstablishing numbers for IKE SA; rtimeIs the response time; n is a radical ofallocNumber of assigned IPSec requests; n is a radical ofconnThe number of established IPSec SA; beta is aa、βb、βc、βdIs a weight coefficient, betaa+βb+βc+βd=1;
Step 2.2: computing the carrying capacity PE of a service node ii;
Wherein, CsiThe CPU frequency of the service node i; msiIs a garmentMemory capacity of service node i; ns (natural gas)iA back-end server network bandwidth for service node i; ac is the average frequency of the CPUs of all the service nodes in the server; am is the average memory capacity of all service nodes in the server; an is the average network bandwidth of the back-end servers of all the service nodes in the server; alpha is alphac、αm、αnIs a weight coefficient, αc+αm+αn=1;
Step 2.3: calculating the weight WE of a service node ii;
Step 2.4: calculating the load degree alpha of the service node ii;
And step 3: degree of load α through each service node iiPredicting the next feedback step length tau;
wherein τ' represents a previous feedback step;andrespectively representing the average load degree of all service nodes in the server in the first two feedback time step lengths;
and 4, step 4: and (3) the load balancer sends a request for obtaining the load attribute state to the server when the next feedback step is long, and the step (2) is returned.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110697486.9A CN113312151A (en) | 2021-06-23 | 2021-06-23 | Load balancing method of IPSecVPN cluster |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110697486.9A CN113312151A (en) | 2021-06-23 | 2021-06-23 | Load balancing method of IPSecVPN cluster |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113312151A true CN113312151A (en) | 2021-08-27 |
Family
ID=77380093
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110697486.9A Pending CN113312151A (en) | 2021-06-23 | 2021-06-23 | Load balancing method of IPSecVPN cluster |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113312151A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115903638A (en) * | 2022-11-08 | 2023-04-04 | 广州鲁邦通智能科技有限公司 | VPN technology-based equipment and method and system for remotely maintaining PLC |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080195755A1 (en) * | 2007-02-12 | 2008-08-14 | Ying Lu | Method and apparatus for load balancing with server state change awareness |
CN101820384A (en) * | 2010-02-05 | 2010-09-01 | 浪潮(北京)电子信息产业有限公司 | Method and device for dynamically distributing cluster services |
US20120109852A1 (en) * | 2010-10-27 | 2012-05-03 | Microsoft Corporation | Reactive load balancing for distributed systems |
CN103220354A (en) * | 2013-04-18 | 2013-07-24 | 广东宜通世纪科技股份有限公司 | Method for achieving load balancing of server cluster |
CN104184675A (en) * | 2014-09-12 | 2014-12-03 | 成都卫士通信息产业股份有限公司 | Load-balanced IPSec VPN device trunking system and working method of load-balanced IPSec VPN device trunking system |
CN105007337A (en) * | 2015-08-20 | 2015-10-28 | 浪潮(北京)电子信息产业有限公司 | Cluster system load balancing method and system thereof |
CN106325976A (en) * | 2016-08-05 | 2017-01-11 | 天河国云(北京)科技有限公司 | Rendering task scheduling processing method and server |
-
2021
- 2021-06-23 CN CN202110697486.9A patent/CN113312151A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080195755A1 (en) * | 2007-02-12 | 2008-08-14 | Ying Lu | Method and apparatus for load balancing with server state change awareness |
CN101820384A (en) * | 2010-02-05 | 2010-09-01 | 浪潮(北京)电子信息产业有限公司 | Method and device for dynamically distributing cluster services |
US20120109852A1 (en) * | 2010-10-27 | 2012-05-03 | Microsoft Corporation | Reactive load balancing for distributed systems |
CN103220354A (en) * | 2013-04-18 | 2013-07-24 | 广东宜通世纪科技股份有限公司 | Method for achieving load balancing of server cluster |
CN104184675A (en) * | 2014-09-12 | 2014-12-03 | 成都卫士通信息产业股份有限公司 | Load-balanced IPSec VPN device trunking system and working method of load-balanced IPSec VPN device trunking system |
CN105007337A (en) * | 2015-08-20 | 2015-10-28 | 浪潮(北京)电子信息产业有限公司 | Cluster system load balancing method and system thereof |
CN106325976A (en) * | 2016-08-05 | 2017-01-11 | 天河国云(北京)科技有限公司 | Rendering task scheduling processing method and server |
Non-Patent Citations (1)
Title |
---|
田绍亮;左明;吴绍伟;: "一种改进的基于动态反馈的负载均衡算法", 计算机工程与设计, no. 03 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115903638A (en) * | 2022-11-08 | 2023-04-04 | 广州鲁邦通智能科技有限公司 | VPN technology-based equipment and method and system for remotely maintaining PLC |
CN115903638B (en) * | 2022-11-08 | 2023-12-08 | 广州鲁邦通智能科技有限公司 | Equipment based on VPN technology and method and system for remotely maintaining PLC |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Piao et al. | A network-aware virtual machine placement and migration approach in cloud computing | |
CN113110933B (en) | System with Nginx load balancing technology | |
Vignesh et al. | Resource management and scheduling in cloud environment | |
Shah et al. | Static load balancing algorithms in cloud computing: challenges & solutions | |
Ahn et al. | Competitive partial computation offloading for maximizing energy efficiency in mobile cloud computing | |
Ma et al. | Real-time virtual machine scheduling in industry IoT network: A reinforcement learning method | |
CN108416465B (en) | Workflow optimization method in mobile cloud environment | |
US20060200469A1 (en) | Global session identifiers in a multi-node system | |
Meskar et al. | Fair multi-resource allocation with external resource for mobile edge computing | |
Chhabra et al. | A probabilistic model for finding an optimal host framework and load distribution in cloud environment | |
CN115718644A (en) | Computing task cross-region migration method and system for cloud data center | |
JP2015011365A (en) | Provisioning apparatus, system, provisioning method, and provisioning program | |
CN113312151A (en) | Load balancing method of IPSecVPN cluster | |
Hayat et al. | Efficient energy utilization in cloud fog environment | |
CN112954012B (en) | Cloud task scheduling method based on improved simulated annealing algorithm of load | |
Chatterjee et al. | A new clustered load balancing approach for distributed systems | |
Ashalatha et al. | Dynamic load balancing methods for resource optimization in cloud computing environment | |
Yang et al. | Enhancement of anticipative recursively adjusting mechanism for redundant parallel file transfer in data grids | |
CN112685167A (en) | Resource using method, electronic device and computer program product | |
KR100478346B1 (en) | Load balancing method based on graph-partitioning for cluster sever and the system | |
CN113377544A (en) | Web cluster load balancing method based on load data dynamic update rate | |
Ali et al. | A hybrid elephant herding optimization and harmony search algorithm for potential load balancing in cloud environments | |
Singhi et al. | A load balancing approach for increasing the resource utilization by minimizing the number of active servers | |
Patel et al. | Analysis of various task scheduling algorithms in cloud computing | |
Aggarwal et al. | Dynamic load balancing based on CPU utilization and data locality in distributed database using priority policy |
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 |