WO2015078498A1 - Procédé et système d'équilibrage de charge dans un réseau sdn - Google Patents

Procédé et système d'équilibrage de charge dans un réseau sdn Download PDF

Info

Publication number
WO2015078498A1
WO2015078498A1 PCT/EP2013/074888 EP2013074888W WO2015078498A1 WO 2015078498 A1 WO2015078498 A1 WO 2015078498A1 EP 2013074888 W EP2013074888 W EP 2013074888W WO 2015078498 A1 WO2015078498 A1 WO 2015078498A1
Authority
WO
WIPO (PCT)
Prior art keywords
sdn controller
cluster
primary
application
sdn
Prior art date
Application number
PCT/EP2013/074888
Other languages
English (en)
Inventor
Dror MIZRACHI
Original Assignee
Huawei Technologies Co., Ltd.
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 Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Priority to CN201380077828.0A priority Critical patent/CN105340241A/zh
Priority to PCT/EP2013/074888 priority patent/WO2015078498A1/fr
Publication of WO2015078498A1 publication Critical patent/WO2015078498A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1025—Dynamic adaptation of the criteria on which the server selection is based

Definitions

  • the present invention relates to a method and system for balancing load in a cluster of software defined network (SDN) controllers.
  • SDN software defined network
  • the method and system of the present invention balances a load of applications and/or devices intending to connect to the cluster of SDN controllers.
  • SDN is an emerging network technology addressing customization and optimization concerns.
  • the data plane is decoupled from the control plane, and thus modern communication networks may be simplified.
  • the control plane functions are normally implemented in one or more SDN controllers.
  • SDN controllers may be grouped into clusters.
  • a cluster of distributed SDN controllers has to simultaneously handle a large number of connected applications and devices in the SDN.
  • one or more devices such as a virtual switch (vSwitch), and one or more applications, such as deep packet inspection (DPI), firewall (FW), web application firewalls (WAF) and more, may be connected to a cluster of distributed SDN controllers.
  • DPI deep packet inspection
  • FW firewall
  • WAF web application firewalls
  • a cluster of SDN controllers is preferably designed such that the load on the individual SDN controllers in the cluster is balanced.
  • each SDN controller should serve the same number of devices and/or applications, irrespectively of the actual total load of the cluster.
  • LB load balancer
  • NBI Northbound Interface
  • SBI Southbound Interface
  • each of these LBs is a bottleneck, and actually each LB is thus a potential single point of failure (SPOF).
  • SPOF single point of failure
  • the term availability refers in the present invention to the ability of a user and/or application and/or device to obtain a service via a cluster of SDN controllers or to access the cluster. If the cluster is inaccessible, or if a service may not be obtained via the cluster, the cluster is not available to the user and/or application and/or device.
  • the cluster and specifically the services and applications running on the SDN controllers in the cluster, are preferably available at all times.
  • an object of the present invention is to balance a load between applications and/or devices and SDN controllers contained in a cluster of SDN controllers, without creating a SPOF, which endangers the whole cluster to become unavailable.
  • the above-mentioned object is achieved by the solution provided in the enclosed independent claims.
  • Advantageous implementations are defined in the respective dependent claims.
  • the core of the invention is to mitigate the SPOFs of traditional LBs by a distributed load balancing mechanism, which is implemented in a software based manner on each SDN controller of a cluster using a coordination service for synchronizing the loads between the individual SDN controllers.
  • a first aspect of the present invention provides a method for balancing load in a cluster of SDN controllers, the method comprising defining for an application and/or device, intending to connect to the cluster, a primary SDN controller, connecting, by the application and/or device, to the primary SDN controller, determining, by the primary SDN controller, a best SDN controller based on load data of the cluster, instructing, by the primary SDN controller, the application and/or device to connect to the best SDN controller, connecting, by the application and/or device, to the best SDN controller.
  • each application or device connecting to the cluster connects particularly to the primary SDN controller instead of connecting to a common LB as in the state of the art.
  • the LB is thus eliminated as SPOF.
  • the method of the present invention is able to automatically balance the load among the SDN controllers in the cluster.
  • Each new application and/or device is preferably instructed to connect to the best SDN controller in view of the load distribution in the cluster, i.e. preferably to the SDN controller currently handling the least amount of load, for example, the lowest number of applications and/or devices.
  • the method further comprises storing load data of the cluster using a predetermined path of a coordination cluster, wherein the step of determining the best SDN controller comprises locking, by the primary SDN controller, said predetermined path of the coordination cluster, reading, by the primary SDN controller, the load data of the cluster stored in the predetermined path of the coordination cluster, calculating, by the primary SDN controller, the best SDN controller, updating, by the primary SDN controller, the estimated load data of the cluster in the coordination cluster according to a load estimation based on the determined best SDN controller, unlocking, by the primary SDN controller, the predetermined path of the coordination cluster, and reconfiguring, by the SDN controller, the application and/or device by the calculated best SDN controller.
  • the coordination cluster is preferably based on Apache Zookeeper and is used for a synchronizing and locking mechanism. Thereby, it is guaranteed that each SDN controller always operates based on the newest and most relevant load data.
  • the coordination cluster provides the load data with high availability to the SDN controllers by using several nodes, wherein read/write can be carried out over any node.
  • the load balancing in the cluster of SDN controllers can be carried out more efficiently and more accurate.
  • the primary SDN controller instructs the application and/or device to connect to the best SDN controller.
  • the primary SDN controller reconfigures the device, preferably by using OF-Config, by the best SDN controller the primary SDN controller found.
  • the application and/or device connect to the best SDN controller.
  • the best SDN controller updates the real load data in the coordination cluster.
  • the method further comprises updating, by the best SDN controller, the load data of the cluster in the coordination cluster, when the application and/or device connects to the best SDN controller.
  • the coordination cluster is implemented by Apache Zookeeper.
  • the coordination cluster is provided with a plurality of nodes for the SDN controller of the cluster.
  • Apache Zookeeper is a popular and scalable coordination cluster, which promises coordination and consistency between multiple of its nodes.
  • the SDN controllers can obtain the load data from any node of Apache Zookeeper. Thus, high availability of the load data can be guaranteed.
  • the cluster of SDN controllers includes 2 to 20 SDN controllers.
  • the coordination controller preferably includes 3 to 5 nodes.
  • the method further comprises defining for the application and/or device, intending to connect to the cluster, a secondary SDN controller, and substituting the primary SDN controller for the secondary SDN controller in all relevant method steps, in case of a connection failure to the primary SDN controller.
  • the method is robust against a failure of the primary SDN controller.
  • the related secondary SDN controller immediately takes over all functions and steps. SPOFs of the cluster of SDN controllers are thereby eliminated.
  • the cluster provides high availability.
  • each application and/or device can also be able to connect to an initial SDN controller, which can then determine the best SDN controller.
  • the step of defining the primary SDN controller is carried out by one or more cloud management system, CMS.
  • the one or more CMS is, for example, and Open Stack, Cloud Stack or the like.
  • the CMS can use a round robin algorithm per cluster to define the primary and the secondary SDN controllers for each device and/or application. Thereby, initial load balancing is guaranteed during handshake. It is possible to use multiple CMS for redundancy, in order to achieve high availability.
  • the load data of the cluster comprises the number of applications and/or devices connected to each SDN controller of the cluster and/or comprises memory utilization, disk space, CPU utilization and/or I/O rates of each SDN controller.
  • the load data can alternatively or additionally be calculated based on a weighted logistic regression formula based on features like memory, CPU, I/O or the like.
  • the memory utilization of each SDN controllers, the processing power (CPU utilization) of each SDN controller, or the I/O rate of each SDN controller in the cluster can be taken into account.
  • the step of determining the best SDN controller comprises using a weighted round robin algorithm.
  • the method comprises defining a plurality of primary SDN controllers, a different primary SDN controller for each application and/or device, connecting, by each application and/or device, to its primary SDN controller, determining, by each of the plurality of primary SDN controllers, one primary SDN controller after the other according to an order, a best SDN controller for the application and/or device connected to it based on load data of the cluster, instructing, by each of the plurality of primary SDN controller, the application and/or device connected to it to connect to the best SDN controller determined for it, connecting, by each application and/or device, to its best SDN controller.
  • the plurality of primary SDN controllers, the different primary SDN controller for each application and/or device can be defined by using a round robin algorithm.
  • each SDN controller can read/write from any node of the coordination cluster. However, if locked it is ensured that the load data in the predetermined path of the coordination cluster is only by accessed by one SDN controller at a time.
  • Each application is associated with one primary SDN controller, wherein the SDN controllers of different applications and/or devices differ from each other.
  • the CMS calculates, preferably based on a round robin algorithm, the primary SDN controller of each application and/or device.
  • the CMS carries out proactive load balancing when defining the plurality of SDN controllers for the plurality of applications and/or devices.
  • determining the best SDN controllers comprises waiting, by each primary SDN controller, for its turn in the order, preferably waiting in a stand-by mode.
  • the method of the present invention can be carried out in an efficient and energy- saving manner.
  • the method further comprises defining a plurality of SDN controllers, a different SDN controller for each application and/or device, and substituting each primary SDN controller exhibiting a failure for its corresponding secondary SDN controller in all relevant method steps.
  • the substitution of the primary SDN controller to the secondary SDN controller ensures high availability of the cluster.
  • the present invention provides a system comprising a cluster of software defined network, SDN, controllers, the system being configured to define for an application and/or device intending to connect to the cluster a primary SDN controller, connect the application and/or device to the primary SDN controller, determine, by the primary SDN controller, a best SDN controller based on load data of the cluster, instruct, by the primary SDN controller, the application and/or device to connect to the best SDN controller, connect the application and/or device to the best SDN controller.
  • the system further comprises a coordination cluster for storing load data of the cluster using a predetermined path, wherein for determining the best SDN controller the system is configured to lock, by the primary SDN controller, said predetermined path of the coordination cluster, read, by the primary SDN controller, the load data of the cluster stored in the predetermined path of the coordination cluster, calculate, by the primary SDN controller, the best SDN controller, update, by the primary SDN controller, the load data of the cluster in the coordination cluster according to a load estimation based on the determined best SDN controller, unlock, by the primary SDN controller, the predetermined path of the coordination cluster, and reconfigure, by the SDN controller, the application and/or device by the calculated best SDN controller.
  • system is further configured to update, by the best SDN controller, the load data of the cluster in the coordination cluster, when the application and/or device is connected to the best SDN controller.
  • the system of the present invention achieves the same advantages as described above for the method.
  • the system of the present invention may be further configured to carry out the method steps according to all implementation forms described above.
  • Fig. 1 shows a basic flow of a method according to an embodiment of the present invention.
  • Figs. 2 to 9 show a specific example of the basic flow of the method shown in Fig. 1, which is carried out in a system according to an embodiment of the present invention.
  • Fig. 1 shows a sequence diagram, which demonstrates a basic flow of an embodiment of the method of the present invention.
  • the method implements a distributed load balancing mechanism for a cluster 1 of SDN controllers 2.
  • the method is particularly able to handle huge traffic and requests in both directions to and from the cluster 1 of SDN controllers 2.
  • the method is preferably carried out in an SDN environment, which preferably includes the cluster 1 of SDN controllers 2, at least one SDN application 3, like DPI, WAF etc., at least one devices 4, like a vSwitch, and a coordination cluster 5.
  • a primary SDN controller 2 is defined for and configured in the application 3 and/or device 4.
  • a cloud management system (CMS) 8 is responsible for defining and configuring the primary SDN controller 2 for each application 3 and/or device 4.
  • a secondary SDN controller 2 is defined for and configured in each application 3 and/or device 4 by the same CMS 8.
  • CMS cloud management system
  • Each secondary SDN controller 2 functions as a backup for its related primary SDN controller 2 (i.e. the primary SDN controller 2 determined for and configured in the same application 3 and/or device 4), in order to replace the related primary SDN controller 2, for example, in case said related primary SDN 2 controller fails or in case a connection to the related primary SDN controller 2 fails.
  • the CMS 8 uses a round robin algorithm to define for each application 3 and/or device 4 the primary SDN controller 2 and preferably the secondary SDN controller 2.
  • a round robin algorithm an initial load balancing is guaranteed during the handshake of an application 3 and/or device 4 and the cluster 1 can be provided. It is also possible to use more than one CMS 8 for reasons of redundancy, and thus in order to guarantee a high availability of the cluster 1.
  • the further method steps are described in relation to a primary SDN controller 2 defined per application 3 and/or device 4. However, the method automatically switches to one or more secondary SDN controllers 2 at any time or any step in the method, if the connection to one or more related primary SDN controllers 2 fails.
  • an application 3 and/or device 4 for which the CMS 8 has defined and configured the primary SDN controller 2 and optionally the secondary SDN controller 2, connects to its primary SDN controller 2.
  • a best SDN controller 2 is determined for the application 3 and/or device 4 taking into account the load data 6 of the cluster 1.
  • the coordination cluster 5 is preferably implemented by Apache Zookeeper, and comprises a plurality of nodes for communicating with the cluster 1. Apache Zookeeper promises good coordination capabilities and reliable consistency between its individual nodes.
  • the predetermined path can, for example, be identified by "clusterlD/SDNcontroller ID/applicationsldevices".
  • the load data 6 can, for example, include the number of applications 3 and/or devices
  • the load data 6 can, however, also be based on different characteristics of the SDN controllers 2, for example, their memory and CPU utilization, processing capabilities, and/or I/O rate.
  • a weighted formula may be used to calculate the load data 6 based on these different characteristics.
  • the primary SDN controller 2 can then determine a best SDN controller 2 for the application 3 and/or device 4 currently intending to connect to the cluster 1. For example, the primary SDN controller 2 defines and unutilized SDN controller 2, which the application 3 and/or device 4 may connect to, as the best SDN controller 2. Preferably, a basic weighted round robin algorithm is used to find the best SDN controller 2.
  • the primary SDN controller 2 updates in a further step of the method the estimated load data 6 into the predetermined path of the coordination cluster 5, i.e. the load data 6 assuming that the application and/or device will connect to the determined best SDN controller 2, and unlocks the predetermined path. Furthermore, the primary SDN controller 2 sends an instruction message to the application 3 and/or device 4 (preferably via OF-Config protocol), which intends to connect to the cluster 1, in order to instruct it to connect to the determined best SDN controller 2.
  • the application 3 and/or device 4 preferably via OF-Config protocol
  • the application 3 and/or device 4 updates in a further method step its configuration, i.e. it is reconfigured, in order to substitute the preconfigured primary SDN controller 2 by the best SDN controller 2. Then the application 3 and/or device 4 connect to its best SDN controller 2. When the application 3 and/or device 4 has connected to its best SDN controller 2, said best SDN controller 2 locks the load data 6 on the predetermined path of the coordination cluster 5 in a further step, and updates the real new load data 6 onto the predetermined path. Finally, the best SDN controller 2 unlocks the predetermined path. Within the coordination controller 5, the information stored in the predetermined path may be synched between a leader node and a follower node for reasons of data integrity and security.
  • Figures 2 to 9 show an example for how an embodiment of the method of the present invention is performed in an embodiment of a system of the present invention.
  • the system comprises a cluster 1 of SDN controllers 2, and preferably comprises a coordination cluster 5 for storing load data 6 of the cluster 1 using a predetermined path.
  • the cluster 1, i.e. particularly the SDN controllers 2 of the cluster 1, can connect and communicate with the coordination cluster 5 via a plurality of nodes 7 of the coordination cluster 5.
  • a plurality of applications 3 and/or devices 4 may be connected or may intend to connect.
  • three SDN controllers 2 (in the following distinguished by 2a, 2b and 2c) are included in the cluster 1.
  • the coordination cluster 5 is realized by Apache Zookeeper, which comprises three Zookeeper nodes 7.
  • the load data 6, in this case the number of applications 3 and devices 4 in the cluster 1 for each SDN controller 2 is stored.
  • two applications 3 are connected to each SDN controller 2a, 2b and 2c.
  • the SDN controller 2a and the SDN controller 2c each handle nine devices 4, while the SDN controller 2b handles eight devices 4.
  • the new devices 4a and 4b are both already configured, preferably by the CMS 8 described in relation to Fig. 1, with a primary SDN controller 2 and optionally with a secondary SDN controller 2 for each device 4a and 4b.
  • the device 4a shown in Fig. 3 is configured with SDN controller 2a as its primary SDN controller
  • the right device 4b is configured with SDN controller 2c as its primary SDN controller.
  • both devices 4a and 4b may be configured with a secondary SDN controller 2 for redundancy in case of a failure of a primary SDN controller 2.
  • the device 4a is configured with SDN controller 2b as its secondary SDN controller
  • the device 4b is configured with SDN controller 2c as its secondary SDN controller.
  • the two devices 4a and 4b intending to connect to the cluster 1 connect simultaneously to their pre-defined primary SDN controllers 2a and 2c, respectively, and send a request for being instructed with a best SDN controller 2.
  • the lock mechanism of the coordination cluster 5 is employed.
  • the SDN controller 2c locks the predetermined path on the coordination cluster 5, and obtains the load data 6 stored therein. Then, the SDN controller 2c calculates the best SDN controller 2 to use for its device 4b. Thereby, it preferably uses a weighted round robin algorithm. While the SDN controller 2c determines the best SDN controller 2, the SDN controller 2a cannot read data from the coordination cluster 5 and is preferably in a waiting state, more preferably in a stand-by state.
  • the SDN controller 2c determines that the best SDN controller 2 for the device 4b is in the SDN controller 2b, because it has currently only eight devices 4 connected to it, while the other SDN controllers 2a and 2c handle each nine connected devices 4. Accordingly, the SDN controller 2c updates the estimated load data 6 in the predetermined path of the coordination cluster 5, i.e. a load estimation based on the assumption that the device 4b will connect to the SDN controller 2b, and sends a response to the device 4b, in order to instruct it to use the SDN controller 2b as best SDN controller to connect to. Then, the SDN controller 2c unlocks the data load 6 on the predetermined path of the coordination cluster 5. In Fig.
  • the SDN controller 2a can read the load data 6 from the coordination cluster 5, and therefore locks the predetermined path. Similar as the SDN controller 2c before, it defines the best SDN controller 2 for the device 4a, preferably again by using a weighted round robin algorithm. In Fig. 6 the best SDN controller 2 for the device 4a is the SDN controller 2a, which is also the primary SDN controller of the device 4a.
  • the SDN controller 2a updates the estimated load data 6 on the predetermined path of the coordination cluster 5, i.e. the load estimated assuming that the device 4a will stay connected with the SDN controller 2a, and instructs the device 4a to use the SDN controller 2a as its best SDN controller.
  • Fig. 8 the device 4a reconnects to SDN controller 2a (actually it is already connected so it preferably just changes the state flag from "initial” to "real”). Then each of the best SDN controllers 2a and 2b update the real load data on the predetermined path of the coordination cluster 5.
  • Fig. 9 shows a final state of the system, after the two devices 4a and 4b have connected to the cluster 1.
  • the cluster 1 is load balanced.
  • the SDN controller 2a now handles ten devices 4, while the SDN controller 2b and the SDN controller 2c handle each nine devices 4.
  • the figs. 2 to 9 demonstrate a method and a system of an embodiment of the present invention for a distributed load balancing mechanism.
  • the load in a cluster 1 of SDN controllers 2 is balanced, without creating a SPOF or availability bottleneck.
  • a SDN with high availability can be supported. It is noted that the flow described within the figs.
  • the method and the system of the present invention achieve load balancing mechanism for a cluster 1 of SDN controllers 2.
  • the load balancing is implemented such that no SPOF is created. Since no SPOF is created, the SDN performance is increased, and high availability can be guaranteed at all times.
  • the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality.
  • a single processor or other unit may fulfill the functions of several items recited in the claims.
  • the mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage.
  • a computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Hardware Redundancy (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

La présente invention concerne un procédé d'équilibrage de charge dans un groupe (1) de contrôleurs SDN (2). Initialement, pour une application (3) et/ou un dispositif (4) intentant de se connecter au groupe (1), un contrôleur SDN primaire (2) est défini par un CMS. L'application (3) et/ou le dispositif (4) se connectent alors au contrôleur SDN primaire (2). Le contrôleur SDN primaire (2) détermine ensuite le meilleur contrôleur SDN (2) pour l'application (3) et/ou le dispositif (4) d'après des données de charge (6) du groupe (1) et/ou d'autres caractéristiques telles que l'utilisation (CPU, mémoire, E/S, etc.). Le contrôleur SDN primaire (2) commande à l'application (3) et/ou au dispositif (4) de se connecter au meilleur contrôleur SDN (2), ce qui amène l'application (3) et/ou le dispositif (4) à se connecter finalement au meilleur contrôleur SDN (2). Pour exécuter le procédé, le groupe (1) utilise de préférence un groupe de coordination (5) pour stocker et verrouiller les données de charge (6) dans un chemin prédéterminé du groupe de coordination (5). La présente invention concerne également un système comprenant un groupe (1) de contrôleurs SDN (2) configuré pour exécuter ledit procédé.
PCT/EP2013/074888 2013-11-27 2013-11-27 Procédé et système d'équilibrage de charge dans un réseau sdn WO2015078498A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201380077828.0A CN105340241A (zh) 2013-11-27 2013-11-27 用于均衡在sdn网络中的负载的方法和系统
PCT/EP2013/074888 WO2015078498A1 (fr) 2013-11-27 2013-11-27 Procédé et système d'équilibrage de charge dans un réseau sdn

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/EP2013/074888 WO2015078498A1 (fr) 2013-11-27 2013-11-27 Procédé et système d'équilibrage de charge dans un réseau sdn

Publications (1)

Publication Number Publication Date
WO2015078498A1 true WO2015078498A1 (fr) 2015-06-04

Family

ID=49683713

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2013/074888 WO2015078498A1 (fr) 2013-11-27 2013-11-27 Procédé et système d'équilibrage de charge dans un réseau sdn

Country Status (2)

Country Link
CN (1) CN105340241A (fr)
WO (1) WO2015078498A1 (fr)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105159767A (zh) * 2015-09-06 2015-12-16 北京京东尚科信息技术有限公司 用于基于zookeeper实现分布式调度的方法和装置
CN106559254A (zh) * 2015-12-29 2017-04-05 国网智能电网研究院 基于双端口交换机的sdn多域网络装置及实现方法
CN106713378A (zh) * 2015-07-30 2017-05-24 北京京东尚科信息技术有限公司 实现多个应用服务器提供服务的方法和系统
CN106936857A (zh) * 2015-12-29 2017-07-07 中国电信股份有限公司 一种混合云的连接管理方法、sdn控制器及混合云系统
CN107094119A (zh) * 2017-07-07 2017-08-25 广州市品高软件股份有限公司 一种基于云计算和sdn网络的负载均衡控制方法及系统
CN107579857A (zh) * 2017-09-29 2018-01-12 烽火通信科技股份有限公司 一种基于云的sdn控制器的冗余热备份保护的方法
CN108228581A (zh) * 2016-12-09 2018-06-29 阿里巴巴集团控股有限公司 Zookeeper兼容通信方法、服务器及系统
CN109451065A (zh) * 2018-12-26 2019-03-08 中电福富信息科技有限公司 一种软负载均衡分流自动化系统及其运行方法
CN110865993A (zh) * 2019-11-04 2020-03-06 紫光云技术有限公司 一种sdn控制器集群系统
CN111522665A (zh) * 2020-04-24 2020-08-11 北京思特奇信息技术股份有限公司 一种基于zookeeper实现Influxdb-proxy的高可用及负载均衡方法
CN112751789A (zh) * 2021-01-05 2021-05-04 浪潮云信息技术股份公司 一种非对称sdn控制器集群的实现方法及系统

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109716732B (zh) * 2016-08-03 2021-07-27 施耐德电器工业公司 用于软件定义的自动化系统中的部署的工业软件定义的网络架构
CN106330558A (zh) * 2016-08-31 2017-01-11 哈尔滨工业大学（威海） 应用于软件定义网络的控制器负载预测系统及方法
CN106603288A (zh) * 2016-12-15 2017-04-26 中国科学院沈阳自动化研究所 一种面向工业控制网络的集中式多控制器管理方法、装置和系统
CN107682410A (zh) * 2017-09-14 2018-02-09 广州西麦科技股份有限公司 一种分布式sdn控制器集群的控制方法及装置
CN112887110A (zh) * 2019-11-29 2021-06-01 中盈优创资讯科技有限公司 Sdn控制器下设备控制方法及装置
CN112350937B (zh) * 2020-08-20 2021-11-19 山西大学 一种高效的融合负载均衡和路由节能的路由计算方法
CN112667409A (zh) * 2020-11-25 2021-04-16 紫光云技术有限公司 一种可重入的分布式排它锁实现方法

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120044532A1 (en) * 2010-08-17 2012-02-23 Fujitsu Limited Management device, file server system, execution method and management program

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101729592B (zh) * 2008-10-29 2013-08-07 中国移动通信集团公司 一种分布式通信网络、设备及通信网络分离方法
CN103248724A (zh) * 2013-04-19 2013-08-14 中国（南京）未来网络产业创新中心 一种基于sdn控制器的dhcp广播处理方法

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120044532A1 (en) * 2010-08-17 2012-02-23 Fujitsu Limited Management device, file server system, execution method and management program

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ADVAIT DIXIT ET AL: "Towards an elastic distributed SDN controller", HOT TOPICS IN SOFTWARE DEFINED NETWORKING, ACM, 2 PENN PLAZA, SUITE 701 NEW YORK NY 10121-0701 USA, 16 August 2013 (2013-08-16), pages 7 - 12, XP058030691, ISBN: 978-1-4503-2178-5, DOI: 10.1145/2491185.2491193 *
FLORIN BALUS DIMITRI STILIADIS NUAGE NETWORKS NABIL BITAR VERIZON KENICHI OGAKI KDDI: "Federated SDN-based Controllers for NVO3; draft-sb-nvo3-sdn-federation-01.txt", FEDERATED SDN-BASED CONTROLLERS FOR NVO3; DRAFT-SB-NVO3-SDN-FEDERATION-01.TXT, INTERNET ENGINEERING TASK FORCE, IETF; STANDARDWORKINGDRAFT, INTERNET SOCIETY (ISOC) 4, RUE DES FALAISES CH- 1205 GENEVA, SWITZERLAND, 23 October 2012 (2012-10-23), pages 1 - 16, XP015085123 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106713378A (zh) * 2015-07-30 2017-05-24 北京京东尚科信息技术有限公司 实现多个应用服务器提供服务的方法和系统
CN106713378B (zh) * 2015-07-30 2020-07-31 北京京东尚科信息技术有限公司 实现多个应用服务器提供服务的方法和系统
CN105159767B (zh) * 2015-09-06 2019-07-02 北京京东尚科信息技术有限公司 用于基于zookeeper实现分布式调度的方法和装置
CN105159767A (zh) * 2015-09-06 2015-12-16 北京京东尚科信息技术有限公司 用于基于zookeeper实现分布式调度的方法和装置
CN106559254A (zh) * 2015-12-29 2017-04-05 国网智能电网研究院 基于双端口交换机的sdn多域网络装置及实现方法
CN106936857B (zh) * 2015-12-29 2020-05-19 中国电信股份有限公司 一种混合云的连接管理方法、sdn控制器及混合云系统
CN106936857A (zh) * 2015-12-29 2017-07-07 中国电信股份有限公司 一种混合云的连接管理方法、sdn控制器及混合云系统
CN108228581A (zh) * 2016-12-09 2018-06-29 阿里巴巴集团控股有限公司 Zookeeper兼容通信方法、服务器及系统
CN108228581B (zh) * 2016-12-09 2022-06-28 阿里云计算有限公司 Zookeeper兼容通信方法、服务器及系统
CN107094119B (zh) * 2017-07-07 2019-10-25 广州市品高软件股份有限公司 一种基于云计算和sdn网络的负载均衡控制方法及系统
CN107094119A (zh) * 2017-07-07 2017-08-25 广州市品高软件股份有限公司 一种基于云计算和sdn网络的负载均衡控制方法及系统
CN107579857A (zh) * 2017-09-29 2018-01-12 烽火通信科技股份有限公司 一种基于云的sdn控制器的冗余热备份保护的方法
CN109451065A (zh) * 2018-12-26 2019-03-08 中电福富信息科技有限公司 一种软负载均衡分流自动化系统及其运行方法
CN110865993A (zh) * 2019-11-04 2020-03-06 紫光云技术有限公司 一种sdn控制器集群系统
CN110865993B (zh) * 2019-11-04 2022-09-09 紫光云技术有限公司 一种sdn控制器集群系统
CN111522665A (zh) * 2020-04-24 2020-08-11 北京思特奇信息技术股份有限公司 一种基于zookeeper实现Influxdb-proxy的高可用及负载均衡方法
CN112751789A (zh) * 2021-01-05 2021-05-04 浪潮云信息技术股份公司 一种非对称sdn控制器集群的实现方法及系统

Also Published As

Publication number Publication date
CN105340241A (zh) 2016-02-17

Similar Documents

Publication Publication Date Title
WO2015078498A1 (fr) Procédé et système d'équilibrage de charge dans un réseau sdn
Paliwal et al. Controllers in SDN: A review report
CN106464528B (zh) 用于通信网络中的无接触编配的方法、介质、及装置
Xie et al. Control plane of software defined networks: A survey
US10057109B2 (en) Defining interdependent virtualized network functions for service level orchestration
US9219718B2 (en) System and method for supporting sub-subnet in an infiniband (IB) network
EP2972855B1 (fr) Configuration automatique de services externes sur la base d'une activité de réseau
US9716746B2 (en) System and method using software defined continuity (SDC) and application defined continuity (ADC) for achieving business continuity and application continuity on massively scalable entities like entire datacenters, entire clouds etc. in a computing system environment
EP3353952B1 (fr) Gestion de groupes de serveurs
US9350682B1 (en) Compute instance migrations across availability zones of a provider network
US20120233315A1 (en) Systems and methods for sizing resources in a cloud-based environment
CN106027270B (zh) 在联网计算环境中的按需功率管理方法及装置
US11601365B2 (en) Wide area networking service using provider network backbone network
US20150143470A1 (en) Managing an interface between an application and a network
US20160006642A1 (en) Network-wide service controller
GB2407887A (en) Automatically modifying fail-over configuration of back-up devices
US11824773B2 (en) Dynamic routing for peered virtual routers
US9350811B1 (en) Load balancing networks and load balancing methods
US20120117246A1 (en) Method And System For The Efficient And Automated Management of Virtual Networks
CN113014611B (zh) 一种负载均衡方法及相关设备
US7966394B1 (en) Information model registry and brokering in virtualized environments
JP7003876B2 (ja) 通信システム及び通信方法
WO2021108652A1 (fr) Configuration et gestion de réseaux privés globaux évolutifs
CN112655185B (zh) 软件定义网络中的服务分配的设备、方法和存储介质
Romanov et al. Principles of Building Modular Control Plane in Software-Defined Network

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 201380077828.0

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13798998

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13798998

Country of ref document: EP

Kind code of ref document: A1