CN117857369A - Hybrid cloud network intelligent system based on SDN technology and control method - Google Patents
Hybrid cloud network intelligent system based on SDN technology and control method Download PDFInfo
- Publication number
- CN117857369A CN117857369A CN202311722378.8A CN202311722378A CN117857369A CN 117857369 A CN117857369 A CN 117857369A CN 202311722378 A CN202311722378 A CN 202311722378A CN 117857369 A CN117857369 A CN 117857369A
- Authority
- CN
- China
- Prior art keywords
- network
- sdn
- node
- control node
- sdn control
- 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 64
- 238000005516 engineering process Methods 0.000 title claims abstract description 26
- 230000008859 change Effects 0.000 claims abstract description 54
- 206010047289 Ventricular extrasystoles Diseases 0.000 claims abstract description 25
- 238000005129 volume perturbation calorimetry Methods 0.000 claims abstract description 25
- 238000011084 recovery Methods 0.000 claims abstract description 14
- 238000007726 management method Methods 0.000 claims description 26
- 238000012508 change request Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000002159 abnormal effect Effects 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 12
- 101000884385 Homo sapiens Arylamine N-acetyltransferase 1 Proteins 0.000 claims description 11
- 101001034811 Homo sapiens Eukaryotic translation initiation factor 4 gamma 2 Proteins 0.000 claims description 11
- 101000639975 Homo sapiens Sodium-dependent noradrenaline transporter Proteins 0.000 claims description 11
- 102100033769 Sodium-coupled neutral amino acid transporter 3 Human genes 0.000 claims description 11
- 108020005224 Arylamine N-acetyltransferase Proteins 0.000 claims description 8
- 238000012986 modification Methods 0.000 claims description 7
- 230000004048 modification Effects 0.000 claims description 7
- 101000884399 Homo sapiens Arylamine N-acetyltransferase 2 Proteins 0.000 claims description 6
- 101000713305 Homo sapiens Sodium-coupled neutral amino acid transporter 1 Proteins 0.000 claims description 6
- 102100036916 Sodium-coupled neutral amino acid transporter 1 Human genes 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 6
- 238000004806 packaging method and process Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 4
- 108020005544 Antisense RNA Proteins 0.000 claims description 3
- 239000000470 constituent Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000000737 periodic effect Effects 0.000 claims description 3
- 230000002085 persistent effect Effects 0.000 claims description 3
- 230000002688 persistence Effects 0.000 abstract description 6
- 238000004321 preservation Methods 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000002955 isolation Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
- H04L41/122—Discovery or management of network topologies of virtualised topologies, e.g. software-defined networks [SDN] or network function virtualisation [NFV]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/40—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using virtualisation of network functions or resources, e.g. SDN or NFV entities
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention belongs to the technical field of big data and cloud computing, and particularly relates to a hybrid cloud network intelligent system and a control method based on SDN technology, wherein the system consists of an SDN master controller, distributed SDN control nodes, a configuration center and a network prediction center; the system realizes regional autonomy in different VPCs by selecting SDN control nodes nearby with the VPCs through the multi-stage SDN controller. The historical network topology structure is subjected to persistence preservation through the configuration center, so that disaster tolerance and recovery capacity of the system are improved, network conditions in a short period in the future are predicted by means of historical data, immediate and predictable network system change is realized, and flexibility and stability of the system are greatly improved. In addition, the invention also designs an intelligent control method of the hybrid cloud network based on the SDN technology, and the working states of the system under different scenes are perfectly controlled through three different sub-methods, so that the fluency and the safety of the system are further improved.
Description
Technical Field
The invention belongs to the technical field of big data and cloud computing, and particularly relates to a hybrid cloud network intelligent system and a control method based on SDN technology.
Background
According to different network architectures, cloud computing can be divided into three use forms of public cloud, private cloud and hybrid cloud. Public clouds are cloud computing services provided by third party providers that users can use as needed, which is advantageous in that it has a low cost and can flexibly expand or contract computing resources as needed. Private clouds refer to computing resources that are independently owned and managed by a single organization, which has the advantage of allowing an organization to fully master its data and computing environment and to be able to meet specific security and compliance requirements. The hybrid cloud absorbs the advantages of public cloud and private cloud, becomes a cloud-up preference gradually, is the combination between an internal data center of the private cloud and one or more public cloud resource pools, organically combines local facilities with third-party public cloud services, has the advantage of lower cost than the public cloud, and has the advantage of protecting highly sensitive information of the private cloud; the core of the hybrid cloud is embodied as how to manage the network connection between the local data center to the public cloud. However, since many different types of network devices and services are included in a hybrid cloud, management and control of its networks is also facing significant challenges. How to realize the intelligent control of the hybrid cloud network becomes an important problem in the current cloud computing field.
At present, a single machine or a distributed SDN controller is mainly deployed under the mixed cloud scene of other cloud manufacturers, network traffic among nodes is monitored, data forwarding and routing rules of a switch and a router are manually controlled through a north-south interface, when network equipment of two layers and three layers cannot determine a data forwarding path and a data forwarding mode, protocol messages are sent to the SDN controller, and the SDN controller sends the forwarding rules to specific equipment. Specifically, the controller may modify its flow table by sending a command to the switch via the OpenFlow protocol, thereby changing the forwarding path of the packet. In this process, the controller first needs to establish a connection with the switch and issue its own flow table to the switch so that the switch can forward the data packets according to these rules. And when a new data packet enters the switch, the switch matches the header information of the data packet with a flow table of the switch, and forwards the data packet to a corresponding port according to the operation appointed by the rule after finding out the corresponding rule.
The existing SDN hybrid cloud deployment and network control mode mainly has the following three defects:
1. the network control method of the SDN controller is single, and all messages need to be transmitted through the north-south interfaces of the SDN controller.
In the current method, the private cloud and the public cloud share the same SDN controller or SDN controller cluster, and the whole network is regulated and changed by the controller. When different areas of the network need to be regulated and controlled by different networks, the load pressure on the SDN controller is high, and the SDN controller directly controls the network across the VPCs, so that the isolation and safety requirements of each VPC in the hybrid cloud network architecture are reduced.
2. Lack of unified persistence of network history data and poor system stability and disaster recovery capability.
In the current SDN network architecture, flow tables and routing tables of each network node are usually stored by different network nodes or SDN controllers, a unified persistence scheme is not established, when a part of network nodes fail, the network is difficult to quickly reorganize, and the capacity of the SDN controller for processing network messages is also affected by storing the network messages in the SDN controller.
3. Lacking predictions of network conditions, network control messages have hysteresis.
In a network architecture based on an SDN controller, real-time control of a network is established on real-time communication between a network node and the SDN controller, the SDN controller can regulate and control the network only after a part of network nodes have network problems and notify the SDN controller, and the mode can not predict the condition of the network in a future short period or long period, and has hysteresis on the real-time control of the network.
The invention patent with the application publication number of CN115825027A discloses a hybrid cloud connection management method, an SDN controller and a hybrid cloud system, wherein the method comprises the steps that the SDN controller receives a hybrid cloud connection management request sent by a hybrid cloud management platform through a northbound interface; the SDN controller determines network equipment to be configured and network configuration information corresponding to the network equipment based on the hybrid cloud connection management request; the SDN controller encapsulates the network configuration information into a preset configuration protocol message and sends the preset configuration protocol message to network equipment needing configuration through a southbound interface.
The above prior art has the problems of the background art, and the network architecture of the single SDN controller proposed by the patent CN106936857a has poor disaster recovery capability and flexibility of processing network change requests due to the single SDN controller. In order to solve the problems, the invention discloses a hybrid cloud network intelligent system and a control method based on SDN technology.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a hybrid cloud network intelligent system and a control method based on SDN technology, which realize interconnection and intercommunication of multi-layer SDN control networks and automatic control and self-adaptive change of the hybrid cloud networks through an SDN main controller, a distributed SDN control node, a configuration center and a network prediction center, enhance the instant control and prediction capability of the networks and ensure the flexibility, stability and reliability of the networks.
In order to achieve the above purpose, the present invention provides the following technical solutions:
the hybrid cloud network intelligent system based on the SDN technology comprises an SDN main controller, a distributed SDN control node, a configuration center and a network prediction center;
the SDN master controller is used for overall management of the whole network and constructing a network topology structure;
the distributed SDN control node is used for issuing data statistics, flow table and routing table modification commands of each network node under the corresponding VPC1 and VPC 2;
the configuration center is used for persisting network control information such as routing rules, flow tables, network topological structures and the like;
and the network prediction center is used for predicting the network change condition in a short period in the future.
Specifically, the distributed SDN control nodes include SDN control node 1 and SDN control node 2, and SDN control node 1 corresponds to VPC1 and SDN control node 2 corresponds to VPC 2.
Specifically, the construction modules of the SDN master controller include, but are not limited to, a north interface module, a flow table management module, a network topology module, a statistics calculation module and a south interface module;
the SDN control node 1 comprises a north interface module, a flow table management module, a network topology module, a statistical calculation module and a south interface module;
the SDN control node 2 and the SDN control node 1 have the same constituent modules;
the north interface module is used for communicating with the SDN master controller;
the southbound interface module is used for communicating with the corresponding VPC1, VPC2, NAT1 and NAT 2; wherein SDN control node 1 corresponds to NAT1, SDN control node 2 corresponds to NAT 2;
NAT1, which is used for communication between each network node under VPC1 and Internet network; NAT2, which is used for communication between each network node under VPC2 and Internet network;
the flow table management module is used for maintaining flow rules stored on the switch, detecting network flow and issuing a command for changing and deleting the flow table;
the network topology module is used for generating a network topology structure diagram and a weight roadmap;
and the statistical calculation module is used for collecting and analyzing network operation data, carrying out statistical calculation and providing data support for network optimization and management.
Specifically, the specific communication modes of the SDN master controller, the SDN control node 1, the SDN control node 2, the configuration center and the network prediction center are as follows:
the SDN master controller is communicated with the configuration center, the network prediction center, the SDN control node 1 and the SDN control node 2;
the configuration center communicates with the SDN master controller, the SDN control node 1 and the SDN control node 2
The network prediction center only communicates with the SDN master controller;
SDN control node 1 communicates with SDN master controller, each network node under VPC1, NAT1 and configuration center;
the SDN control node 2 communicates with the SDN master controller, each network node under the VPC2, the NAT2 and the configuration center;
specifically, the specific workflows of the SDN master controller, SDN control node 1, SDN control node 2, configuration center and network prediction center are as follows:
each network node under the VPC1 and each network node under the VPC2 simultaneously send real-time network conditions to the corresponding SDN control node 1 and SDN control node 2 based on a nearby principle;
the SDN control node 1 and the SDN control node 2 send the network condition summary to the SDN master controller through a MapReduce process;
the SDN master controller gathers and transmits the network status information to the network prediction center, the network prediction center predicts the received network status based on a time-varying network prediction algorithm and transmits the prediction result to the SDN master controller,
and the SDN master controller determines the change condition of the configuration center according to the prediction result and completes the change of the network.
Specifically, the specific process of the network prediction center predicting the received network condition based on the time-varying network prediction algorithm includes:
a1: the network prediction center calculates and obtains a node state matrix at the time t+1 by inputting a similarity matrix, a weight vector of each side and a feature vector between nodes at the time t into a time-varying network prediction algorithm;
a2: and (3) obtaining a node state matrix at the time t+1 through calculation, and reconstructing a network topological structure and a weight route map.
The intelligent control method of the hybrid cloud network based on the SDN technology comprises a periodic automatic control sub-method based on time sequence characteristics, a network abnormal rapid recovery automatic control sub-method and a manual active control sub-method.
Specifically, the automatic cycle control sub-method based on the time sequence characteristic is a control method based on the normal working condition of the system, and specifically comprises the following steps:
b1: calculating the instantaneous flow Q by counting the message receiving and forwarding quantity PPS of each node of the network under the VPC1 and each node of the network under the VPC2, counting the total flow Q and the maximum instantaneous flow in a period according to a fixed period T, sequentially adding a flow table and a routing table of the total flow and the maximum instantaneous flow, sequentially packaging the total flow and the maximum instantaneous flow into protocol message bodies, and sending the protocol message bodies to SDN control nodes 1 and SDN control nodes 2 corresponding to the VPC1 and the VPC2 according to the period T;
b2: according to the information sent by the VPC1 and the VPC2, the SDN control node 1 and the SDN control node 2 establish a mapping relation by taking node_id of each network node under the VPC1 and the VPC2 as a key, count data of each network node as a value, count network conditions in the VPC1 and the VPC2, package the collected key value pair set into an information body and send the information body to the SDN main controller;
b3: after receiving the information of SDN control node 1 and SDN control node 2, SDN master controller executes Reduce process, establishes network topology structure and weight route map according to network condition in key value pair set, and sends the generated network topology structure and weight route map to network prediction center for predicting network condition;
b4: the network prediction center receives real-time network condition information from the SDN main controller, predicts the network conditions of short and long periods in the future, packages the generated network node prediction result set into a prediction result message and sends the prediction result message to the SDN main controller after the prediction is completed;
b5: the SDN master controller receives the prediction result information sent by the network prediction center, reconstructs a network topology map and a weight route map according to the state of each node at the next moment, modifies flow tables, routing tables and the like sent by SDN control nodes 1 and 2 according to comparison results, packages the flow tables, the routing tables and the like into network change request information, and sends the network change request information to the corresponding SDN control nodes under the VPC needing network topology modification;
b6: the SDN control node receives the network change request message from the SDN master controller, confirms whether the network node needing to be modified pointed out in the message exists under the corresponding VPC1 and VPC2, and forwards the message to the corresponding network node if the network node needing to be modified exists;
b7: the network node receives the network change message forwarded by the SDN control node, changes the flow table, the routing table and the like of the network node, and sends a change completion message to the SDN control node after the change is completed;
b8: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the weight route map and the flow table and the route table of each network node to the configuration center, and completes one-time network automation control and self-adaptive change.
Specifically, the automatic control sub-method for quickly recovering network abnormality is a recovery control method for when the flow condition of network nodes in VPC is suddenly changed, and comprises the following specific steps:
c1: the node with abnormal network flow monitors abnormal water level, actively triggers the message reporting, packages the network condition of the node as protocol message, marks the message type as alarm, and sends the message to SDN control node under the same VPC;
c2: SDN control node receives abnormal information of network, modifies the route rule of the correspondent network node according to self flow table management, topology management module, and send flow table, route table, network topology structure after modifying to the network node needing to modify;
and C3: the network node receives the network change message of the SDN control node, changes the flow table or the routing table of the network node, reconstructs the network topology structure, and sends a change completion message to the SDN control node after the change is completed;
and C4: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the changed weight route map, the flow table and the changed route table of each network node to the configuration center, and completes the automatic control of one-time network alarm.
Specifically, the manual active control sub-method is a method for manually checking network conditions and manually changing network topology through the southbound interface module or the northbound interface module provided by the SDN master controller, and the specific steps include:
d1: the south interface module or the north interface module of the SDN master controller receives a network change request message from an operator, the SDN master controller inquires a durable network topology structure and a weight route map in a configuration center, compares network nodes needing to be changed, and issues a network change request to corresponding SDN control nodes if configuration changes exist;
d2: SDN control nodes receive the network change request and forward the network change request to network nodes under corresponding VPCs;
d3: the network node receives the network change message sent by the corresponding SDN control node, changes the flow table, the routing rule and the like of the network node, and sends a change completion message to the corresponding SDN control node after the change is completed;
d4: and the corresponding SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the weight route map and the flow table and the route table of each network node to the configuration center, and completes one-time manual change of the network.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention integrates the defects of the traditional single SDN controller network, designs and realizes the hybrid cloud network intelligent system based on SDN, adopts a multi-level SDN network architecture, and increases the efficiency and flexibility of network control;
2. according to the invention, the configuration center is utilized to persistence the network history topological structure, so that the fault recovery capacity of each node of the network is improved, the prediction capacity of the network condition in a certain period in the future based on the history data is realized, the system stability is further improved, and the disaster recovery capacity of the system is improved;
3. the invention utilizes the network prediction center to infer the characteristic of the time sequence network topology structure according to the historical network topology structure, the server characteristic and the service logic, further avoids the hysteresis of network control and greatly increases the automatic control and self-adaptive capacity of the network system.
Drawings
Other features, objects and advantages of the present invention will become more apparent upon reading of the detailed description of non-limiting embodiments, made with reference to the following drawings in which:
fig. 1 is a diagram of a hybrid cloud network intelligent system structure based on SDN technology in embodiment 1 of the present invention;
FIG. 2 is a flowchart of a cycle automatic control sub-method based on a time sequence feature according to embodiment 2 of the present invention;
fig. 3 is a flowchart of a network anomaly fast recovery automatic control sub-method according to embodiment 2 of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Example 1:
most of hybrid cloud network intelligent systems adopting traditional SDN technology adopt a single SDN controller, which causes the system to lack unified persistence of network history data and prediction capability of network conditions, and also causes the flexibility of the system to be poor when processing network change requests.
To solve the above problems, please refer to fig. 1, an embodiment of the present invention is provided: the hybrid cloud network intelligent system based on the SDN technology comprises an SDN main controller, a distributed SDN control node, a configuration center and a network prediction center;
the SDN master controller is used for overall management of the whole network and constructing a network topology structure;
the distributed SDN control node is used for issuing data statistics, flow table and routing table modification commands of each network node under the corresponding VPC1 and VPC 2; the VPC represents a virtual private cloud, is a customized logic isolation network space on a public cloud, is a network space which can be customized by a user, and is replaced by the abbreviation VPC in the later description.
And the configuration center is used for persisting network control information such as routing rules, flow tables, network topology structures and the like.
And the network prediction center is used for predicting the network change condition in a short period in the future.
Further, the distributed SDN control nodes include SDN control node 1 and SDN control node 2, where SDN control node 1 corresponds to VPC1 and SDN control node 2 corresponds to VPC 2.
Further, the construction modules of the SDN master controller include, but are not limited to, a north interface module, a flow table management module, a network topology module, a statistics calculation module and a south interface module;
the SDN control node 1 comprises a north interface module, a flow table management module, a network topology module, a statistical calculation module and a south interface module;
the SDN control node 2 and the SDN control node 1 have the same constituent modules;
the north interface module is used for communicating with the SDN master controller and sending network statistical information and the like;
the southbound interface module is used for communicating with the corresponding VPC1, VPC2, NAT1 and NAT 2; wherein SDN control node 1 corresponds to 1 and SDN control node 2 corresponds to NAT 2. The NAT is an abbreviation of address translation, and is a device capable of sharing a plurality of computers on a public network to an Internet for connection and conversion, and is replaced by the abbreviation NAT in the later full text;
NAT1, which is used for communication between each network node under VPC1 and Internet network; NAT2, which is used for communication between each network node under VPC2 and Internet network; wherein, each network node is jointly realized by a gateway and a VxLan network virtualization technology, and VPC1 and VPC2 are mutually connected by the gateway and the VxLan to realize complete network intercommunication in the system; in addition, the VxLan is a network virtualization technology and is used for improving the expansion problem of large-scale cloud computing in deployment;
the flow table management module is used for maintaining flow rules stored on the switch, detecting network flow and issuing a command for changing and deleting the flow table;
the network topology module is used for generating a network topology structure diagram and a weight roadmap;
and the statistical calculation module is used for collecting and analyzing network operation data, carrying out statistical calculation and providing data support for network optimization and management.
Further, the specific communication modes of the SDN master controller, the SDN control node 1, the SDN control node 2, the configuration center and the network prediction center are as follows:
the SDN master controller is communicated with the configuration center, the network prediction center, the SDN control node 1 and the SDN control node 2;
the configuration center communicates with the SDN master controller, the SDN control node 1 and the SDN control node 2
The network prediction center only communicates with the SDN master controller;
SDN control node 1 communicates with SDN master controller, each network node under VPC1, NAT1 and configuration center;
the SDN control node 2 communicates with the SDN master controller, each network node under the VPC2, the NAT2 and the configuration center; through the communication connection, interconnection and interworking of the multi-layer SDN control network are realized;
the communication modes among the SDN master controller, the SDN control node 1, the SDN control node 2, the configuration center and the network prediction center are realized based on an OpenFlow protocol or a custom message, and the message header is exemplified as follows:
struct msg_header{
uint8_tversion；
uint8_ttype；
uint16_t length；
uint32_txid；
}
where version represents the current protocol message version, type represents the message type for distinguishing each component communication type, length represents the total message length for maintaining buffer contents, xid represents the corresponding transaction id of the request for distinguishing different request response sequences. The message body is determined based on the content of the message actually sent by each component.
Further, the specific workflow of the SDN master controller, SDN control node 1, SDN control node 2, configuration center and network prediction center is as follows:
each network node under the VPC1 and each network node under the VPC2 simultaneously send real-time network conditions to the corresponding SDN control node 1 and SDN control node 2 based on a nearby principle;
the SDN control node 1 and the SDN control node 2 send the network condition summary to the SDN master controller through a MapReduce process;
the SDN master controller gathers and transmits the network state information to the network prediction center, and the network prediction center predicts the received network state based on a time-varying network prediction algorithm and transmits a prediction result to the SDN master controller;
and the SDN master controller determines the change condition of the configuration center according to the prediction result and completes the change of the network.
Further, the specific process of the network prediction center predicting the received network condition based on the time-varying network prediction algorithm includes:
a1: the network prediction center calculates and obtains a node state matrix at the time t+1 by inputting a similarity matrix, a weight vector of each side and a feature vector between nodes at the time t into a time-varying network prediction algorithm;
a2: and (3) obtaining a node state matrix at the time t+1 through calculation, and reconstructing a network topological structure and a weight route map.
The system realizes regional autonomy in different VPCs by selecting SDN control nodes nearby with the VPCs through a multi-stage SDN controller; the historical network topology structure is subjected to persistence preservation through the configuration center, so that disaster tolerance and recovery capacity of the system are improved, network conditions in a short period in the future are predicted by means of historical data, immediate and predictable network system change is realized, and flexibility and stability of the system are greatly improved.
Example 2:
referring to fig. 2, the present invention provides an embodiment: an intelligent control method for a hybrid cloud network based on SDN technology specifically comprises three methods: a periodic automatic control sub-method based on time sequence characteristics, a network abnormal rapid recovery automatic control sub-method and a manual active control sub-method.
Further, the automatic cycle control sub-method based on the time sequence characteristic is a control method based on the normal working condition of the system, and specifically comprises the following steps:
b1: calculating the instantaneous flow Q by counting the message receiving and forwarding quantity PPS of each node of the network under the VPC1 and each node of the network under the VPC2, counting the total flow Q and the maximum instantaneous flow in a period according to a fixed period T (a constant value can be set manually or can be changed according to a monitoring water level), sequentially adding own flow table and routing table to the total flow and the maximum instantaneous flow, sequentially packaging the total flow and the maximum instantaneous flow into protocol message bodies, and sending the protocol message bodies to SDN control nodes 1 and SDN control nodes 2 corresponding to the VPC1 and the VPC2 according to the period T; examples of message bodies are as follows:
struct node_info{
uint32_tnode_id；
uint8_tnode_qm；
uint16_tnode_Q；
uint8_tnode_type；
struct flow_entry*fe；
struct route_entry*re；
}；
where node_id represents the unique id of each network node, node_qm represents the maximum instantaneous traffic, node_q represents the total traffic in period T, node_type represents the node type, which may be a switch or a router, etc., fe is the pointer to the flow table structure, re is the pointer to the routing table structure. flow_entry and route_entry structures are exemplified as follows:
wherein flow_entry represents a flow table item, match_fields represents a matching field, which contains packet header information such as source address, destination address, port number, etc. to be matched, actions represents actions to be executed, priority represents priority, and flow table items with high priority are preferably matched and executed;
b2: according to the information sent by the VPC1 and the VPC2, the SDN control node 1 and the SDN control node 2 establish a mapping relation by taking node_id of each network node under the VPC1 and the VPC2 as a key, count data of each network node as a value, count network conditions in the VPC1 and the VPC2, package the collected key value pair set into an information body and send the information body to the SDN main controller; the mapping procedure described above is as follows:
M(x i )→(k i ,v i )，
wherein k is i Key representing node_id construction of i-th network node, v i Representing the value of the ith statistical data construct, M (x i ) Information representing the collection of the ith network node;
b3: after receiving the messages of each SDN control node, the SDN master controller executes a Reduce process, establishes a network topology structure and a weight route diagram according to the network conditions in the key value pair set, executes the Reduce process after receiving the messages of the SDN control node 1 and the SDN control node 2, and establishes the network topology structure and the weight route diagram according to the network conditions in the key value pair set, as follows:
R({(k 1 ,v 1 ),(k 2 ,v 2 ),(k 3 ,v 3 ),…})→u t ，
wherein R represents a Reduce mapping process, u t Representing a network topology and a weight roadmap at a time t, comprising a set of nodes V at time t t Edge set E t Edge weight set W t The method can be realized through an adjacency matrix or an adjacency table, and then the generated network topology structure and weight roadmap are sent to a network prediction center for predicting network conditions;
b4: the network prediction center receives real-time network condition information from the SDN main controller, predicts the network conditions of short and long periods in the future, and firstly calculates the similarity of any two nodes i and j as follows:
wherein s is ij (t) Representing similarity of node i and node j at time t, V i (t) And V j (t) Respectively representing the neighbor node sets of node i and node j at time t. Similar calculation sectionThe point similarity matrix is as follows:
wherein N represents the number of nodes, S t Representing a node similarity matrix.
The node weight vector is calculated as follows:
wherein W is t Node weight vector, w i (t) Representing the weight of node i at time t, and then computing a node feature vector, as follows:
wherein X is t Calculating node feature vector, x i (t) The characteristic vector of the node i at the time t is represented, and the state of the node i at the next moment can be calculated according to the characteristic vector, the similarity and the weight, as follows:
after the prediction is completed, packaging the generated prediction result set of each network node into a prediction result message, and sending the prediction result message to the SDN master controller;
b5: the SDN master controller receives the prediction result information sent by the network prediction center, and reconstructs a network topology map and a weight route map u according to the state of each node at the next moment t+1 Modifying flow tables, routing tables and the like sent by SDN control nodes 1 and 2 by the comparison result, packaging the flow tables, the routing tables and the like as network change request messages, and transmitting the network change request messages to corresponding SDN control nodes under the VPC needing network topology structure modification;
b6: the SDN control node receives the network change request message from the SDN master controller, confirms whether the network node needing to be modified pointed out in the message exists under the corresponding VPC1 and VPC2, and forwards the message to the corresponding network node if the network node needing to be modified exists;
b7: the network node receives the network change message forwarded by the SDN control node, changes the flow table, the routing table and the like of the network node, and sends a change completion message to the SDN control node after the change is completed;
b8: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the weight route map and the flow table and the route table of each network node to the configuration center, and completes one-time network automation control and self-adaptive change.
Further, referring to fig. 2, in some cases, when a traffic condition of a part of network nodes in VPC1 or VPC2 is suddenly changed (suddenly increased or reduced), the embodiment provides a method for automatically controlling abnormal recovery of a network, which specifically includes the following steps:
c1: the node with abnormal network flow monitors abnormal water level, actively triggers the message reporting, packages the network condition of the node as protocol message, marks the message type as alarm, and sends the message to SDN control node under the same VPC;
c2: SDN control node receives abnormal information of network, modifies the route rule of the correspondent network node according to self flow table management, topology management module, and send flow table, route table, network topology structure after modifying to the network node needing to modify;
and C3: the network node receives the network change message of the SDN control node, changes the flow table or the routing table of the network node, reconstructs the network topology structure, and sends a change completion message to the SDN control node after the change is completed;
and C4: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the changed weight route map, the flow table and the changed route table of each network node to the configuration center, and completes the automatic control of one-time network alarm.
By the control method, abnormal points can be rapidly identified, the existing problems are solved, the fluency and safety of the system can be ensured, and the execution efficiency of the system is improved.
Further, in order to further enhance the operability of the system, the embodiment provides a manual active control sub-method, which is a method for manually checking the network condition and manually changing the network topology through the southbound interface module or the northbound interface module provided by the SDN master controller, and specifically includes the steps of:
d1: the south interface module or the north interface module of the SDN master controller receives a network change request message from an operator, the SDN master controller inquires a durable network topology structure and a weight route map in a configuration center, compares network nodes needing to be changed, and issues a network change request to corresponding SDN control nodes if configuration changes exist;
d2: SDN control nodes receive the network change request and forward the network change request to network nodes under corresponding VPCs;
d3: the network node receives the network change message sent by the corresponding SDN control node, changes the flow table, the routing rule and the like of the network node, and sends a change completion message to the corresponding SDN control node after the change is completed;
d4: and the corresponding SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the weight route map and the flow table and the route table of each network node to the configuration center, and completes one-time manual change of the network.
The method can improve the controllability of an operator to the system, so that the system can better cooperate with the operator to finish specific tasks, and the integrity and the intelligence of the overall control method are further enhanced.
While embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the invention.
Claims (10)
1. The hybrid cloud network intelligent system based on the SDN technology is characterized by comprising an SDN master controller, a distributed SDN control node, a configuration center and a network prediction center;
the SDN master controller is used for overall management of the whole network and constructing a network topology structure;
the distributed SDN control node is used for issuing data statistics, flow table and routing table modification commands of each network node under the corresponding VPC1 and VPC 2;
the configuration center is used for persisting network control information such as routing rules, flow tables, network topological structures and the like;
the network prediction center is used for predicting network change conditions in a short period in the future.
2. The hybrid cloud network intelligent system based on SDN technology of claim 1, wherein the distributed SDN control nodes include SDN control node 1 and SDN control node 2;
the SDN control node 1 corresponds to the VPC1, and the SDN control node 2 corresponds to the VPC 2.
3. The hybrid cloud network intelligent system based on SDN technology of claim 2, wherein the SDN master controller building modules include, but are not limited to, a north interface module, a flow table management module, a network topology module, a statistics calculation module and a south interface module;
the SDN control node 1 comprises a north interface module, a flow table management module, a network topology module, a statistical calculation module and a south interface module;
the SDN control node 2 and the SDN control node 1 have the same constituent modules;
the north interface module is used for communicating with the SDN master controller;
the southbound interface module is used for communicating with corresponding VPC1, VPC2, NAT1 and NAT 2; wherein, the SDN control node 1 corresponds to NAT1, and the SDN control node 2 corresponds to NAT 2;
the NAT1 is used for communication between each network node under the VPC1 and an Internet network;
the NAT2 is used for communication between each network node under the VPC2 and the Internet;
the flow table management module is used for maintaining flow rules stored on the switch, detecting network flow and issuing a command for changing and deleting the flow table;
the network topology module is used for generating a network topology structure diagram and a weight roadmap;
the statistical calculation module is used for collecting and analyzing network operation data, carrying out statistical calculation and providing data support for network optimization and management.
4. The hybrid cloud network intelligent system based on SDN technology as set forth in claim 3, wherein the specific communication modes of the SDN master controller, SDN control node 1, SDN control node 2, configuration center and network prediction center are as follows:
the SDN master controller is communicated with the configuration center, the network prediction center, the SDN control node 1 and the SDN control node 2;
the configuration center is communicated with the SDN master controller, the SDN control node 1 and the SDN control node 2;
the network prediction center only communicates with the SDN master controller;
the SDN control node 1 communicates with the SDN master controller, each network node under the VPC1, the NAT1 and the configuration center;
the SDN control node 2 communicates with the SDN master controller, each network node under the VPC2, the NAT2 and the configuration center.
5. The hybrid cloud network intelligent system based on SDN technology of claim 4, wherein the specific workflows of the SDN master controller, SDN control node 1, SDN control node 2, configuration center and network prediction center are as follows:
each network node under the VPC1 and each network node under the VPC2 simultaneously send real-time network conditions to the corresponding SDN control node 1 and SDN control node 2 based on a nearby principle;
the SDN control node 1 and the SDN control node 2 send the network condition summary to the SDN master controller through a MapReduce process;
the SDN master controller gathers and transmits the network status information to the network prediction center, and the network prediction center predicts the received network status based on a time-varying network prediction algorithm and transmits a prediction result to the SDN master controller;
and the SDN master controller determines the change condition of the configuration center according to the prediction result and completes the change of the network.
6. The hybrid cloud network intelligent system based on SDN technology of claim 5, wherein the specific process of predicting the received network condition by the network prediction center based on a time-varying network prediction algorithm includes:
a1: the network prediction center calculates and obtains a node state matrix at the time t+1 by inputting a similarity matrix, a weight vector of each side and a feature vector between nodes at the time t into a time-varying network prediction algorithm;
a2: and (3) obtaining a node state matrix at the time t+1 through calculation, and reconstructing a network topological structure and a weight route map.
7. An intelligent control method of a hybrid cloud network based on an SDN technology, which is implemented based on an intelligent system of the hybrid cloud network based on the SDN technology as set forth in any one of claims 1-6, wherein the control method includes a periodic automatic control sub-method based on a timing characteristic, a network anomaly fast recovery automatic control sub-method, and a manual active control sub-method.
8. The intelligent control method of the hybrid cloud network based on the SDN technology of claim 7, wherein the automatic cycle control sub-method based on the time sequence features is a control method based on the normal working condition of the system, and the specific steps comprise:
B1：VPC1calculating the instantaneous flow Q by counting the message receiving and forwarding quantity PPS of each node of the lower network and each node of the VPC2 lower network, and counting the total flow Q and the maximum instantaneous flow Q in a period according to a fixed period T max Sequentially adding a flow table and a routing table of the total flow and the maximum instantaneous flow, sequentially packaging the flow table and the routing table into a protocol message body, and sending the protocol message body to SDN control nodes 1 and 2 corresponding to VPC1 and VPC2 according to a period T;
b2: according to the information sent by the VPC1 and the VPC2, the SDN control node 1 and the SDN control node 2 establish a mapping relation by taking node_id of each network node under the VPC1 and the VPC2 as a key, count data of each network node as a value, count network conditions in the VPC1 and the VPC2, package the collected key value pair set into an information body and send the information body to the SDN main controller;
b3: after receiving the information of SDN control node 1 and SDN control node 2, SDN master controller executes Reduce process, establishes network topology structure and weight route map according to network condition in key value pair set, and sends the generated network topology structure and weight route map to network prediction center for predicting network condition;
b4: the network prediction center receives real-time network condition information from the SDN main controller, predicts the network conditions of short and long periods in the future, packages the generated network node prediction result set into a prediction result message and sends the prediction result message to the SDN main controller after the prediction is completed;
b5: the SDN master controller receives the prediction result information sent by the network prediction center, and reconstructs a network topology map and a weight route map u according to the state of each node at the next moment t+1 Modifying flow tables, routing tables and the like sent by SDN control nodes 1 and 2 by the comparison result, packaging the flow tables, the routing tables and the like as network change request messages, and transmitting the network change request messages to corresponding SDN control nodes under the VPC needing network topology structure modification;
b6: the SDN control node receives the network change request message from the SDN master controller, confirms whether the network node needing to be modified pointed out in the message exists under the corresponding VPC1 and VPC2, and forwards the message to the corresponding network node if the network node needing to be modified exists;
b7: the network node receives the network change message forwarded by the SDN control node, changes the flow table, the routing table and the like of the network node, and sends a change completion message to the SDN control node after the change is completed;
b8: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the weight route map and the flow table and the route table of each network node to the configuration center, and completes one-time network automation control and self-adaptive change.
9. The hybrid cloud network intelligent control method based on the SDN technology as set forth in claim 8, wherein the network anomaly fast recovery automatic control sub-method is a recovery control method for when a network node traffic condition in a VPC is suddenly changed, and the specific steps include:
c1: the node with abnormal network flow monitors abnormal water level, actively triggers the message reporting, packages the network condition of the node as protocol message, marks the message type as alarm, and sends the message to SDN control node under the same VPC;
c2: SDN control node receives abnormal information of network, modifies the route rule of the correspondent network node according to self flow table management, topology management module, and send flow table, route table, network topology structure after modifying to the network node needing to modify;
and C3: the network node receives the network change message of the SDN control node, changes the flow table or the routing table of the network node, reconstructs the network topology structure, and sends a change completion message to the SDN control node after the change is completed;
and C4: and the SDN control node receives the change completion message sent by the network node, persists the changed network topology structure, the changed weight route map, the flow table and the changed route table of each network node to the configuration center, and completes the automatic control of one-time network alarm.
10. The hybrid cloud network intelligent control method based on SDN technology of claim 9, wherein the manual active control sub-method is a method for manually checking network conditions and manually changing network topology through the southbound interface module or the northbound interface module provided by the SDN master controller, and the specific steps include:
d1: the south interface module or the north interface module of the SDN master controller receives a network change request message from an operator, the SDN master controller inquires a durable network topology structure and a weight route map in a configuration center, compares network nodes needing to be changed, and issues a network change request to corresponding SDN control nodes if configuration changes;
d2: SDN control nodes receive the network change request and forward the network change request to network nodes under corresponding VPCs;
d3: the network node receives the network change message sent by the corresponding SDN control node, changes the flow table, the routing rule and the like of the network node, and sends a change completion message to the corresponding SDN control node after the change is completed;
d4: and the corresponding SDN control node receives the change completion message sent by the network node, and persists the changed network topology structure, the changed weight route map, the flow table and the changed route table of each network node to a configuration center to complete one-time manual network change.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311722378.8A CN117857369A (en) | 2023-12-14 | 2023-12-14 | Hybrid cloud network intelligent system based on SDN technology and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311722378.8A CN117857369A (en) | 2023-12-14 | 2023-12-14 | Hybrid cloud network intelligent system based on SDN technology and control method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117857369A true CN117857369A (en) | 2024-04-09 |
Family
ID=90537530
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311722378.8A Pending CN117857369A (en) | 2023-12-14 | 2023-12-14 | Hybrid cloud network intelligent system based on SDN technology and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117857369A (en) |
-
2023
- 2023-12-14 CN CN202311722378.8A patent/CN117857369A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230362249A1 (en) | Systems and methods for routing data to a parallel file system | |
| CN103179046B (en) | Based on data center's flow control methods and the system of openflow | |
| CN103516782B (en) | The distributed treatment of network equipment task | |
| CN108540876A (en) | Service path choosing method, SDN controllers, storage medium and electronic equipment | |
| CN106470155B (en) | Virtual machine data forwarding method, SDN controller and SDN system | |
| CN103795805A (en) | Distributed server load balancing method based on SDN | |
| CN104303467A (en) | Integrated heterogeneous software-defined network | |
| CN104272656A (en) | Network feedback in software-defined networks | |
| CN102770852A (en) | Information and communication processing system, method, and network node | |
| CN103763367A (en) | Method and system for designing distributed virtual network in cloud calculating data center | |
| CN105376154A (en) | Progressive MAC address learning | |
| CN103069756A (en) | Communication system, controller, node controlling method and program | |
| CN106209615B (en) | A kind of Dynamic Route Control method and system calculating forward-path based on SPFA algorithm | |
| CN105391635A (en) | Network virtualization method based on software defined network (SDN) | |
| CN105052113A (en) | Common agent framework for network devices | |
| CN104662838A (en) | Transmit specific traffic along blocked link | |
| CN111147372B (en) | Downlink message sending and forwarding method and device | |
| CN104065553B (en) | Virtual network moving method and relevant device | |
| JP2008042665A (en) | Network virtualization apparatus and network virtualization program | |
| Tadesse et al. | Energy-efficient traffic allocation in SDN-basec backhaul networks: Theory and implementation | |
| Gadasin et al. | Routing Management System Formation for Machine-to-Machine Interaction in a Decentralized Environment | |
| CN109218181A (en) | A kind of method, apparatus and electronic equipment of data transmission | |
| KR101586474B1 (en) | Apparatus and method for openflow routing | |
| EP3399424B1 (en) | Using unified api to program both servers and fabric for forwarding for fine-grained network optimizations | |
| US6301252B1 (en) | Control and management of ATM networks from server computers |
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 |