Load balancing strategy method adaptive to air-sea cross-domain network control plane under SDN (Software Defined Network) architecture

Abstract

The invention relates to the technical field of wireless communication, and discloses a load balancing strategy method adaptive to an air-sea cross-domain network control plane under an SDN architecture. The invention provides a load balancing strategy method adaptive to an air-sea cross-domain network control plane under an SDN (Software Defined Network) architecture, aiming at the problems that the network service life is greatly shortened due to frequent routing redirection, migration is difficult due to overload of multiple controllers, performance loss and migration conflict exist, the response time of the control plane is too long and the like in a traditional control plane load balancing strategy. According to the method, the particularity and complexity of the air-sea cross-domain network are fully considered, and the constraint condition is taken as the premise of switch migration, so that an intelligent migration strategy for various QoS requirements is realized, and a more intelligent, efficient and reliable management scheme is provided for an air-sea cross-domain network control plane based on the SDN.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a load balancing strategy method for an air-sea cross-domain network control plane adapted to an SDN architecture. Background Art

[0002] With the rise of emerging technologies such as cloud computing, big data, and the Internet of Things, traditional network architectures can no longer meet the high performance, high scalability, and high security requirements of modern data centers and cross-domain networks in the air and sea. As a new type of network architecture, SDN separates the control plane from the data plane of the network, achieving flexible scheduling and management of network traffic. It has gradually become a key technology to change the rigid architecture of existing cross-domain communication networks in the air and sea, and continuously promote the transformation of networks to meet differentiated QoS requirements.

[0003] In the SDN architecture, the control plane is responsible for centrally managing network resources and providing flexible network traffic scheduling and control capabilities. However, in the air-sea cross-domain network environment, due to the network scale and network environment diversity, the load balancing problem of the control plane becomes particularly prominent. Traditional load balancing strategies are often based on hardware or simple algorithms, which cannot meet the requirements of dynamic, intelligent and efficient management of network resources under the SDN architecture.

[0004] Therefore, it is necessary to provide a load balancing strategy method for the air-sea cross-domain network control plane adapted to the SDN architecture, aiming to achieve efficient utilization of network resources and controller load balancing by dynamically adjusting network traffic or control relationships, thereby improving network performance and stability; at the same time, the strategy also takes into account the particularity and complexity of the air-sea cross-domain network, such as the diversity of network topology, transmission delay, and bandwidth limitation, to ensure the effectiveness and feasibility of the strategy. Summary of the invention

[0005] In order to solve the technical problems of existing control plane load balancing strategies such as difficulty in coping with multi-controller overload and migration conflicts, the present invention provides a load balancing strategy method for air-sea cross-domain network control plane adapted to SDN architecture. The method is based on switch migration and limits the switch migration objects during controller allocation through constraints, thereby avoiding the difficulties of migration conflicts and new load problems caused by migration, realizing dynamic allocation of controllers, and realizing intelligent load balancing strategies based on user and task requirements by changing constraints.

[0006] The load balancing strategy method for the air-sea cross-domain network control plane adapted to the SDN architecture provided by the present invention includes an SDN control plane structure design and a corresponding load balancing strategy.

[0007] The SDN control plane structure is designed as follows: Generally speaking, the air-sea cross-domain network based on SDN adopts a single controller structure. For small underwater networks, a single controller can indeed manage the entire system, thereby reducing costs. However, the large-scale air-sea cross-domain communication network has a large number of nodes. If only a single controller is used to manage the system, the following problems will arise: all traffic will be forwarded to one controller. As the number of nodes increases, the traffic of the controller will also increase rapidly, causing system performance bottlenecks; secondly, due to the large coverage of large-scale networks, nodes far away from the controller will not be able to receive feedback in time, causing difficulties in data exchange and synchronization; finally, since the air-sea cross-domain network forwarding path under the SDN architecture is controlled by the flow table distributed by the controller, if a single controller fails or is attacked, almost all nodes will lose communication functions, causing system paralysis. Therefore, a multi-controller architecture is a better choice for underwater sensor acoustic networks based on SDN.

[0008] The control plane of the present invention adopts the design concept of "physically distributed and logically centralized", that is, the SDN control layer functions are realized by multiple devices (mostly buoys in the air-sea cross-domain communication environment) equipped with SDN controllers, which can enjoy the scalability of the distributed structure and retain the simplicity of the centralized structure. A relatively clear control layer concept is composed of a master controller, a master-slave controller, and a secondary slave controller mounted on different devices.

[0009] Divide the data layer into subdomains, each containing Switches (It should be noted that, unlike the network devices with clear division of labor in traditional networks, the switch nodes of the air-sea cross-domain network are not limited to data forwarding functions, but are multifunctional nodes that can collect and preliminarily process environmental information and realize the forwarding function of the data link layer like ordinary switches), each subdomain corresponds to a main controller, and To improve the utilization of controller performance, a controller that serves as a master-slave controller of a network can also be a secondary slave controller of another network with a similar physical location.

[0010] In the control plane of a multi-controller structure, load balancing is an indispensable link. Generally speaking, there are two solutions to achieve load balancing: traffic migration and switch migration. Traffic migration migrates traffic based on the granularity of traffic problems, and achieves a more balanced distribution of network traffic through the migration of control plane traffic, but the process requires continuous rerouting that occupies a large amount of network resources. In contrast, switch migration considers transferring the entire switch from a controller with a larger load to a controller with a lighter load, and achieves load balancing through dynamic allocation. Since the cross-domain communication environment in the air and sea includes an underwater environment with tight network resources, switch migration seems to be a more effective solution.

[0011] Based on the above control plane structure, the node set of each subdomain can be expressed as: , in, Indicates a switch. Indicates the main controller, Indicates the slave controller, then the entire air-sea cross-domain network has controllers, the collection is ,as well as switches, the set is . It is assumed that any switch belongs to only one subdomain.

[0012] First, the control relationship between the controller and the switch can be represented by a matrix, called the control matrix. , When By controller Control; In order to achieve load balancing, when a master controller reaches a load threshold, it is necessary to change the control matrix so that the control authority is transferred to the slave controller. This process can also be represented by a matrix, called the (switch) migration matrix: , The migration matrix represents the switch The corresponding controller can migrate state when When the master controller When the load reaches a critical value, the switch can migrate to .

[0013] Assume that the switch The load is , controller The capacity is Since switch migration often results in additional controller load, a certain amount of redundancy should be left when setting the migration threshold. ,Right now: , in For controller Total load: , The model can be obtained as follows: ① Solve the optimal migration matrix according to the constraints ,example: Constraints: , , The parameters are as follows: , , , , , Rt is the decision parameter, , , , , They are all weighted quality of service parameters (QoS). Rt is changed by adjusting the weight of QoS, and the control relationship matrix is ​​solved by the constraint condition that minimizes Rt. In this way, the load balancing strategy can be changed according to the QoS required by the task (this is achieved by adjusting the weight of each QoS). , , , , They are the relative transmission delay, relative queuing delay, packet loss rate, bandwidth utilization, and relative available bandwidth of the current link; ②New control matrix after migration: ; ③It should be noted that the controller load redundancy is always required during the calculation process, that is: , Each switch is assigned only one controller, namely: , And only 0 and 1 are used in the control matrix to represent the control relationship, that is: .

[0014] Compared with the related art, the load balancing strategy method for the air-sea cross-domain network control plane adapted to the SDN architecture provided by the present invention has the following beneficial effects: The present invention proposes a load balancing strategy under a "physically distributed, logically centralized" control plane structure, which avoids the migration conflicts, long response time and impaired routing performance that are easily encountered in traditional multi-controller load balancing strategies, thereby reducing service quality and weakening the effectiveness of load balancing strategies. It reduces the control plane response delay and migration conflict probability, significantly improves the performance of air-sea cross-domain networks, and provides a load balancing strategy based on changing constraints to adapt to user / task requirements, providing an efficient and flexible load balancing strategy for the air-sea cross-domain network control plane under the SDN architecture. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of the initial allocation of switch control relations in the present invention; Figure 2 This is a schematic diagram of the control relationship after load balancing. DETAILED DESCRIPTION

[0016] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0017] The specific implementation method of the load balancing strategy method of the air-sea cross-domain network control plane adapted to the SDN architecture proposed in the present invention is as follows: After the network is deployed, the data plane is evenly divided into multiple subdomains D, and then all subdomains are traversed to complete the allocation of controllers, thereby constructing the corresponding control matrix and Switch Migration Matrix In each subdomain, the master controller is assigned according to the standard that the controller with the largest maximum load capacity and the smallest average control path delay is the master controller of the subdomain. The initial assignment is as follows: Figure 1 As shown. Subsequently, the remaining controllers will also become the priority slave controllers of the subdomain and the secondary slave controllers of the adjacent subdomain according to this standard. In this way, a migration matrix can be formed for each switch as a reference when there is a migration demand.

[0018] Before executing a network task, you need to set load redundancy for the controller to prevent the controller from being overloaded due to the additional load generated during switch migration, so as to ensure the stable operation of the network. Among them, since the main controller is the center of the load of each sub-domain switch, it needs to reserve more redundancy than the slave controller and the secondary slave controller.

[0019] During the execution of network tasks, and Under the action of , whenever the load of the switch exceeds the capacity of its main controller, the network will solve the migration matrix according to the constraints. , thereby changing the control matrix , thus migrating it to the corresponding slave controller under the constraints of the switch migration optimization model, achieving overall load balancing of the control plane and maintaining stable and efficient network operation, such as Figure 2 shown.

[0020] In the case given in the present invention, the switch migration constraint is a weighted parameter In actual applications, the network performance required by different tasks is often different, that is, the requirements for service quality parameters (QoS) such as latency, stability, and bit error rate are different, such as the communication requirements of three types of application scenarios given in Table 1. Therefore, when executing tasks in the actual network, it is necessary to flexibly change the constraints according to the task requirements, implement differentiated and intelligent load balancing strategies based on QoS, and adopt the form of weighted parameters to better adapt to actual needs.

[0021] Table 1 Communication requirements of three application scenarios .

[0022] Compared with the related art, the load balancing strategy method for the air-sea cross-domain network control plane adapted to the SDN architecture provided by the present invention has the following beneficial effects: 1. Reduce migration conflicts: Limit switch migration objects through migration matrix constraints to avoid conflicts and controller overload.

[0023] 2. Improve network performance: dynamically adjust controller load, reduce response delay, and optimize resource utilization.

[0024] 3. Enhanced flexibility: Support differentiated QoS requirements and adapt to different application scenarios (such as cross-domain communication and wide-area perception) by adjusting weight parameters.

[0025] 4. High reliability: Master-slave controller redundant design and load redundancy mechanism improve network fault tolerance.

[0026] 5. Simplified management: The logical centralized architecture retains the simplicity of centralized management while supporting distributed expansion.

[0027] The above only describes the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the above specific implementation methods. Although the present invention has been disclosed as above in the preferred embodiments, it is not used to limit the present invention. Any technician familiar with the profession can make some changes or modify the technical contents disclosed above into equivalent embodiments with equivalent changes without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent replacement and improvement made to the above embodiments without departing from the content of the technical solution of the present invention, based on the technical essence of the present invention, within the spirit and principles of the present invention, still fall within the protection scope of the technical solution of the present invention.

Claims

1. A load balancing strategy method for air-sea cross-domain network control plane under SDN architecture, characterized in that: The following steps are involved: S1, divide the data plane into multiple subdomains, each subdomain contains several switches; S2. Allocate a master controller and multiple slave controllers to each subdomain. The master controller and the slave controllers are physically distributed on different devices and logically centrally managed. S3, monitor the controller load in real time. When the load of the main controller exceeds the threshold, generate a migration matrix based on the constraint conditions and dynamically adjust the control relationship between the switch and the controller; The step of dividing the subdomains in step S1 includes: Divide the data plane into subdomains, each containing Switches, corresponding to 1 main controller and A slave controller, which consists of a master slave controller and a secondary slave controller; the slave controller also serves as a secondary slave controller of the adjacent subdomain to achieve cross-domain redundant control; Based on the above control plane structure, the node set of each subdomain can be expressed as: , in, Indicates a switch. Indicates the main controller, Indicates the slave controller, then the entire air-sea cross-domain network has controllers, the collection is ,as well as switches, the set is ; and assume that any switch belongs to only one subdomain; The step of generating a migration matrix comprises: P1. Define the control matrix , , used to represent the control relationship between the controller and the switch, where When By controller control; P2. Calculate the migration matrix according to the constraints: , It represents the switch The corresponding controller can migrate state when When the master controller When the load reaches a critical value, the switch can migrate to .

2. According to the load balancing strategy method for the air-sea cross-domain network control plane adapted to the SDN architecture in claim 1, it is characterized in that: The method further comprises: Switch-based migration will cause additional load on the controller, so when setting the migration threshold, you need to leave a certain amount of redundancy compared to the maximum load of the controller. ,Right now: , in For controller Total load: , in, For Switch The load, For controller Capacity; The model can be obtained as follows: ① Solve the optimal migration matrix according to the constraints ,example: Constraints: , ， The parameters are as follows: ， ， ， ， ， Rt is the decision parameter, , , , , They are the relative transmission delay, relative queuing delay, packet loss rate, bandwidth utilization, and relative available bandwidth of the current link; ②The new control matrix after migration is: ; ③It should be noted that the controller load redundancy is always required during the calculation process, that is: , Each switch is assigned only one controller, namely: , And only 0 and 1 are used in the control matrix to represent the control relationship, that is: .

3. According to the load balancing strategy method for air-sea cross-domain network control plane under the SDN architecture adaptation according to claim 1, it is characterized in that: The method further comprises: In each subdomain, the master controller is allocated according to the standard that the controller with the largest maximum load capacity and the smallest average control path delay is the master controller of the subdomain, and the remaining controllers are allocated as the master-slave controllers of the subdomain and the secondary slave controllers of the adjacent subdomain according to the redundancy rules.

Images