CN111988220B - Multi-target disaster backup method and system among data centers based on reinforcement learning - Google Patents

Multi-target disaster backup method and system among data centers based on reinforcement learning Download PDF

Info

Publication number
CN111988220B
CN111988220B CN202010820767.4A CN202010820767A CN111988220B CN 111988220 B CN111988220 B CN 111988220B CN 202010820767 A CN202010820767 A CN 202010820767A CN 111988220 B CN111988220 B CN 111988220B
Authority
CN
China
Prior art keywords
backup
data
routing
disaster
link
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.)
Expired - Fee Related
Application number
CN202010820767.4A
Other languages
Chinese (zh)
Other versions
CN111988220A (en
Inventor
王�华
燕嘉鑫
伊善文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202010820767.4A priority Critical patent/CN111988220B/en
Publication of CN111988220A publication Critical patent/CN111988220A/en
Application granted granted Critical
Publication of CN111988220B publication Critical patent/CN111988220B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/22—Alternate routing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/24—Multipath
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/28—Routing or path finding of packets in data switching networks using route fault recovery
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The utility model discloses a multi-target disaster backup method and system among data centers based on reinforcement learning, which comprises: expanding the network among the data centers to obtain a time expansion network; acquiring data to be backed up in a time expansion network; inputting data to be backed up into a backup routing selection model to obtain an optimal backup routing scheme; transmitting backup of data to be backed up in a time expansion network by an optimal backup routing scheme; the backup routing selection model comprises a fitness function of each link in the multicast tree in the time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load. When disaster backup is carried out, the factors of the backup cost and the load balance are fully considered, a backup routing scheme which is better in the backup cost and the load balance is obtained through a backup routing selection model, and the maximum link congestion is relieved on the basis of reducing broadband waste.

Description

Multi-target disaster backup method and system among data centers based on reinforcement learning
Technical Field
The disclosure relates to a reinforcement learning-based multi-objective disaster backup method and system between data centers.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
In recent years, many large enterprises, such as amazon, google, microsoft, have deployed large data centers in multiple geographic locations to provide various services to millions of users around the globe. Due to natural disasters, artificial damages and the like, data security also draws more and more attention of people. In order to realize data redundancy and ensure data safety, the data from the TB to the PB needs to be periodically copied in a data center-to-data center network and distributed to three or more other remote data centers, namely disaster backup.
At present, for redundant disaster backup, most researches adopt multicast routing to reduce backup bandwidth consumption, but most of the researches do not consider the load balancing problem of network links, which easily causes that when local some links are seriously congested, daily service of a data center is also seriously influenced.
Disclosure of Invention
The invention provides a multi-target disaster backup method and a multi-target disaster backup system among data centers based on reinforcement learning, which fully consider the factors of backup cost and load balance when carrying out disaster backup, and obtain a backup routing scheme which is better in both the backup cost and the load balance through a backup routing selection model, thereby relieving the maximum link congestion on the basis of reducing broadband waste.
In order to achieve the purpose, the following technical scheme is adopted in the disclosure:
in one or more embodiments, a reinforcement learning-based multi-objective disaster backup method between data centers is provided, which includes:
expanding the network among the data centers to obtain a time expansion network;
acquiring data to be backed up in a time expansion network;
inputting data to be backed up into a backup routing selection model to obtain an optimal backup routing scheme;
transmitting backup of data to be backed up in a time expansion network by an optimal backup routing scheme;
the backup routing selection model comprises a fitness function of each link in a multicast tree in a time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load.
In one or more embodiments, a reinforcement learning-based inter-data center multi-objective disaster backup system is provided, including:
the acquisition module acquires data to be backed up;
the backup routing selection model comprises a fitness function of each link in a multicast tree in the time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load;
and the calculation module is used for inputting the data to be backed up into the backup routing selection model to obtain the optimal backup routing scheme.
In one or more embodiments, a computer-readable storage medium stores computer instructions that, when executed by a processor, perform the steps of the reinforcement learning-based inter-data-center multi-objective disaster backup method.
Compared with the prior art, the beneficial effect of this disclosure is:
1. the method and the device construct a plurality of disaster backup multicast trees by using factors such as sharing degree of the multicast trees, thereby reducing bandwidth waste; sending data in different time slots by using a store-and-forward mechanism so as to relieve the congestion of the maximum link; and continuously constructing a routing path by utilizing multi-objective reinforcement learning, thereby obtaining a backup routing scheme which is better in both backup cost and load balancing.
2. The optimal backup routing scheme of the data to be backed up is calculated through a backup routing selection model, wherein the backup routing selection model comprises a fitness function of each link in a multicast tree in a time expansion network to the multicast tree and a congestion factor function of each link, and when the optimal backup routing scheme is solved by taking the minimized backup cost and the link load balance as targets, the two factors of the backup cost and the load balance are fully considered, so that the maximum link congestion is relieved on the basis of reducing broadband waste by the obtained optimal backup routing scheme.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application.
Fig. 1 is a flow chart of example 1 of the present disclosure;
fig. 2 is an example of backup routing in the time-expanding network in embodiment 1 of the present disclosure.
The specific implementation mode is as follows:
the present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
In the present disclosure, terms such as "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", "side", "bottom", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only relational terms determined for convenience in describing structural relationships of the parts or elements of the present disclosure, and do not refer to any parts or elements of the present disclosure, and are not to be construed as limiting the present disclosure.
In the present disclosure, terms such as "fixedly connected", "connected", and the like are to be understood in a broad sense, and mean either a fixed connection or an integrally connected or detachable connection; may be directly connected or indirectly connected through an intermediate. The specific meanings of the above terms in the present disclosure can be determined on a case-by-case basis by persons skilled in the relevant art or technicians, and are not to be construed as limitations of the present disclosure.
Example 1
At present, for redundant disaster backup, most researches adopt multicast routing to reduce backup bandwidth consumption, but most of the researches do not consider the load balancing problem of network links, which easily causes that when local some links are seriously congested, daily service of a data center is also seriously influenced. And the problem of link congestion can be better solved by applying a store-and-forward mechanism in the time expansion network, so that link load balancing is achieved. Due to the rise of the software defined network, the traffic can be explicitly routed and scheduled in the software defined network, so that the traffic scheduling can be more flexibly performed.
In the embodiment, a multi-target disaster backup method among data centers based on reinforcement learning is disclosed, wherein in a time expansion network after the network among the data centers is expanded, a multicast routing and store-and-forward mechanism is used for transmitting backup data, so that the minimum total backup cost and load balance are realized; a backup routing scheme which is excellent in both backup cost and load balance is obtained by utilizing a multicast backup multi-target reinforcement learning algorithm, and the backup routing scheme comprises the following steps:
expanding the network among the data centers to obtain a time expansion network;
acquiring data to be backed up in a time expansion network;
inputting data to be backed up into a backup routing selection model to obtain an optimal backup routing scheme;
transmitting backup of data to be backed up in a time expansion network by an optimal backup routing scheme;
the backup routing selection model comprises a fitness function of each link in a multicast tree in a time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load.
The data center regularly performs disaster backup of data for preventing disasters and ensuring data safety, and in order to ensure data redundancy, the same data needs to perform at least three redundant backups in data centers in different geographic positions.
In the time expansion network after the network among the data centers is expanded, the multicast routing and the storage forwarding mechanism are used for transmitting the data to be backed up, so that the minimum total backup cost and load balance are realized; and obtaining a backup routing scheme which is better in both backup cost and load balance by utilizing a multicast backup multi-target reinforcement learning algorithm.
The time topology network is composed of network topologies among the same data centers of a plurality of different time slots and converts a dynamic network flow problem into a static flow problem; the edge of the topology after the network topology is copied among the data centers is a transmission edge, and the edge connecting the same node in the adjacent time slots is a reserved edge.
The time slot division is related to the total time length of disaster backup, the total data volume of backup and the network link bandwidth, so that the flow of different time slots can be scheduled as fine as possible in bearable running time.
The data to be backed up is routed in a multicast mode, each backup requirement corresponds to a disaster backup multicast tree, the starting point of the disaster backup multicast tree is a source data center with backup requirements, the end points of the disaster backup multicast tree are a plurality of backup data centers with enough storage space, and the routing path is the routing path of the data to be backed up;
all backup requirements correspond to a disaster backup multicast forest comprised of all disaster backup multicast trees.
The multicast routing is carried out in a time expansion network, the time expansion network simultaneously contains the spatial information and the time information of the routing, and with the help of a software defined network, the routing can be explicitly scheduled in different time slots.
In order to enable data to be backed up to be transmitted and backed up in a time expansion network according to an optimal backup routing scheme, a backup routing selection model is constructed, the optimal backup routing scheme is obtained through the backup routing selection model, and the scheme fully considers the factors of backup cost and load balance, so that when the data are transmitted and backed up according to the optimal backup routing scheme, broadband waste is reduced, and the maximum link congestion is relieved.
The backup routing selection model comprises a fitness function of each link in a multicast tree in a time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is solved by taking minimized backup cost and link load balance as targets.
When the backup data selects the next hop in the routing path, the congestion degree of the link, the bandwidth cost and the multicast tree sharing degree need to be considered.
The data center as a routing node can temporarily store data, and the data can be sent when a link of the next hop is idle.
And independently generating a disaster backup multicast tree for each backup requirement, and selecting a routing node by using an epsilon greedy strategy, so that the disaster backup multicast forests corresponding to all the backup requirements obtain a solution set with higher quality on two targets of backup cost and load balance.
And calculating and solving a backup routing selection model by using an epsilon greedy strategy, obtaining each routing node in the disaster backup multicast tree, forming multicast tree paths by each routing node, and forming an optimal backup routing scheme by all the multicast tree paths.
When selecting the next routing node, the congestion degree of the link, the bandwidth cost and the multicast tree sharing degree need to be considered. Multicast tree sharing degree tsd (t)iAnd e) defining a link e in a multicast tree tiDegree of polymerization of (1). The higher the sharing degree of the multicast tree is, the higher the sharing degree of the multicast tree isiThe higher the polymerization degree in (1), the more bandwidth is saved, and the fitness (t) is usediAnd e) to represent the link e to the multicast tree tiThe fitness of the link e is used as a comprehensive standard for judging whether the link e is selected. Fitness is defined as follows:
Figure BDA0002634343470000071
parameter alpha1And alpha2Is the weight value of the unit bandwidth cost of the link e and the multicast tree sharing degree function. Variables avgcost and avgtsd represent multicast tree t, respectivelyiAverage unit bandwidth cost and average multicast tree sharing degree of the medium link.
Defining a link congestion factor mueThe following were used:
Figure BDA0002634343470000081
wherein f iseCurrent traffic for link e, beFor the maximum capacity of link e, both may get data from the controller of the SDN.
Using vectors
Figure BDA0002634343470000083
To represent the reward factor, using fitness (t)iE) and a link congestion factor mueAs optimization targets for minimizing backup costs and load balancing, respectively, i.e.
Figure BDA0002634343470000084
In particular, the total cost is reduced by copying the backup data as late as possible by walking the common link as much as possible. After the reward factor is determined, the iterative updating formula of the Q table is as follows:
Figure BDA0002634343470000082
wherein alpha isi∈(0,1]Represents the learning rate of step i, r is the reward for taking action a in state s, and γ ∈ (0, 1)]Is a discount factor. Through a certain number of iterations, a group of pareto approximation sets S can be converged and obtained finallypEach solution is a set of disaster backup multicast trees tiE.g. T. Each disaster backup multicast tree tiMulticast tree path corresponding to a backup transmission requirement, including all disaster backup multicast trees tiThe set T of (a) corresponds to the routing schedule of all backup transmission needs.
The process of obtaining the optimal backup routing scheme is as follows:
initializing a backup routing model, wherein the backup routing model comprises a pareto approximation set and a Q table;
(II) initializing an adjacency matrix of a route;
initializing the state of the intelligent agent as a current node, calculating an optional action set of the node in the initialized state, selecting a next node according to an epsilon greedy strategy, calculating a reward factor, updating global information and a Q table, and acquiring an available backup terminal point;
judging whether the backup number meets the backup requirement number, if so, executing the step (five); if not, executing the step (II);
adding the solution of the obtained backup routing model into a pareto approximate set;
and (VI) outputting the pareto approximate set as an optimal backup routing scheme.
Judging whether the solution in the pareto approximate set meets an iteration condition or not, and if so, outputting the solution; if not, executing the step (II).
Specifically, as shown in fig. 1, the following are:
the pareto approximation set is initialized to the empty set and the Q table is initialized to a value of 0. And setting the iteration number, starting an iteration loop, and storing the route path by using the adjacency matrix. Initializing an adjacent matrix to be a 0 matrix, initializing the state of an agent to be a current node, calculating an optional action set in the current state, selecting a next node in the optional action set according to an epsilon greedy strategy, updating global information such as the adjacent matrix, calculating a reward factor and updating a Q table. When a node arrives, whether the node is an available backup node is judged, if not, the state of initializing the agent is returned to, and the next node is continuously selected in one step; and if so, executing the next judgment. Judging whether the backup number meets the backup requirement number, if not, returning to the routing starting point to continue circulating until the backup number meets the backup requirement number; and if so, executing the next judgment. Judging whether the generated solution is dominated by the solution of the approximation set or not, and if so, discarding the solution; if not, the solution is added to the approximate set. Judging whether an iteration condition is met, if not, continuing iteration; if so, terminating the iteration and outputting the pareto approximation set.
The output pareto approximate set is the optimal backup routing scheme.
In the multi-target disaster backup method between data centers based on reinforcement learning disclosed in the embodiment, a plurality of disaster backup multicast trees are constructed by using factors such as sharing degree of the multicast trees, so that bandwidth waste is reduced; sending data in different time slots by using a store-and-forward mechanism so as to relieve the congestion of the maximum link; and continuously constructing a routing path by utilizing multi-objective reinforcement learning, thereby obtaining a backup routing scheme which is better in both backup cost and load balancing.
The method calculates the optimal backup routing scheme of the data to be backed up through a backup routing model, wherein the backup routing model takes the fitness of each link in a multicast tree in a time-expanded network to the multicast tree and the congestion factor of each link as constraints, and takes the minimized backup cost and the link load balance as targets to solve the optimal backup routing scheme, and the factors of the backup cost and the load balance are fully considered, so that the obtained optimal backup routing scheme relieves the maximum link congestion on the basis of reducing broadband waste.
Example 2
In this embodiment, a reinforcement learning-based inter-data center multi-objective disaster backup system is disclosed, which includes:
the acquisition module acquires data to be backed up;
the backup routing selection model comprises a fitness function of each link in a multicast tree in the time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load;
and the calculation module is used for inputting the data to be backed up into the backup routing selection model to obtain the optimal backup routing scheme.
Example 3
In this embodiment, a computer-readable storage medium is disclosed for storing computer instructions that, when executed by a processor, perform the steps of the reinforcement learning-based inter-data-center multi-objective disaster backup method described in embodiment 1.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Although the present disclosure has been described with reference to specific embodiments, it should be understood that the scope of the present disclosure is not limited thereto, and those skilled in the art will appreciate that various modifications and changes can be made without departing from the spirit and scope of the present disclosure.

Claims (9)

1. The multi-target disaster backup method among the data centers based on the reinforcement learning is characterized by comprising the following steps:
expanding the network among the data centers to obtain a time expansion network;
acquiring data to be backed up in a time expansion network;
inputting data to be backed up into a backup routing selection model to obtain an optimal backup routing scheme;
transmitting backup of data to be backed up in a time expansion network by an optimal backup routing scheme;
the backup routing selection model comprises a fitness function of each link in a multicast tree in a time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load;
the specific process for obtaining the optimal backup routing scheme is as follows:
initializing a backup routing model, wherein the backup routing model comprises a pareto approximation set and a Q table;
(II) initializing an adjacency matrix of a route;
initializing the state of the intelligent agent as a current node, calculating an optional action set of the node in the initialized state, selecting a next node according to an epsilon greedy strategy, calculating a reward factor, updating global information and a Q table, and acquiring an available backup terminal point;
judging whether the backup number meets the backup requirement number, if so, executing the step (five); if not, executing the step (II);
adding the solution of the obtained backup routing model into a pareto approximate set;
and (VI) outputting the pareto approximate set as an optimal backup routing scheme.
2. The reinforcement learning-based inter-data-center multi-objective disaster backup method as claimed in claim 1, wherein the time topology network is composed of the same inter-data-center network topology of a plurality of different time slots; the edge of the topology after the network topology is copied among the data centers is a transmission edge, and the edge connecting the same node in the adjacent time slots is a reserved edge.
3. The reinforcement learning-based inter-data-center multi-objective disaster backup method according to claim 1, wherein multicast routing and store-and-forward mechanisms are used to transmit data to be backed up in the time-expanding network.
4. The reinforcement learning-based multi-objective disaster backup method among data centers as claimed in claim 1, wherein the data centers are used as routing nodes, and can temporarily store data and send data when the link of the next hop is idle.
5. The reinforcement learning-based multi-objective disaster backup method among data centers as claimed in claim 1, wherein the data to be backed up is routed in a multicast manner, each backup requirement corresponds to a disaster backup multicast tree, the starting point of the disaster backup multicast tree is a source data center with backup requirement, and the end points are a plurality of backup data centers;
all backup requirements correspond to a disaster backup multicast forest comprised of all disaster backup multicast trees.
6. The reinforcement learning-based multi-objective disaster backup method among data centers as claimed in claim 1, wherein an epsilon greedy strategy is used to solve a backup route selection model, to obtain each route node in a multicast tree, and to form multicast tree paths by each route node, and all multicast tree paths form an optimal backup route scheme.
7. The reinforcement learning-based multi-objective disaster backup method among data centers as claimed in claim 1, wherein, judging whether the solution in pareto approximate concentration meets the iteration condition, if yes, outputting; if not, executing the step (II).
8. The system for multi-target disaster backup among data centers based on reinforcement learning is characterized by comprising the following steps:
the acquisition module acquires data to be backed up;
the backup routing selection model comprises a fitness function of each link in a multicast tree in the time expansion network to the multicast tree and a congestion factor function of each link, and an optimal backup routing scheme is obtained by solving with the aim of minimizing backup cost and balancing link load;
the computing module is used for inputting the data to be backed up into the backup routing selection model to obtain an optimal backup routing scheme;
the specific process for obtaining the optimal backup routing scheme is as follows:
initializing a backup routing model, wherein the backup routing model comprises a pareto approximation set and a Q table;
(II) initializing an adjacency matrix of a route;
initializing the state of the intelligent agent as a current node, calculating an optional action set of the node in the initialized state, selecting a next node according to an epsilon greedy strategy, calculating a reward factor, updating global information and a Q table, and acquiring an available backup terminal point;
judging whether the backup number meets the backup requirement number, if so, executing the step (five); if not, executing the step (II);
adding the solution of the obtained backup routing model into a pareto approximate set;
and (VI) outputting the pareto approximate set as an optimal backup routing scheme.
9. A computer-readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the reinforcement learning-based inter-data-center multi-objective disaster backup method according to any one of claims 1 to 7.
CN202010820767.4A 2020-08-14 2020-08-14 Multi-target disaster backup method and system among data centers based on reinforcement learning Expired - Fee Related CN111988220B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010820767.4A CN111988220B (en) 2020-08-14 2020-08-14 Multi-target disaster backup method and system among data centers based on reinforcement learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010820767.4A CN111988220B (en) 2020-08-14 2020-08-14 Multi-target disaster backup method and system among data centers based on reinforcement learning

Publications (2)

Publication Number Publication Date
CN111988220A CN111988220A (en) 2020-11-24
CN111988220B true CN111988220B (en) 2021-05-28

Family

ID=73433984

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010820767.4A Expired - Fee Related CN111988220B (en) 2020-08-14 2020-08-14 Multi-target disaster backup method and system among data centers based on reinforcement learning

Country Status (1)

Country Link
CN (1) CN111988220B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114244854B (en) * 2021-12-21 2022-11-22 山东大学 Data center data backup method and system based on reinforcement learning

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108900419A (en) * 2018-08-17 2018-11-27 北京邮电大学 Route decision method and device based on deeply study under SDN framework
CN109218180A (en) * 2018-08-24 2019-01-15 刘昱 Multicast control method and device based on Local Area Network
CN110807069A (en) * 2019-10-23 2020-02-18 华侨大学 Entity relationship joint extraction model construction method based on reinforcement learning algorithm
CN111147387A (en) * 2019-12-19 2020-05-12 北京邮电大学 Flow control method and device for hybrid SDN network
CN111367282A (en) * 2020-03-09 2020-07-03 山东大学 Robot navigation method and system based on multimode perception and reinforcement learning

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090252033A1 (en) * 2008-04-08 2009-10-08 At&T Knowledge Ventures, L.P. System and method of distributing media content
CN101478801B (en) * 2008-12-23 2011-03-16 华南理工大学 Wireless Ad Hoc network multicast routing method based on dynamic source tree
CN101511054B (en) * 2009-03-23 2010-12-29 山东大学 Novel IPv6 network mobile multicast protocol
US8462621B2 (en) * 2009-07-27 2013-06-11 At&T Intellectual Property I, L.P. Systems and methods of multicast reconfiguration using cross-layer information
CN102571527A (en) * 2010-12-09 2012-07-11 上海杉达学院 Backup and load balancing system based on virtual router redundancy
CN102868602B (en) * 2012-10-12 2014-11-19 南京邮电大学 Improved routing method based on social correlation degree in delay tolerant network
CN102970231B (en) * 2012-11-20 2018-05-01 中兴通讯股份有限公司 Multicast data flow forwards implementation method and route-bridge（RB）
CN103618675B (en) * 2013-11-11 2017-01-18 西安交通大学 Content-network-oriented content-influence-based caching method
CN105897605B (en) * 2016-04-08 2018-08-28 重庆邮电大学 A kind of power line carrier communication network network jamming control method based on IPv6
CN110798841B (en) * 2018-08-02 2022-06-24 富士通株式会社 Multi-hop wireless network deployment method, network capacity determination method and device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108900419A (en) * 2018-08-17 2018-11-27 北京邮电大学 Route decision method and device based on deeply study under SDN framework
CN109218180A (en) * 2018-08-24 2019-01-15 刘昱 Multicast control method and device based on Local Area Network
CN110807069A (en) * 2019-10-23 2020-02-18 华侨大学 Entity relationship joint extraction model construction method based on reinforcement learning algorithm
CN111147387A (en) * 2019-12-19 2020-05-12 北京邮电大学 Flow control method and device for hybrid SDN network
CN111367282A (en) * 2020-03-09 2020-07-03 山东大学 Robot navigation method and system based on multimode perception and reinforcement learning

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于多目标优化的组播路由问题研究;全文明;《中国优秀硕士学位论文全文数据库》;20181015;全文 *

Also Published As

Publication number Publication date
CN111988220A (en) 2020-11-24

Similar Documents

Publication Publication Date Title
CN110493131B (en) Design method of spatial information network routing strategy under SDN architecture
AU691260B2 (en) Method for efficient aggregation of link metrics
Buragohain et al. Power aware routing for sensor databases
CN101888331B (en) Method and apparatus for providing fast reroute of a unicast packet
CN103561445B (en) A kind of wireless sense network multi-path routing method based on network code
CN112600609B (en) Network capacity estimation method of satellite network system
Verma et al. Integrated routing protocol for opportunistic networks
CN112350949B (en) Rerouting congestion control method and system based on flow scheduling in software defined network
CN102510404A (en) Nondestructive continuous extensible interconnection structure for data center
CN109586785A (en) Low-track satellite network routing policy based on K shortest path first
CN111988220B (en) Multi-target disaster backup method and system among data centers based on reinforcement learning
CN103888976A (en) Link selection method combined with network scheduling and routing
CN112543151A (en) SDN controller deployment method and device, electronic equipment and storage medium
CN102711125A (en) Method for improving transmission capability of wireless mesh network
CN107040466B (en) Multi-domain cooperative data transmission path selection method based on layered architecture of Internet of things
CN115865761B (en) Network entity task activity cost evaluation method considering blocking feedback
CN101986727A (en) Multipath improvement method of network on-demand distance vector multicast routing protocol and corresponding device
WO2020134713A1 (en) Network node election method and node device
CN106028415A (en) Routing method of delay tolerant network for applications of Internet of Things
CN103338490A (en) Network data routing method and network node
Chenji et al. Pareto optimal cross layer lifetime optimization for disaster response networks
Manfredi et al. Small world effects in networks: An engineering interpretation
CN108199890B (en) A kind of software defined network resilient controller dispositions method
Lee et al. A new adaptive routing scheme based on the traffic characteristics in mobile ad hoc networks
Sun et al. Novel dynamic ant genetic algorithm for QoS routing in wireless mesh networks

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
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210528

CF01 Termination of patent right due to non-payment of annual fee