WO2014023255A1 - 数据中心虚拟化网络地址的分发方法、系统及目录服务器 - Google Patents

数据中心虚拟化网络地址的分发方法、系统及目录服务器 Download PDF

Info

Publication number
WO2014023255A1
WO2014023255A1 PCT/CN2013/081118 CN2013081118W WO2014023255A1 WO 2014023255 A1 WO2014023255 A1 WO 2014023255A1 CN 2013081118 W CN2013081118 W CN 2013081118W WO 2014023255 A1 WO2014023255 A1 WO 2014023255A1
Authority
WO
WIPO (PCT)
Prior art keywords
network
network address
address information
directory server
address
Prior art date
Application number
PCT/CN2013/081118
Other languages
English (en)
French (fr)
Inventor
胡方伟
金利忠
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2014023255A1 publication Critical patent/WO2014023255A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00Network arrangements, protocols or services for addressing or naming
    • H04L61/45Network directories; Name-to-address mapping

Definitions

  • the present invention relates to cloud computing technologies, and in particular, to a data center virtualization network address distribution method, system, and directory server. Background technique
  • cloud computing With the widespread application of cloud computing technology, the cloud computing data center network is very different from the traditional data center network.
  • the cloud computing data center network is larger than the traditional data center network, so the traditional two-tier technology is in the cloud.
  • cloud computing uses a large number of virtualization technologies.
  • For a cloud computing data center one server can virtualize multiple virtual machines, and each virtual machine is assigned a virtual MAC. Addresses, and due to disaster recovery backups, traffic balancing, etc. between data centers, virtual machines may migrate between a data center and even across data centers.
  • Figure 1 shows the existing data center network virtualization overlay network diagram.
  • the VM is a virtual machine
  • NVE is the network virtual boundary
  • the dotted line indicates the data flow
  • the solid line indicates the connection relationship between the devices
  • the original data is on the NVE.
  • the overlay network adopts the layer 2 Over Layer 3 mode.
  • the inner layer encapsulates the packet as an Ethernet packet, and the outer layer can be a network layer packet header, a header, or a TCP packet header, and an inner layer packet address.
  • the outer address is the IP address.
  • Data is transmitted in an overlay network depending on the IP address for routing.
  • the implementation of the overlay network solves the scalability problem of the traditional Layer 2 data center network, and the network interworking between the data centers and the rapid migration of VMs.
  • Traditional media access control for this overlay cloud computing data center network
  • MAC Media Access Control
  • the address self-learning method cannot meet the requirements. The reasons are as follows: (1) In the MAC address self-learning mode, the MAC address table update depends on the aging time of the MAC address, and the MAC update is slow. The loss of data packets, but in the cloud computing data center network, because the VM migration application is more, and the data center network has higher reliability requirements, this slow MAC address update cannot satisfy the cloud computing data center network. Claim;
  • the first packet forwarding is broadcast or multicast. Since the network size of the cloud computing data center is much larger than that of the traditional Layer 2 data center, this first packet forwarding The transmission mode will cause a large number of broadcast or multicast packets to be formed in the cloud computing data center, which wastes network bandwidth and reduces the bandwidth utilization of the network.
  • the industry proposes a directory server ( Directory Server) way to store the user's MAC address and IP address mapping, the directory server can work in push (pull) and pull (Pull) In the two modes, the pull mode is the address information required to reply to the virtual boundary of the network after receiving the request packet of the network virtual boundary, and the Push mode is that the directory server actively sends the address information to one or more network virtual boundaries.
  • the pull mode is the address information required to reply to the virtual boundary of the network after receiving the request packet of the network virtual boundary
  • the Push mode is that the directory server actively sends the address information to one or more network virtual boundaries.
  • the directory server may send address information to multiple network virtual boundaries. Unnecessary distribution of messages and messages can result in wasted network bandwidth. Summary of the invention
  • the main purpose of the present invention is to provide a data center virtualized network address distribution method, system, and directory server, which can reduce address information sent by the directory server to multiple network virtual boundaries, thereby saving network bandwidth.
  • a method for distributing a data center virtualized network address according to an embodiment of the present invention includes:
  • the directory server receives and saves the network address registration message that is sent by the network virtual boundary and carries the tenant identifier and the network address information.
  • the network address information includes: Address information, address information of the virtual boundary of the source network, and address information of the virtual boundary of the destination network;
  • the address information of the virtual machine includes: a media access control (MAC) address of the virtual machine, or a virtual MAC address of the virtual machine, or an IP address of the virtual machine, or a virtual IP address of the virtual machine.
  • MAC media access control
  • the directory server receives and stores a network address registration message that is sent by the network virtual boundary and carries the tenant identifier and the network address information: when the directory server is a centralized server, the directory server receives the network virtual boundary.
  • the sent network address registration message carrying the tenant identification and network address information saves the network address information and the tenant identification in the network address registration message in the form of a mapping entry.
  • the directory server receives and stores a network address registration message that is sent by the network virtual boundary and carries the tenant identifier and the network address information: when the directory server is a distributed server, receiving the network address registration message.
  • the directory server sends the message to the discovered directory information and the tenant ID according to the network address information and the tenant ID in the distributed algorithm address registration message.
  • the network address information when the network address information is distributed to the virtual boundary of the network, the network address information is only distributed to the network virtual boundary related to the tenant identifier, and the directory server is mapped to the directory server in the Push mode. After the network address information and the tenant ID in the network address registration message are saved in the form of the entry, the notification message is sent to the data center network.
  • the network address information when the network address information is distributed to the virtual boundary of the network, the network address information is only distributed to the network virtual boundary related to the tenant identifier, and the directory server receives the network virtuality when the directory server works in the Pull mode.
  • the method further includes: after the virtual machine is started, sending an attached message to the virtual boundary of the network, where the attached message carries the address information of the virtual machine and the tenant identifier;
  • the network virtual boundary After receiving the attached message, the network virtual boundary locally stores the address information of the virtual machine and the tenant identifier, and sends a network address registration message carrying the tenant identifier and the network address information to the directory server.
  • the method further includes: receiving a network virtual boundary of the network address information distributed by the directory server, using the address information of the virtual boundary of the destination network in the network address information as the outer destination address, and sending the data packet to the virtual machine After encapsulation, the encapsulated data packet is forwarded and routed by the router of the overlay network depending on the outer destination address.
  • a data center virtualization network address distribution system provided by the embodiment of the present invention includes: a network virtual boundary, a directory server;
  • a network virtual boundary configured to send a network address registration message carrying the tenant identifier and the network address information to the directory server, and receive network address information distributed by the directory server;
  • the directory server is configured to receive and save a network address registration message that carries the tenant identifier and the network address information sent by the network virtual boundary, and distributes the network address information to the network virtual boundary.
  • the network address information includes: Address information, address information of the virtual boundary of the source network, and address information of the virtual boundary of the destination network;
  • the address information of the virtual machine includes: a MAC address of the virtual machine, or a virtual MAC address of the virtual machine, or an IP address of the virtual machine, or a virtual IP address of the virtual machine.
  • the system further includes: a virtual machine, configured to send an attach message to the virtual border of the network, where the attached message carries the address information of the virtual machine and the tenant identifier; the virtual boundary of the network is configured to receive After the message is attached, the address information of the virtual machine and the tenant ID are locally saved, and a network address registration message carrying the tenant ID and the network address information is sent to the directory server.
  • a virtual machine configured to send an attach message to the virtual border of the network, where the attached message carries the address information of the virtual machine and the tenant identifier
  • the virtual boundary of the network is configured to receive After the message is attached, the address information of the virtual machine and the tenant ID are locally saved, and a network address registration message carrying the tenant ID and the network address information is sent to the directory server.
  • the system further includes a router that superimposes the network, and is configured to transmit data packets between the virtual borders of the network.
  • the directory server includes: an address registration module and an address distribution module;
  • the address registration module is configured to receive and save a network address registration message that carries the tenant identification and network address information sent by the network virtual boundary;
  • the address distribution module is configured to distribute network address information only to network virtual boundaries related to the tenant identification when distributing network address information to the virtual boundary of the network.
  • the address registration module is configured to receive a network address registration message that is sent by the network virtual boundary and carries the tenant identifier and the network address information, and save the network in the form of a mapping entry. Network address information and tenant ID in the address registration message.
  • the address registration module of the directory server that receives the network address registration message is configured to search for the network address information and the tenant identifier in the registration message according to the distributed algorithm.
  • the text is sent to the discovered directory server;
  • the address registration module of the discovered directory server is configured to be in the form of a mapping entry
  • the network address information and the tenant ID are saved.
  • the address distribution module is configured to work in the Push mode, and after the address registration module saves the network address information and the tenant identifier in the network address registration message in the form of a mapping entry, the notification message is sent to the data.
  • the network virtual boundary associated with the tenant identification in the central network distributes the network address information.
  • the address distribution module is configured to receive a network address request message carrying a tenant identifier sent by the network virtual border, and the corresponding to the tenant identifier carried in the network address request message. a mapping entry, the network address information in the mapping entry is distributed to the network virtual boundary by a reply message.
  • a directory server provided by the embodiment of the present invention, the directory server includes: an address registration module and an address distribution module;
  • the address registration module is configured to receive and save a network address registration message that carries the tenant identification and network address information sent by the network virtual boundary;
  • the address distribution module is configured to distribute network address information only to network virtual boundaries related to the tenant identification when distributing network address information to the virtual boundary of the network.
  • the embodiment of the invention provides a method, a system and a directory server for distributing a data center virtualized network address, and the directory server receives and saves a network address registration message carried by the network virtual boundary and carries the tenant identifier and the network address information, in the network
  • the virtual boundary distributes the network address information, only the network address information is distributed to the network virtual boundary related to the tenant identifier; thus, the address information sent by the directory server to the multiple virtual borders of the network can be reduced, thereby saving network bandwidth.
  • Figure 1 is a diagram of an existing data center network virtualization overlay network
  • FIG. 2 is a schematic flowchart of a method for distributing a data center virtualized network address according to an embodiment of the present invention
  • FIG. 3 is a structural diagram of a data center virtualization network address distribution system according to an embodiment of the present invention. Schematic diagram
  • FIG. 4 is a schematic flowchart of a first packet forwarding according to an embodiment of the present invention. detailed description
  • the directory server receives and stores the network address registration message that carries the tenant identifier and the network address information sent by the network virtual boundary, and the present invention is further described below with reference to the accompanying drawings and specific embodiments. Detailed description.
  • the embodiment of the invention implements a data center virtualized network address distribution method. As shown in FIG. 2, the method includes the following steps:
  • Step 101 The directory server receives and saves a network address registration message that is sent by the network virtual boundary and carries the tenant identifier and the network address information.
  • the network address information includes: address information of the virtual machine, address information of the virtual boundary of the source network, and address information of the virtual boundary of the destination network;
  • the address information of the virtual machine includes: a MAC address of the virtual machine, or a virtual MAC address of the virtual machine, or an IP address of the virtual machine, or a virtual IP address of the virtual machine;
  • the address information of the source network virtual boundary is an IP address of a network virtual boundary of the sending network address registration message
  • the address information of the virtual boundary of the destination network is generally an IP address of a virtual boundary of the destination network
  • the tenant identifier may be a tenant's name, or number, or a serial number
  • the directory server When the directory server is a centralized server, one data center has one or several directory servers that are redundantly backed up, and the directory server receives the network address registration carried by the network virtual boundary and carries the tenant identification and network address information. a message, in the form of a mapping entry, the network address information and the tenant identifier in the network address registration message are saved; further, when the directory server has saved the mapping entry corresponding to the tenant identifier, the directory server is configured according to The network address information and the tenant identifier in the network address registration message update the saved mapping entry;
  • a data server When the directory server is a distributed server, a data server has a plurality of directory servers receiving a network address registration message sent by the network virtual boundary and carrying the tenant identification and network address information, and is used according to a distributed algorithm.
  • a directory server that saves the mapping entry corresponding to the tenant identifier, and sends the network address information and the tenant identifier in the network address registration message to the discovered directory server by using the packet, where the found directory server maps the entry Forming the network address information and the tenant identification;
  • the distributed algorithm includes a distributed hash (DHT) algorithm or the like.
  • DHT distributed hash
  • the method further includes: after the virtual machine is started, sending an attach message to the virtual boundary of the network, where the attached message carries the address information of the virtual machine and the tenant identifier;
  • the network virtual boundary After receiving the attached message, the network virtual boundary locally stores the address information of the virtual machine and the tenant identifier, and sends a network address registration message carrying the tenant identifier and the network address information to the directory server.
  • the virtual machine when the virtual machine leaves the network, if the virtual machine migrates, the virtual machine sends a detach message to the network virtual boundary; after receiving the detach message, the network virtual boundary sends a deregistration to the directory server. a message, the directory server revokes the mapping entry related to the virtual machine according to the deregistration message; where the detach message carries the address information of the virtual machine and the tenant identifier; the message includes: Network address information and tenant ID.
  • the network virtual boundary may be an independent Ethernet switch device, or may be embedded as a function block in the terminal server. If the function block is embedded in the terminal server, the virtual machine and the network virtual There is no need for standardized protocols for attachment and detachment operations between boundaries;
  • the directory server may be a separate server device, or may be stored as a function block in a network device such as an Ethernet switch or a router of the data center network. Distributing network address information to network virtual boundaries associated with tenant identification;
  • the directory server saves the network address information and the tenant identifier in the network address registration message in the form of a mapping entry, and sends the notification message to the tenant in the data center network. Identifying the network virtual boundary to distribute the network address information; where the advertisement packet carries the tenant identifier, and only distributes to the network virtual boundary related to the tenant identifier, reducing the scope of the network address information advertisement;
  • the network virtual boundary related to the tenant identifier for example, when the tenant identifier is 100, all network virtual borders supporting the tenant identifier being 100 are network virtual boundaries related to the tenant identifier;
  • the directory server If the directory server works in the pull mode, the directory server receives the network address request (request) message that is sent by the network virtual border and carries the tenant identifier, and searches for the corresponding mapping entry according to the tenant identifier carried in the network address request message. The network address information in the mapping entry is distributed to the network virtual boundary by a reply message.
  • request the network address request
  • the network address information in the mapping entry is distributed to the network virtual boundary by a reply message.
  • the method further includes: receiving a network virtual boundary of the network address information distributed by the directory server, and using the address information of the virtual boundary of the destination network in the network address information as the outer destination address, and encapsulating and encapsulating the data packet sent by the virtual machine.
  • the subsequent data message depends on the outer destination address being forwarded and routed by the router of the overlay network.
  • the embodiment of the present invention further provides a data center virtualization network address distribution system.
  • the system includes: a network virtual boundary 11 and a directory server 12;
  • the network virtual boundary 11 is configured to send a network address registration message carrying the tenant identification and network address information to the directory server 12, and receive network address information distributed by the directory server 12;
  • the directory server 12 is configured to receive and save a network address registration message that carries the tenant identifier and the network address information sent by the network virtual border 11, and when the network address information is distributed to the network virtual border 11, only the network virtuality related to the tenant identifier is The boundary 11 distributes network address information.
  • the network address information includes: address information of a virtual machine, address information of a virtual boundary of a source network, and address information of a virtual boundary of a destination network;
  • the address information of the virtual machine includes: a MAC address of the virtual machine, or a virtual MAC address of the virtual machine, or an IP address of the virtual machine, or a virtual IP address of the virtual machine;
  • the address information of the source network virtual boundary is an IP address of a network virtual boundary of the sending network address registration message
  • the address information of the virtual boundary of the destination network is generally an IP address of a virtual boundary of the destination network.
  • the system further includes: after the virtual machine 13 is configured to be started, sending an attached message to the network virtual boundary 11, where the attached message carries the address information of the virtual machine and the tenant identifier;
  • the network virtual boundary 11 is configured to locally save the address information of the virtual machine and the tenant identifier after receiving the attach message, and send a network address registration message carrying the tenant identifier and the network address information to the directory server 12.
  • the system also includes a router 14 for overlaying the network configured to transmit data messages between the network virtual boundaries 11.
  • the directory server 12 includes: an address registration module 121, an address distribution module 122;
  • the address registration module 121 is configured to receive and save a network address registration message that is sent by the network virtual boundary 11 and carries the tenant identifier and the network address information.
  • the address distribution module 122 is configured to distribute the network address information only to the network virtual boundary 11 associated with the tenant identification when distributing the network address information to the network virtual boundary 11;
  • the address registration module 121 has The physical configuration is configured to receive a network address registration message that is sent by the network virtual boundary 11 and that carries the tenant identifier and the network address information, and save the network address information and the tenant identifier in the network address registration message in the form of a mapping entry.
  • the address registration module 121 is specifically configured to update the network address information and the tenant identifier in the network address registration message. Saved map entry;
  • the address registration module 121 of the directory server is configured to receive the network address registration message that carries the tenant identifier and the network address information sent by the network virtual boundary 11 and is distributed according to the distributed
  • the algorithm finds a directory server for saving the mapping entry corresponding to the tenant identifier, and sends the network address information and the tenant identifier in the network address registration message to the discovered directory server by using the packet; the found directory
  • the address registration module 121 of the server is specifically configured to save the network address information and the tenant identifier in the form of a mapping entry.
  • the distributed algorithm includes a DHT algorithm and the like.
  • the address distribution module 122 is configured to work in the Push mode, and after the address registration module saves the network address information and the tenant identifier in the network address registration message in the form of a mapping entry, the notification message is sent to the data center network. Distributing the network address information in a network virtual boundary 11 related to the tenant identifier;
  • the address distribution module 122 is configured to receive a network address request message carrying a tenant identifier sent by the virtual border of the network, and find a corresponding mapping according to the tenant identifier carried in the network address request message. An entry, the network address information in the mapping entry is distributed to the network virtual boundary 11 by a reply message.
  • the embodiment of the present invention further provides a directory server.
  • the directory server 12 includes: an address registration module 121 and an address distribution module 122.
  • the address registration module 121 is configured to receive and save. a network address registration message carried by the network virtual boundary 11 carrying the tenant identification and network address information;
  • the address distribution module 122 is configured to distribute the network address information only to the network virtual boundary 11 related to the tenant identification when distributing the network address information to the network virtual boundary 11;
  • the address registration module 121 is configured to receive a network address registration message that carries the tenant identifier and network address information sent by the network virtual border 11, and save the network in the form of a mapping entry. Network address information and tenant ID in the address registration message;
  • the address registration module 121 is specifically configured to update the network address information and the tenant identifier in the network address registration message. Saved map entry;
  • the address registration module 121 of the directory server is configured to receive the network address registration message that carries the tenant identifier and the network address information sent by the network virtual boundary 11 and is distributed according to the distributed
  • the algorithm finds a directory server for saving the mapping entry corresponding to the tenant identifier, and sends the network address information and the tenant identifier in the network address registration message to the discovered directory server by using the packet; the found directory
  • the address registration module 121 of the server is specifically configured to save the network address information and the tenant identifier in the form of a mapping entry.
  • the distributed algorithm includes a DHT algorithm and the like.
  • the address distribution module 122 is configured to work in the Push mode, and after the address registration module saves the network address information and the tenant identifier in the network address registration message in the form of a mapping entry, the notification message is sent to the data center network. Distributing the network address information in a network virtual boundary 11 related to the tenant identifier;
  • the address distribution module 122 is configured to receive a network address request message carrying a tenant identifier sent by the virtual border of the network, and find a corresponding mapping according to the tenant identifier carried in the network address request message. An entry, the network address information in the mapping entry is distributed to the network virtual boundary 11 by a reply message. Reason.
  • This embodiment is the first packet forwarding process of data forwarding.
  • the directory server works in the pull mode. As shown in FIG. 4, the following steps are included:
  • Step 201 The virtual machine VM 1 sends a data packet to the corresponding network virtual boundary NVE 1;
  • the layer destination address when not found in the local cache, confirms that the data packet is the first packet, and sends a network address request message to the directory server, where the network address request message carries the destination address information of the virtual machine VM1, and the tenant Identifying, etc.;
  • Step 203 When the directory server receives the network address request message, searching for a corresponding mapping entry according to the tenant identifier in the network address request message, and using the response message to set the network address information in the mapping entry Sending to the network virtual boundary NVE1, the network address information includes an IP address of the network virtual boundary NVE2;
  • Step 204 After receiving the response message, the network virtual boundary NVE1 saves the network address information and the tenant identifier, and sends the IP address of the network virtual boundary NVE2 in the network address information as the outer destination address to the virtual machine VM1.
  • the data packet is encapsulated, and the encapsulated data packet is forwarded and routed by the router of the overlay network according to the outer destination address, and sent to the network virtual boundary NVE2;
  • Step 205 The network virtual boundary NVE2 decapsulates the received data packet, and sends the decapsulated data packet to the virtual machine VM2.
  • the network virtual boundary NVE1 in the forwarding process of the non-first packet, can locally find the outer destination address corresponding to the destination address of the non-first packet, and does not need to search the directory server for the outer destination address.
  • the directory server receives and saves a network address registration message sent by the network virtual boundary carrying the tenant identification and network address information, and distributes the network address to the virtual boundary of the network. Less directory servers send address information to multiple network virtual boundaries, saving network bandwidth.

Landscapes

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

Abstract

本发明公开了一种数据中心虚拟化网络地址的分发方法,目录服务器接收并保存网络虚拟边界发送的携带有租户标识和网络地址信息的网络地址注册消息,在向网络虚拟边界分发网络地址信息时,只向与租户标识相关的网络虚拟边界分发网络地址信息;本发明同时还公开了一种数据中心虚拟化网络地址的分发系统及目录服务器,通过本发明实施例的方案,能够减少目录服务器向多个网络虚拟边界发送的地址信息,节省网络带宽。

Description

数据中心虚拟化网络地址的分发方法、 系统及目录服务器 技术领域
本发明涉及云计算技术, 尤其涉及一种数据中心虚拟化网络地址的分 发方法、 系统及目录服务器。 背景技术
随着云计算技术的广泛应用, 云计算数据中心网络跟传统的数据中心 网络有很大的不同, 云计算数据中心网络的规模比传统的数据中心网络要 大, 所以传统的二层技术在云计算数据中心网络中存在扩展性问题, 另夕卜, 云计算会大量使用虚拟化技术, 对于云计算数据中心, 一台服务器可以虚 拟化出多个虚拟机,每个虚拟机都分配有虚拟 MAC地址, 而且由于数据中 心之间的容灾备份, 流量平衡等要求, 虚拟机可能会在一个数据中心甚至 跨数据中心之间进行迁移。
图 1为现有的数据中心网络虚拟化叠加网络图, 图 1中 VM为虚拟机, NVE为网络虚拟边界, 虚线表示数据流, 实线表示各设备间的连接关系, 在 NVE上对原始数据进行封装, 并依赖于中间叠加网络的路由器( Router ) 进行数据传输, NVE上需要保存外层地址和内层地址的映射关系, 采用自 学习的方式进行 MAC 地址表的建立和更新。 叠加网络采用 layer 2 Over Layer3的方式, 内层的封装报文为以太网报文, 外层可以为网络层报文头, 头, 也可以封装类 TCP报文头等形式, 内层的报文地址为 MAC地址, 外 层的地址为 IP地址。数据在叠加网络中传输依赖于 IP地址进行路由。这种 叠加网络的实现解决了传统二层数据中心网络的扩展性问题, 以及跨数据 中心的网络互通、 VM的快速迁移等问题。 对于这种叠加层的云计算数据中心网络, 采用传统的介质访问控制
( MAC, Media Access Control )地址自学习的方式已经不能满足要求, 原 因如下: ( 1 ) MAC地址自学习的方式下, MAC地址表的更新依赖于 MAC 地址的老化时间, MAC更新很慢, 造成数据报文的丟失, 而云计算数据中 心网络中, 由于 VM迁移应用较多, 而数据中心网络对于可靠性的要求较 高, 所以这种慢速的 MAC地址更新不能满足云计算数据中心网络的要求;
( 2 ) MAC地址学习方式下, 对于首包转发采用的是广播或者组播的方式, 由于云计算数据中心的网络规模比传统的二层数据中心的规模要大得多, 这种首包转发的传输方式会造成云计算数据中心中形成大量的广播或者组 播报文, 浪费网络带宽, 减少了网络的带宽利用率。
为了解决以上的云计算数据中心网络的问题, 业界提出了一种目录服 务器( Directory Server )的方式保存用户的 MAC地址和 IP地址映射, 目录 服务器可以工作于推送(Push )和拉取(Pull ) 两种方式, Pull方式为收到 网络虚拟边界的请求报文后向该网络虚拟边界回复请求所需要的地址信 息, Push方式为目录服务器主动向一个或者多个网络虚拟边界发送地址信 息。
但是, 这种使用目录服务器存储用户 MAC地址和 IP地址映射的方式 不能进行更精细的地址区分, 比如, 在 Push方式下, 目录服务器可能向多 个网络虚拟边界发送地址信息, 这种分发地址方式会产生不必要的分发消 息和报文, 造成网络带宽的浪费。 发明内容
有鉴于此, 本发明的主要目的在于提供一种数据中心虚拟化网络地址 的分发方法、 系统及目录服务器, 能够减少目录服务器向多个网络虚拟边 界发送的地址信息, 节省网络带宽。
为达到上述目的, 本发明实施例的技术方案是这样实现的: 本发明实施例提供的一种数据中心虚拟化网络地址的分发方法, 该方 法包括:
目录服务器接收并保存网络虚拟边界发送的携带有租户标识和网络地 址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址信息时, 只 上述方案中, 所述网络地址信息包括: 虚拟机的地址信息、 源网络虚 拟边界的地址信息和目的网络虚拟边界的地址信息;
所述虚拟机的地址信息包括: 虚拟机的介质访问控制 (MAC )地址、 或虚拟机的虚拟 MAC地址、或虚拟机的 IP地址、或虚拟机的虚拟 IP地址。
上述方案中, 所述目录服务器接收并保存网络虚拟边界发送的携带有 租户标识和网络地址信息的网络地址注册消息为: 所述目录服务器为集中 式的服务器时, 所述目录服务器接收网络虚拟边界发送的携带有租户标识 和网络地址信息的网络地址注册消息, 以映射条目的形式保存网络地址注 册消息中的网络地址信息和租户标识。
上述方案中, 所述目录服务器接收并保存网络虚拟边界发送的携带有 租户标识和网络地址信息的网络地址注册消息为: 所述目录服务器为分布 式的服务器时, 接收所述网络地址注册消息的目录服务器根据分布式算法 地址注册消息中的网络地址信息和租户标识通过报文发送到查找到的目录 息和租户标识。
上述方案中, 所述在向网络虚拟边界分发网络地址信息时, 只向与租 户标识相关的网络虚拟边界分发网络地址信息, 为: 对于目录服务器工作 在 Push方式的情况下, 目录服务器在以映射条目的形式保存网络地址注册 消息中的网络地址信息和租户标识之后, 通过通告报文向数据中心网络中 上述方案中, 所述在向网络虚拟边界分发网络地址信息时, 只向与租 户标识相关的网络虚拟边界分发网络地址信息, 为: 对于目录服务器工作 在 Pull方式的情况下, 目录服务器接收网络虚拟边界发送的携带租户标识 的网络地址请求(request )消息,根据所述网络地址请求消息中携带的租户 标识查找到对应的映射条目, 将所述映射条目中的网络地址信息通过应答 ( reply ) 消息分发给所述网络虚拟边界。
上述方案中, 该方法还包括: 虚拟机启动后, 向网络虚拟边界发送附 着消息, 所述附着消息携带有虚拟机的地址信息和租户标识;
所述网络虚拟边界收到所述附着消息后, 在本地保存虚拟机的地址信 息和租户标识, 并向目录服务器发送携带有租户标识和网络地址信息的网 络地址注册消息。
上述方案中, 该方法还包括: 收到目录服务器分发的网络地址信息的 网络虚拟边界, 以网络地址信息中的目的网络虚拟边界的地址信息为外层 目的地址, 对虚拟机发送的数据报文进行封装, 封装后的数据报文依赖于 外层目的地址由叠加层网络的路由器转发和路由。
本发明实施例提供的一种数据中心虚拟化网络地址的分发系统, 该系 统包括: 网络虚拟边界、 目录服务器; 其中,
网络虚拟边界, 配置为向目录服务器发送携带有租户标识和网络地址 信息的网络地址注册消息, 并接收目录服务器分发的网络地址信息;
目录服务器, 配置为接收并保存网络虚拟边界发送的携带有租户标识 和网络地址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址信 上述方案中, 所述网络地址信息包括: 虚拟机的地址信息、 源网络虚 拟边界的地址信息和目的网络虚拟边界的地址信息; 所述虚拟机的地址信息包括: 虚拟机的 MAC 地址、 或虚拟机的虚拟 MAC地址、 或虚拟机的 IP地址、 或虚拟机的虚拟 IP地址。
上述方案中, 该系统还包括: 虚拟机, 配置为启动后, 向网络虚拟边 界发送附着消息, 所述附着消息携带有虚拟机的地址信息和租户标识; 所述网络虚拟边界, 配置为收到所述附着消息后, 在本地保存虚拟机 的地址信息和租户标识, 并向目录服务器发送携带有租户标识和网络地址 信息的网络地址注册消息。
上述方案中, 该系统还包括叠加网络的路由器, 配置为在网络虚拟边 界之间传输数据报文。
上述方案中, 所述目录服务器包括: 地址注册模块、 地址分发模块; 其中,
地址注册模块, 配置为接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息;
地址分发模块, 配置为在向网络虚拟边界分发网络地址信息时, 只向 与租户标识相关的网络虚拟边界分发网络地址信息。
上述方案中, 所述目录服务器为集中式的服务器时, 所述地址注册模 块, 配置为接收网络虚拟边界发送的携带有租户标识和网络地址信息的网 络地址注册消息, 以映射条目的形式保存网络地址注册消息中的网络地址 信息和租户标识。
上述方案中, 所述目录服务器为分布式的服务器时, 接收到所述网络 地址注册消息的目录服务器的地址注册模块, 配置为根据分布式算法查找 注册消息中的网络地址信息和租户标识通过报文发送到查找到的目录服务 器;
所述查找到的目录服务器的地址注册模块, 配置为以映射条目的形式 保存所述网络地址信息和租户标识。
上述方案中, 所述地址分发模块, 配置为工作在 Push方式的情况下, 在地址注册模块以映射条目的形式保存网络地址注册消息中的网络地址信 息和租户标识之后, 通过通告报文向数据中心网络中与所述租户标识相关 的网络虚拟边界分发所述网络地址信息。
上述方案中, 所述地址分发模块, 配置为工作在 Pull方式的情况下, 接收网络虚拟边界发送的携带租户标识的网络地址请求消息, 根据所述网 络地址请求消息中携带的租户标识查找到对应的映射条目, 将所述映射条 目中的网络地址信息通过应答消息分发给所述网络虚拟边界。
本发明实施例提供的一种目录服务器, 该目录服务器包括: 地址注册 模块、 地址分发模块; 其中,
地址注册模块, 配置为接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息;
地址分发模块, 配置为在向网络虚拟边界分发网络地址信息时, 只向 与租户标识相关的网络虚拟边界分发网络地址信息。
本发明实施例提供了一种数据中心虚拟化网络地址的分发方法、 系统 及目录服务器, 目录服务器接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址 信息时, 只向与租户标识相关的网络虚拟边界分发网络地址信息; 如此, 能够减少目录服务器向多个网络虚拟边界发送的地址信息, 节省网络带宽。 附图说明
图 1为现有的数据中心网络虚拟化叠加网络图;
图 2 为本发明实施例实现数据中心虚拟化网络地址的分发方法的流程 示意图;
图 3 为本发明实施例实现数据中心虚拟化网络地址的分发系统的结构 示意图;
图 4为本发明实施例提供的首包转发的流程示意图。 具体实施方式
本发明实施例中, 目录服务器接收并保存网络虚拟边界发送的携带有 租户标识和网络地址信息的网络地址注册消息, 在向网络虚拟边界分发网 下面通过附图及具体实施例对本发明^:进一步的详细说明。
本发明实施例实现一种数据中心虚拟化网络地址的分发方法, 如图 2 所示, 该方法包括以下几个步骤:
步骤 101 : 目录服务器接收并保存网络虚拟边界发送的携带有租户标识 和网络地址信息的网络地址注册消息;
具体的, 所述网络地址信息包括: 虚拟机的地址信息、 源网络虚拟边 界的地址信息和目的网络虚拟边界的地址信息;
所述虚拟机的地址信息包括: 虚拟机的 MAC 地址、 或虚拟机的虚拟 MAC地址、 或虚拟机的 IP地址、 或虚拟机的虚拟 IP地址等;
所述源网络虚拟边界的地址信息为发送网络地址注册消息的网络虚拟 边界的 IP地址;
所述目的网络虚拟边界的地址信息一般为目的网络虚拟边界的 IP 地 址;
所述租户标识可以是租户的名称、 或编号、 或序列号等;
所述目录服务器为集中式的服务器时, 一个数据中心有一个或者几个 互为冗余备份的目录服务器, 所述目录服务器接收网络虚拟边界发送的携 带有租户标识和网络地址信息的网络地址注册消息, 以映射条目的形式保 存网络地址注册消息中的网络地址信息和租户标识; 进一步的, 当所述目 录服务器已保存有所述租户标识对应的映射条目时, 所述目录服务器根据 网络地址注册消息中的网络地址信息和租户标识更新所述已保存的映射条 目;
所述目录服务器为分布式的服务器时, 一个数据中心有多个目录服务 务器接收网络虚拟边界发送的携带有租户标识和网络地址信息的网络地址 注册消息, 并根据分布式算法查找到用于保存所述租户标识对应的映射条 目的目录服务器, 将所述网络地址注册消息中的网络地址信息和租户标识 通过报文发送到查找到的目录服务器, 所述查找到的目录服务器以映射条 目的形式保存所述网络地址信息和租户标识; 所述分布式算法包括分布式 哈希 (DHT )算法等。
本步骤之前, 还包括: 虚拟机启动后, 向网络虚拟边界发送附着消息, 所述附着消息携带有虚拟机的地址信息和租户标识;
所述网络虚拟边界收到所述附着消息后, 在本地保存虚拟机的地址信 息和租户标识, 并向目录服务器发送携带有租户标识和网络地址信息的网 络地址注册消息。
进一步的, 当虚拟机离开网络时, 如虚拟机发生迁移时, 所述虚拟机 向网络虚拟边界发送去附着消息; 所述网络虚拟边界收到所述去附着消息 后, 向目录服务器发送撤销注册消息, 所述目录服务器根据所述撤销注册 消息撤销与所述虚拟机相关的映射条目; 这里, 所述去附着消息携带有虚 拟机的地址信息和租户标识; 所述 4敦销注册消息包括: 网络地址信息和租 户标识。
本步骤中, 所述网络虚拟边界可以为独立的以太网交换机设备, 也可 以作为一个功能块嵌入到终端服务器中, 如果为功能块嵌入到终端服务器 中, 则所述虚拟机和所述网络虚拟边界之间不需要标准化的协议进行附着 和去附着操作; 所述目录服务器可以为单独服务器设备, 也可以作为一个功能块驻存 于数据中心网络的以太网交换机或者路由器等网络设备中。 向与租户标识相关的网络虚拟边界分发网络地址信息;
具体的, 对于目录服务器工作在 Push方式的情况下, 目录服务器在以 映射条目的形式保存网络地址注册消息中的网络地址信息和租户标识之 后, 通过通告报文向数据中心网络中与所述租户标识相关的网络虚拟边界 分发所述网络地址信息; 这里, 所述通告报文中携带有租户标识, 并且只 向与所述租户标识相关的网络虚拟边界分发, 减少了网络地址信息通告的 范围; 所述与所述租户标识相关的网络虚拟边界, 如: 所述租户标识为 100 时, 所有支持租户标识为 100 的网络虚拟边界都是与所述租户标识相关的 网络虚拟边界;
对于目录服务器工作在 Pull方式的情况下, 目录服务器接收网络虚拟 边界发送的携带租户标识的网络地址请求(request )消息,根据所述网络地 址请求消息中携带的租户标识查找到对应的映射条目, 将所述映射条目中 的网络地址信息通过应答(reply ) 消息分发给所述网络虚拟边界。
上述方法还包括: 收到目录服务器分发的网络地址信息的网络虚拟边 界, 以网络地址信息中的目的网络虚拟边界的地址信息为外层目的地址, 对虚拟机发送的数据报文进行封装, 封装后的数据报文依赖于外层目的地 址由叠加层网络的路由器转发和路由。
为了实现上述方法, 本发明实施例还提供一种数据中心虚拟化网络地 址的分发系统, 如图 3所示, 该系统包括: 网络虚拟边界 11、 目录服务器 12; 其中,
网络虚拟边界 11,配置为向目录服务器 12发送携带有租户标识和网络 地址信息的网络地址注册消息,并接收目录服务器 12分发的网络地址信息; 目录服务器 12,配置为接收并保存网络虚拟边界 11发送的携带有租户 标识和网络地址信息的网络地址注册消息, 在向网络虚拟边界 11分发网络 地址信息时, 只向与租户标识相关的网络虚拟边界 11分发网络地址信息。
所述网络地址信息包括: 虚拟机的地址信息、 源网络虚拟边界的地址 信息和目的网络虚拟边界的地址信息;
所述虚拟机的地址信息包括: 虚拟机的 MAC 地址、 或虚拟机的虚拟 MAC地址、 或虚拟机的 IP地址、 或虚拟机的虚拟 IP地址等;
所述源网络虚拟边界的地址信息为发送网络地址注册消息的网络虚拟 边界的 IP地址;
所述目的网络虚拟边界的地址信息一般为目的网络虚拟边界的 IP 地 址。
该系统还包括: 虚拟机 13, 配置为启动后, 向网络虚拟边界 11发送附 着消息, 所述附着消息携带有虚拟机的地址信息和租户标识;
所述网络虚拟边界 11, 具体配置为收到所述附着消息后, 在本地保存 虚拟机的地址信息和租户标识, 并向目录服务器 12发送携带有租户标识和 网络地址信息的网络地址注册消息。
该系统还包括叠加网络的路由器 14,配置为在网络虚拟边界 11之间传 输数据报文。
所述目录服务器 12包括: 地址注册模块 121、 地址分发模块 122; 其 中,
地址注册模块 121, 配置为接收并保存网络虚拟边界 11发送的携带有 租户标识和网络地址信息的网络地址注册消息;
地址分发模块 122,配置为在向网络虚拟边界 11分发网络地址信息时, 只向与租户标识相关的网络虚拟边界 11分发网络地址信息;
所述目录服务器 12为集中式的服务器时, 所述地址注册模块 121, 具 体配置为接收网络虚拟边界 11发送的携带有租户标识和网络地址信息的网 络地址注册消息, 以映射条目的形式保存网络地址注册消息中的网络地址 信息和租户标识;
进一步的, 当所述目录服务器 12已保存有所述租户标识对应的映射条 目时, 所述地址注册模块 121, 具体配置为才艮据网络地址注册消息中的网络 地址信息和租户标识更新所述已保存的映射条目;
所述目录服务器 12为分布式的服务器时, 就近的目录服务器的地址注 册模块 121, 具体配置为接收网络虚拟边界 11发送的携带有租户标识和网 络地址信息的网络地址注册消息, 并根据分布式算法查找到用于保存所述 租户标识对应的映射条目的目录服务器, 将所述网络地址注册消息中的网 络地址信息和租户标识通过报文发送到查找到的目录服务器; 所述查找到 的目录服务器的地址注册模块 121,具体配置为以映射条目的形式保存所述 网络地址信息和租户标识; 所述分布式算法包括 DHT算法等。
所述地址分发模块 122, 具体配置为工作在 Push方式的情况下, 在地 址注册模块以映射条目的形式保存网络地址注册消息中的网络地址信息和 租户标识之后, 通过通告报文向数据中心网络中与所述租户标识相关的网 络虚拟边界 11分发所述网络地址信息;
所述地址分发模块 122,具体配置为工作在 Pull方式的情况下,接收网 络虚拟边界发送的携带租户标识的网络地址请求消息, 根据所述网络地址 请求消息中携带的租户标识查找到对应的映射条目, 将所述映射条目中的 网络地址信息通过应答消息分发给所述网络虚拟边界 11。
基于上述系统, 本发明实施例还提供一种目录服务器, 如图 3 所示, 所述目录服务器 12包括: 地址注册模块 121、 地址分发模块 122; 其中, 地址注册模块 121, 配置为接收并保存网络虚拟边界 11发送的携带有 租户标识和网络地址信息的网络地址注册消息; 地址分发模块 122,配置为在向网络虚拟边界 11分发网络地址信息时, 只向与租户标识相关的网络虚拟边界 11分发网络地址信息;
所述目录服务器 12为集中式的服务器时, 所述地址注册模块 121, 具 体配置为接收网络虚拟边界 11发送的携带有租户标识和网络地址信息的网 络地址注册消息, 以映射条目的形式保存网络地址注册消息中的网络地址 信息和租户标识;
进一步的, 当所述目录服务器 12已保存有所述租户标识对应的映射条 目时, 所述地址注册模块 121, 具体配置为才艮据网络地址注册消息中的网络 地址信息和租户标识更新所述已保存的映射条目;
所述目录服务器 12为分布式的服务器时, 就近的目录服务器的地址注 册模块 121, 具体配置为接收网络虚拟边界 11发送的携带有租户标识和网 络地址信息的网络地址注册消息, 并根据分布式算法查找到用于保存所述 租户标识对应的映射条目的目录服务器, 将所述网络地址注册消息中的网 络地址信息和租户标识通过报文发送到查找到的目录服务器; 所述查找到 的目录服务器的地址注册模块 121,具体配置为以映射条目的形式保存所述 网络地址信息和租户标识; 所述分布式算法包括 DHT算法等。
所述地址分发模块 122, 具体配置为工作在 Push方式的情况下, 在地 址注册模块以映射条目的形式保存网络地址注册消息中的网络地址信息和 租户标识之后, 通过通告报文向数据中心网络中与所述租户标识相关的网 络虚拟边界 11分发所述网络地址信息;
所述地址分发模块 122,具体配置为工作在 Pull方式的情况下,接收网 络虚拟边界发送的携带租户标识的网络地址请求消息, 根据所述网络地址 请求消息中携带的租户标识查找到对应的映射条目, 将所述映射条目中的 网络地址信息通过应答消息分发给所述网络虚拟边界 11。 理。
本实施例为数据转发的首包转发的流程, 本实施例中目录服务器工作 在 Pull方式下, 如图 4所示, 包括以下几个步骤:
步骤 201 : 虚拟机 VM 1发送数据报文到对应的网络虚拟边界 NVE 1; 步骤 202: 网络虚拟边界 NVE1收到虚拟机 VM1发送的数据报文后, 在本地查找所述数据报文对应的外层目的地址, 当在本地的緩存中没有查 找到时, 确认所述数据报文为首包, 向目录服务器发送网络地址请求消息, 所述网络地址请求消息中携带虚拟机 VM1的目的地址信息、 租户标识等; 步骤 203: 当目录服务器收到所述网络地址请求消息时, 根据所述网络 地址请求消息中的租户标识查找到对应的映射条目, 通过应答消息将所述 映射条目中的网络地址信息发送给网络虚拟边界 NVE1,所述网络地址信息 包括网络虚拟边界 NVE2的 IP地址;
步骤 204: 网络虚拟边界 NVE1收到所述应答消息后,保存所述网络地 址信息和租户标识, 以网络地址信息中的网络虚拟边界 NVE2的 IP地址为 外层目的地址, 对虚拟机 VM1发送的数据报文进行封装, 封装后的数据报 文依赖于外层目的地址由叠加层网络的路由器转发和路由, 发送到网络虚 拟边界 NVE2;
步骤 205: 网络虚拟边界 NVE2解封装接收到的数据报文,将解封装后 的数据报文发送到虚拟机 VM2。
本实施例中, 对于非首包的转发过程中, 所述网络虚拟边界 NVE1 在 本地可以查找到非首包报文的目的地址对应的外层目的地址, 不需要向目 录服务器查找外层目的地址。 工业实用性
综上所述, 目录服务器接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址 少目录服务器向多个网络虚拟边界发送地址信息, 节省网络带宽。
以上所述, 仅为本发明的较佳实施例而已, 并非用于限定本发明的保 护范围。

Claims

权利要求书
1、 一种数据中心虚拟化网络地址的分发方法, 该方法包括:
目录服务器接收并保存网络虚拟边界发送的携带有租户标识和网络地 址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址信息时, 只
2、 根据权利要求 1所述的分发方法, 其中, 所述网络地址信息包括: 虚拟机的地址信息、 源网络虚拟边界的地址信息和目的网络虚拟边界的地 址信息;
所述虚拟机的地址信息包括: 虚拟机的介质访问控制 (MAC )地址、 或虚拟机的虚拟 MAC地址、或虚拟机的 IP地址、或虚拟机的虚拟 IP地址。
3、 根据权利要求 2所述的分发方法, 其中, 所述目录服务器接收并保 存网络虚拟边界发送的携带有租户标识和网络地址信息的网络地址注册消 息为: 所述目录服务器为集中式的服务器时, 所述目录服务器接收网络虚 拟边界发送的携带有租户标识和网络地址信息的网络地址注册消息, 以映
4、 根据权利要求 2所述的分发方法, 其中, 所述目录服务器接收并保 存网络虚拟边界发送的携带有租户标识和网络地址信息的网络地址注册消 息为: 所述目录服务器为分布式的服务器时, 接收所述网络地址注册消息 的目录服务器根据分布式算法查找到用于保存所述租户标识对应的映射条 目的目录服务器, 将所述网络地址注册消息中的网络地址信息和租户标识 通过报文发送到查找到的目录服务器, 所述查找到的目录服务器以映射条 目的形式保存所述网络地址信息和租户标识。
5、 根据权利要求 2所述的分发方法, 其中, 所述在向网络虚拟边界分 息, 为: 对于目录服务器工作在推送(Push ) 方式的情况下, 目录服务器 在以映射条目的形式保存网络地址注册消息中的网络地址信息和租户标识 之后, 通过通告报文向数据中心网络中与所述租户标识相关的网络虚拟边 界分发所述网络地址信息。
6、 根据权利要求 2所述的分发方法, 其中, 所述在向网络虚拟边界分 息, 为: 对于目录服务器工作在拉取(Pull )方式的情况下, 目录服务器接 收网络虚拟边界发送的携带租户标识的网络地址请求(request )消息, 根据 所述网络地址请求消息中携带的租户标识查找到对应的映射条目, 将所述 映射条目中的网络地址信息通过应答(reply ) 消息分发给所述网络虚拟边 界。
7、 根据权利要求 2所述的分发方法, 其中, 该方法还包括: 虚拟机启 动后, 向网络虚拟边界发送附着消息, 所述附着消息携带有虚拟机的地址 信息和租户标识;
所述网络虚拟边界收到所述附着消息后, 在本地保存虚拟机的地址信 息和租户标识, 并向目录服务器发送携带有租户标识和网络地址信息的网 络地址注册消息。
8、根据权利要求 2至 7任一项所述的分发方法,其中,该方法还包括: 收到目录服务器分发的网络地址信息的网络虚拟边界, 以网络地址信息中 的目的网络虚拟边界的地址信息为外层目的地址, 对虚拟机发送的数据报 文进行封装, 封装后的数据报文依赖于外层目的地址由叠加层网络的路由 器转发和路由。
9、 一种数据中心虚拟化网络地址的分发系统, 该系统包括: 网络虚拟 边界、 目录服务器; 其中,
网络虚拟边界, 配置为向目录服务器发送携带有租户标识和网络地址 信息的网络地址注册消息, 并接收目录服务器分发的网络地址信息; 目录服务器, 配置为接收并保存网络虚拟边界发送的携带有租户标识 和网络地址信息的网络地址注册消息, 在向网络虚拟边界分发网络地址信
10、 根据权利要求 9所述的分发系统, 其中, 所述网络地址信息包括: 虚拟机的地址信息、 源网络虚拟边界的地址信息和目的网络虚拟边界的地 址信息;
所述虚拟机的地址信息包括: 虚拟机的 MAC 地址、 或虚拟机的虚拟 MAC地址、 或虚拟机的 IP地址、 或虚拟机的虚拟 IP地址。
11、根据权利要求 10所述的分发系统, 其中, 该系统还包括: 虚拟机, 配置为启动后, 向网络虚拟边界发送附着消息, 所述附着消息携带有虚拟 机的地址信息和租户标识;
所述网络虚拟边界, 配置为收到所述附着消息后, 在本地保存虚拟机 的地址信息和租户标识, 并向目录服务器发送携带有租户标识和网络地址 信息的网络地址注册消息。
12、 根据权利要求 11所述的分发系统, 其中, 该系统还包括叠加网络 的路由器, 配置为在网络虚拟边界之间传输数据报文。
13、 根据权利要求 10所述的分发系统, 其中, 所述目录服务器包括: 地址注册模块、 地址分发模块; 其中,
地址注册模块, 配置为接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息;
地址分发模块, 配置为在向网络虚拟边界分发网络地址信息时, 只向 与租户标识相关的网络虚拟边界分发网络地址信息。
14、 根据权利要求 13所述的分发系统, 其中, 所述目录服务器为集中 式的服务器时, 所述地址注册模块, 配置为接收网络虚拟边界发送的携带 有租户标识和网络地址信息的网络地址注册消息, 以映射条目的形式保存 网络地址注册消息中的网络地址信息和租户标识。
15、 根据权利要求 13所述的分发系统, 其中, 所述目录服务器为分布 式的服务器时, 接收到所述网络地址注册消息的目录服务器的地址注册模 块, 配置为根据分布式算法查找到用于保存所述租户标识对应的映射条目 的目录服务器, 将所述网络地址注册消息中的网络地址信息和租户标识通 过报文发送到查找到的目录服务器;
所述查找到的目录服务器的地址注册模块, 配置为以映射条目的形式 保存所述网络地址信息和租户标识。
16、 根据权利要求 13所述的分发系统, 其中, 所述地址分发模块, 配 置为工作在 Push方式的情况下, 在地址注册模块以映射条目的形式保存网 络地址注册消息中的网络地址信息和租户标识之后, 通过通告报文向数据
17、 根据权利要求 13所述的分发系统, 其中, 所述地址分发模块, 配 置为工作在 Pull方式的情况下, 接收网络虚拟边界发送的携带租户标识的 网络地址请求消息, 根据所述网络地址请求消息中携带的租户标识查找到 对应的映射条目, 将所述映射条目中的网络地址信息通过应答消息分发给 所述网络虚拟边界。
18、 一种目录服务器, 其中, 该目录服务器包括: 地址注册模块、 地 址分发模块; 其中,
地址注册模块, 配置为接收并保存网络虚拟边界发送的携带有租户标 识和网络地址信息的网络地址注册消息;
地址分发模块, 配置为在向网络虚拟边界分发网络地址信息时, 只向 与租户标识相关的网络虚拟边界分发网络地址信息。
19、 根据权利要求 18所述的目录服务器, 其中, 所述目录服务器为集 中式的服务器时, 所述地址注册模块, 配置为接收网络虚拟边界发送的携 带有租户标识和网络地址信息的网络地址注册消息, 以映射条目的形式保 存网络地址注册消息中的网络地址信息和租户标识。
20、 根据权利要求 18所述的目录服务器, 其中, 所述目录服务器为分 布式的服务器时, 接收到所述网络地址注册消息的目录服务器的地址注册 模块, 配置为根据分布式算法查找到用于保存所述租户标识对应的映射条 目的目录服务器, 将所述网络地址注册消息中的网络地址信息和租户标识 通过报文发送到查找到的目录服务器;
所述查找到的目录服务器的地址注册模块, 配置为以映射条目的形式 保存所述网络地址信息和租户标识。
21、 根据权利要求 18所述的目录服务器, 其中, 所述地址分发模块, 配置为工作在 Push方式的情况下, 在地址注册模块以映射条目的形式保存 网络地址注册消息中的网络地址信息和租户标识之后, 通过通告报文向数 据中心网络中与所述租户标识相关的网络虚拟边界分发所述网络地址信 息。
22、 根据权利要求 18所述的目录服务器, 其中, 所述地址分发模块, 配置为工作在 Pull方式的情况下, 接收网络虚拟边界发送的携带租户标识 的网络地址请求消息, 根据所述网络地址请求消息中携带的租户标识查找 到对应的映射条目, 将所述映射条目中的网络地址信息通过应答消息分发 给所述网络虚拟边界。
PCT/CN2013/081118 2012-08-09 2013-08-08 数据中心虚拟化网络地址的分发方法、系统及目录服务器 WO2014023255A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210282370.XA CN103581277A (zh) 2012-08-09 2012-08-09 数据中心虚拟化网络地址的分发方法、系统及目录服务器
CN201210282370.X 2012-08-09

Publications (1)

Publication Number Publication Date
WO2014023255A1 true WO2014023255A1 (zh) 2014-02-13

Family

ID=50052192

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/081118 WO2014023255A1 (zh) 2012-08-09 2013-08-08 数据中心虚拟化网络地址的分发方法、系统及目录服务器

Country Status (2)

Country Link
CN (1) CN103581277A (zh)
WO (1) WO2014023255A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111654452A (zh) * 2020-05-08 2020-09-11 杭州迪普科技股份有限公司 一种报文处理的方法及装置

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104917682B (zh) * 2014-03-14 2017-12-15 新华三技术有限公司 一种叠加网络的配置方法和装置
US20160359720A1 (en) * 2015-06-02 2016-12-08 Futurewei Technologies, Inc. Distribution of Internal Routes For Virtual Networking
CN106559338A (zh) * 2015-09-29 2017-04-05 中国电信股份有限公司 Sdn网络中的租户划分方法、装置及sdn网络系统
CN106572021B (zh) * 2015-10-09 2021-07-06 中兴通讯股份有限公司 一种实现网络虚拟化叠加的方法与网络虚拟化边缘节点
CN114679370B (zh) * 2021-05-20 2024-01-12 腾讯云计算(北京)有限责任公司 一种服务器托管方法、装置、系统及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011142972A2 (en) * 2010-05-14 2011-11-17 Microsoft Corporation Interconnecting members of a virtual network
CN102480404A (zh) * 2010-11-19 2012-05-30 财团法人工业技术研究院 数据中心网络系统及其封包传送方法
CN102577255A (zh) * 2009-09-30 2012-07-11 阿尔卡特朗讯公司 云计算中企业的第2层无缝站点扩展

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2781060C (en) * 2010-05-28 2016-03-08 Huawei Technologies Co., Ltd. Virtual layer 2 and mechanism to make it scalable

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102577255A (zh) * 2009-09-30 2012-07-11 阿尔卡特朗讯公司 云计算中企业的第2层无缝站点扩展
WO2011142972A2 (en) * 2010-05-14 2011-11-17 Microsoft Corporation Interconnecting members of a virtual network
CN102480404A (zh) * 2010-11-19 2012-05-30 财团法人工业技术研究院 数据中心网络系统及其封包传送方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111654452A (zh) * 2020-05-08 2020-09-11 杭州迪普科技股份有限公司 一种报文处理的方法及装置
CN111654452B (zh) * 2020-05-08 2024-02-27 杭州迪普科技股份有限公司 一种报文处理的方法及装置

Also Published As

Publication number Publication date
CN103581277A (zh) 2014-02-12

Similar Documents

Publication Publication Date Title
US11546288B2 (en) Techniques for managing software defined networking controller in-band communications in a data center network
US8996675B2 (en) Interconnecting data centers for migration of virtual machines
US9515930B2 (en) Intelligent handling of virtual machine mobility in large data center environments
US10171357B2 (en) Techniques for managing software defined networking controller in-band communications in a data center network
US9864619B2 (en) Systems and methods for a data center architecture facilitating layer 2 over layer 3 communication
EP2853066B1 (en) Layer-3 overlay gateways
EP2905930B1 (en) Processing method, apparatus and system for multicast
US9448821B2 (en) Method and system for realizing virtual machine mobility
EP2982097B1 (en) Method and apparatus for exchanging ip packets among network layer 2 peers
WO2014023255A1 (zh) 数据中心虚拟化网络地址的分发方法、系统及目录服务器
US9716687B2 (en) Distributed gateways for overlay networks
US20160191462A1 (en) Message forwarding in a virtual local area network
CN105763512A (zh) Sdn虚拟化网络的通信方法和装置
JP2015095784A (ja) 情報処理システム、情報処理システムの制御方法及び情報処理装置の制御プログラム
JP2014021979A (ja) 複数の仮想マシンを管理するための階層システム、方法、及びコンピュータプログラム
WO2013189059A1 (zh) 报文处理方法、装置、主机和网络系统
US9647902B2 (en) Virtualized network for virtualized guests as an independent overlay over a physical network
EP3038296B1 (en) Pool element status information synchronization method, pool register and pool element
WO2014075527A1 (zh) 网络虚拟边界设备间进行冗余备份的方法、设备及系统
US9763135B1 (en) Load balancing with mobile resources
US12021826B2 (en) Techniques for managing software defined networking controller in-band communications in a data center network
Black et al. Internet Engineering Task Force (IETF) T. Narten, Ed. Request for Comments: 7364 IBM Category: Informational E. Gray, Ed.

Legal Events

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

Ref document number: 13827911

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: 13827911

Country of ref document: EP

Kind code of ref document: A1