Network virtualization implementation method and device for cloud data center, equipment and storage medium

Abstract

The invention discloses a network virtualization implementation method, device and equipment oriented to a cloud data center and a storage medium, relates to the field of cloud computing, is applied to a cloud computing management platform, and comprises the following steps: allocating a corresponding target address for each target device based on the device type of each target device of the cloud data center; announcing the target address of each first network device to all network devices based on a preset routing switching mechanism to determine the whole network topology information of the cloud data center; the network equipment comprises first network equipment and second network equipment; the first network equipment is a leaf switch, and the second network equipment is a spin switch; and determining a topological relationship between the source server and the target server based on the target message by using the whole network topological information and the identifier mapping table, so as to forward the target message to the target server by using the first network device and the second network device based on the topological relationship. According to the invention, the overall transmission efficiency of a virtualization technology in a large-scale cloud data center environment is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of cloud computing, and particularly relates to a network virtualization implementation method and device for a cloud data center, equipment and a storage medium. BACKGROUND

[0002] Traditional cloud data centers usually rely on network virtualization schemes based on tunnel encapsulation (such as VXLAN (Virtual Extensible Local Area Network), NVGRE (Network Virtualization using Generic Routing Encapsulation), etc.). Although these methods achieve network isolation and multi-tenant support to some extent, they bring significant CPU processing overhead, encapsulation header overhead, and increased network delay due to the need for message encapsulation and forwarding at a higher protocol layer (network layer or transport layer).

[0003] To sum up, how to improve the overall transmission efficiency of virtualization technology in a large-scale cloud data center environment is a problem to be solved at present. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a network virtualization implementation method and device for a cloud data center, equipment and a storage medium, which can improve the overall transmission efficiency of virtualization technology in a large-scale cloud data center environment. The specific scheme is as follows:

[0005] In a first aspect, the present application discloses a network virtualization implementation method for a cloud data center, applied to a cloud computing management platform, comprising: allocating a corresponding target address to each target device based on the device type of each target device in the cloud data center; the device type includes network devices and edge virtual networks; the target address includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device; announcing the target address of each first network device to all network devices based on a preset routing exchange mechanism to determine the full network topology information of the cloud data center; the network devices include the first network devices and second network devices; the first network devices are leaf switches, and the second network devices are spine switches; determining the topology relationship between the corresponding source server and destination server based on the full network topology information and a pre-acquired identifier mapping table based on a target message, and forwarding the target message to the destination server based on the topology relationship using the first network devices and the second network devices.

[0006] Optionally, the preset routing exchange mechanism is used to announce the target address of each first network device to all the network devices to determine the whole network topology information of the cloud data center, comprising: determining all target second network devices corresponding to each first network device; the target second network device is a second network device connected to the first network device; establishing a neighbor relationship between each first network device and the corresponding target second network device; announcing the first address of the first network device to the corresponding target second network device, so as to record the routing information of the first network device to the local routing table of the target second network device by using the target second network device, and forward the routing information to other first network devices having a neighbor relationship with the target second network device, and record the routing information in the local routing table of the other first network devices; determining the whole network topology information of the cloud data center based on all the routing information.

[0007] Optionally, the identification mapping table is used to record the mapping relationship between the MAC address of the cloud server in the virtual network inside the current cloud data center and the first address of the first network device corresponding to the cloud server.

[0008] Optionally, the network virtualization method for the cloud data center further comprises: if a cloud server startup request is obtained, a target edge server is determined based on the cloud server startup request by using a preset server scheduling strategy; creating and starting a target cloud server corresponding to the cloud server startup request on the target edge server, and mounting the target cloud server to a corresponding target edge virtual network; announcing the virtual network information and MAC address information corresponding to the target cloud server to a first edge network device, so as to send the triple information corresponding to the target cloud server to a second edge network device by using the first edge network device; sending the triple information to other first network devices having a neighbor relationship with the second edge network device by using the second edge network device, and saving the triple information to the identification mapping table in the local of the other first network devices having a neighbor relationship with the second edge network device; The first edge network device is the first network device corresponding to the target edge server, the second edge network device is the second network device in a neighbor relationship with the first edge network device, and the triple information includes virtual network information corresponding to the target cloud server, MAC address information corresponding to the target cloud server, and the first address of the first edge network device.

[0009] Optionally, the topology relationship between the source server and the destination server is determined based on the target message by using the full-network topology information and the pre-acquired identifier mapping table, and the method comprises the following steps. The source server sends a target message to the first network device corresponding to the source server; the target message is a message encapsulated in a VLAN format; The first network device corresponding to the source server processes the target message to extract a second address in the target message from an outer Ethernet header of the target message; The first network device corresponding to the destination server of the target message is determined based on the full-network topology information and the pre-acquired identifier mapping table, so as to determine the topology relationship between the source server and the destination server; The topology relationship comprises a first topology relationship, a second topology relationship and a third topology relationship; the first topology relationship is that the source server and the destination server are in the same edge server; the second topology relationship is that the source server and the destination server are in different edge servers of the same first network device; and the third topology relationship is that the source server and the destination server are in different edge servers of different first network devices.

[0010] Optionally, the target message is forwarded to the destination server by using the first network device and the second network device based on the topology relationship, and the method comprises the following steps. If the topology relationship between the source server and the destination server is the first topology relationship, the target message is directly forwarded in the edge servers corresponding to the source server and the destination server; If the topology relationship between the source server and the destination server is the second topology relationship, the target message is forwarded to the edge server corresponding to the destination server by using the first network device corresponding to the source server, so that the outer encapsulation of the target message is removed by using the edge server corresponding to the destination server, and the obtained inner target message is forwarded to the destination server; if the topological relationship between the source server and the destination server is a third topological relationship, taking out an inner message of the target message by using the first network device corresponding to the source server, encapsulating the inner message based on a first preset message encapsulation format to obtain a new target message, and forwarding the new target message hop by hop to the first network device corresponding to the destination server, so as to re-encapsulate the target message based on a second preset message encapsulation format by using the first network device corresponding to the destination server, and send the re-encapsulated target message to an edge server corresponding to the destination server, so as to send the re-encapsulated target message to the destination server by using the edge server corresponding to the destination server.

[0011] Optionally, the forwarding of the target message to the destination server by using the first network device and the second network device based on the topological relationship comprises: forwarding the target message to the destination server by using the first network device and the second network device based on the topological relationship and a preset load balancing strategy.

[0012] In a second aspect, the present application discloses a network virtualization implementation device for a cloud data center, applied to a cloud computing management platform, comprising: an address allocation module, configured to allocate a corresponding target address to each target device based on a device type of each target device of the cloud data center; the device type comprises a network device and an edge virtual network; the target address comprises a first address used to identify position information of the target device and a second address used to identify edge virtual network information of the target device; a topological information determination module, configured to announce the target address of each first network device to all network devices based on a preset routing switching mechanism to determine full-network topological information of the cloud data center; the network device comprises the first network device and a second network device; the first network device is a leaf switch, and the second network device is a spine switch; a message forwarding module, configured to determine a topological relationship between a corresponding source server and a destination server based on a target message by using the full-network topological information and a pre-acquired identifier mapping table, and forward the target message to the destination server by using the first network device and the second network device based on the topological relationship.

[0013] In a third aspect, the present application discloses an electronic device, comprising: a memory, configured to save a computer program; a processor, configured to execute the computer program to implement the network virtualization implementation method for a cloud data center.

[0014] In a fourth aspect, the present application discloses a computer readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the cloud data center oriented network virtualization implementation method.

[0015] In the present application, when implementing the cloud data center oriented network virtualization, the cloud computing management platform allocates a corresponding target address to each target device based on the device type of each target device in the cloud data center; the device type includes a network device and an edge virtual network; the target address includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device; the target address of each first network device is announced to all network devices based on a preset routing exchange mechanism to determine the full network topology information of the cloud data center; the network device includes the first network device and a second network device; the first network device is a leaf switch, and the second network device is a spine switch; the topology relationship between the source server and the destination server is determined based on a target message by using the full network topology information and a pre-acquired identification mapping table, so that the target message is forwarded to the destination server by using the first network device and the second network device based on the topology relationship. It can be seen that, in the present application, the cloud computing management platform allocates a corresponding target address to each target device based on the device type of each target device in the cloud data center, and the target address includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device, thereby redefining the network address structure inside the cloud data center and supporting fine management and efficient addressing of large-scale virtual networks. In addition, the cloud computing management platform announces the target address of each first network device to all network devices based on a preset routing exchange mechanism to determine the full network topology information of the cloud data center, thereby realizing fast synchronization and automatic routing generation of the full network topology and avoiding the problems of uncertainty and slow convergence caused by MAC learning in the traditional Ethernet. Then, the cloud computing management platform can determine the topology relationship between the source server and the destination server based on a target message by using the full network topology information and an identification mapping table, so as to forward the target message to the destination server based on the topology relationship. Through the above process, the network virtualization function is originally integrated into the network protocol design, the construction and operation process of the virtual network is comprehensively optimized, and the overall transmission efficiency and resource utilization rate of the cloud data center network are improved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings described below are only a part of the embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0017] Figure 1 A flow chart of a network virtualization implementation method for a cloud data center is disclosed in the present application. Figure 2 A specific internal implementation architecture diagram of network virtualization is disclosed in the present application. Figure 3 A specific target address structure diagram is disclosed in the present application. Figure 4 A typical cloud data center network topology diagram is disclosed in the present application. Figure 5 A specific message header format diagram is disclosed in the present application. Figure 6 A specific message format diagram issued by an edge server is disclosed in the present application. Figure 7 A specific encapsulated frame structure diagram is disclosed in the present application. Figure 8 A network protocol stack change diagram is disclosed in the present application. Wherein, Figure 8 a is a network protocol stack diagram of a traditional network virtualization implementation method based on a transport layer tunneling mechanism, Figure 8 b is a network protocol stack diagram of the network virtualization implementation method provided in the present embodiment. Figure 9 A network virtualization implementation device structure diagram for a cloud data center is disclosed in the present application. Figure 10 An electronic device structure diagram is disclosed in the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0019] Traditional cloud data center usually relies on network virtualization scheme based on tunnel encapsulation (such as VXLAN, NVGRE, etc.). These methods, although to some extent, realize network isolation and multi-tenant support, but due to the need for packet encapsulation and forwarding at a higher protocol layer (network layer or transport layer), it brings significant CPU processing overhead, encapsulation header overhead, and increased network delay. In order to solve the above technical problems, the application discloses a network virtualization implementation method for cloud data center, which can improve the overall transmission efficiency of virtualization technology in large-scale cloud data center environment.

[0020] Referring to Figure 1 The embodiment of the application discloses a network virtualization implementation method for cloud data center, applied to a cloud computing management platform, comprising:

[0021] Step S11, based on the device type of each target device of the cloud data center, the target device is allocated a corresponding target address; the device type includes network device and edge virtual network; the target address includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device.

[0022] In this embodiment, as Figure 2 As shown in the figure, in order to simplify the architecture of the internal virtual network of the edge server, only one virtual network, virtual switch virtual switch, is reserved, and the cloud server of the tenant will be directly mounted on the virtual switch, and different local VLAN (Virtual Local Area Network, Virtual Local Area Network) is used to distinguish different virtual networks. This embodiment avoids the problems of complex configuration, loop occurrence and other problems caused by multiple virtual bridge connections by reserving only a single virtual switch in the edge server and distinguishing different virtual networks through VLAN tags, significantly improving the deployment efficiency and network security.

[0023] In this embodiment, the cloud computing management platform allocates a corresponding target address to each target device based on the device type of each target device of the cloud data center, wherein the device type includes network device and edge virtual network; the target address (i.e. target device address) includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device. As Figure 3The structure of a target address is shown, which includes three parts. The first part is a reserved field (Reserved) with 4 bits, which is all 0 by default. The second part is a Locator ID, also called the first address, with 48 bits, which is used for routing and forwarding in the cloud data center. The third part is an Edge ID, also called the second address, with 12 bits. The 48-bit Locator ID follows the semantics of the Ethernet MAC address, including unicast, multicast and broadcast semantics. The 12-bit Edge ID has two cases. One is the network device itself, such as a spine switch and a leaf switch, whose Edge ID is all 0, i.e. 0x000H. The other is an edge virtual network, whose Edge ID is allocated by the leaf switch connected to the device, which is used to identify the identity of the virtual network and needs to be mapped to a globally unique virtual network ID. The Edge ID is only valid in the management domain of the current switch and does not have global semantics. In addition, there is a special case, which is a traditional server that does not support access to the cloud data center network. The format of its network address is similar to that of the switch. The high 48 bits are the MAC address of the device itself, and the low 12 bits are all 0, i.e. 0x000H.

[0024] In this embodiment, the virtual network has three different representations in the protocol architecture of this embodiment. In the edge server, the virtual network is marked by the VLAN on the virtual switch. When the cloud servers in different virtual networks are mounted to the virtual switch, they will be labeled with the VLAN tag corresponding to the virtual network. On the uplink leaf switch of the edge cloud server, the virtual network can be identified by the Edge ID. After leaving the leaf switch, the virtual network is identified by a 32-bit global Network ID, which is allocated by the cloud computing management platform and transmitted with the packet during packet forwarding. The virtual network has three different representations in the protocol architecture of this embodiment. In the edge server, the virtual network is marked by the VLAN on the virtual switch. When the cloud servers in different virtual networks are mounted to the virtual switch, they will be labeled with the VLAN tag corresponding to the virtual network. On the uplink leaf switch of the edge cloud server, the virtual network can be identified by the Edge ID. After leaving the leaf switch, the virtual network is identified by a 32-bit global Network ID, which is allocated by the cloud computing management platform and transmitted with the packet during packet forwarding. By integrating the virtual network identification information in the protocol design stage, only a lightweight encapsulation is required, which significantly reduces the encapsulation header redundancy brought by traditional tunneling protocols (such as VXLAN, NVGRE), thereby improving the effective load ratio and improving the overall network transmission efficiency.

[0025] Step S12, announcing the target address of each first network device to all the network devices based on the preset routing exchange mechanism to determine the full-network topology information of the cloud data center; the network devices include the first network devices and second network devices; the first network devices are leaf switches, and the second network devices are spine switches.

[0026] In the embodiment, as shown in Figure 4 A typical cloud data center topology diagram is shown. Unlike the Internet and enterprise networks and other complex scenarios, in the cloud data center, the routing exchange mechanism can be greatly simplified to obtain a preset routing exchange mechanism due to the determinacy and stability of the network topology. The network devices in the cloud data center have only two roles, i.e., the first network devices leaf switches and the second network devices spine switches, and the ultimate goal of route generation is to announce the location information of each leaf switch to all network devices in the network. On this premise, the target address of each first network device is announced to all network devices based on the preset routing exchange mechanism to determine the full-network topology information of the cloud data center, which can specifically include: determining all target second network devices corresponding to each first network device; the target second network device is a second network device connected to the first network device; establishing a neighbor relationship between each first network device and the corresponding target second network device; announcing the first address of the first network device to the corresponding target second network device, so as to record the routing information of the first network device to the local routing table of the target second network device, and forward the routing information to other first network devices having a neighbor relationship with the target second network device, and record the routing information in the local routing table of the other first network devices; determining the full-network topology information of the cloud data center based on all the routing information.

[0027] That is, in the embodiment, each leaf switch establishes a neighbor relationship with all spine switches connected thereto, and the leaf switch announces its Locator ID to all neighbors, i.e., spine switches; after receiving the address announcement of the leaf switch, the spine switch records it in the local routing table, and continues to announce the address information to all other leaf switches having a neighbor relationship; after receiving the address announcement of other leaf switches of the spine switch, the leaf switch records it in the local routing table. The routing table format is shown in Table 1. The embodiment quickly synchronizes the full-network routing information by using the simplified neighbor relationship establishment and Locator ID announcement mechanism, avoids the uncertainty and slow convergence problem based on MAC table learning, and improves the stability and reliability of the cloud data center network.

[0028] Table 1 Routing table format

[0029]

[0030] It should be noted that the leaf switch does not forward the address announcement from the spine switch to avoid routing loops. Meanwhile, in some special cases, if there is a connection between the spine switches, the spine switches also need to establish a neighbor relationship, which is generally a non-standard network. In addition, when the leaf switch is connected with a traditional device, such as a traditional physical server (which does not support the new cloud data center network protocol), the leaf switch will act as a proxy and announce the 48-bit MAC address of the device as a Locator ID to the spine, so that the entire network can generate routing information to the traditional device. When all leaf switches and spine switches perform routing information announcement according to the above-mentioned preset routing switching mechanism, complete network topology information can be generated on each switch (including all leaf switches and spine switches), and each switch can select routing according to the 48-bit Locator ID, thereby greatly improving the routing scalability and meeting the routing requirements of the cloud data center. By introducing the new addressing structure of 48-bit Locator ID and 12-bit Edge ID, the present application realizes efficient identification and isolation of large-scale nodes and virtual networks in the cloud data center, breaks through the limitation of the number of traditional VLANs, and meets the needs of large-scale cloud environments for multi-tenant isolation and resource management.

[0031] It can be understood that in the embodiment, if the cloud computing management platform obtains a cloud server startup request, the target edge server is determined based on the cloud server startup request using a preset server scheduling strategy; the target cloud server corresponding to the cloud server startup request is created and started on the target edge server, and the target cloud server is mounted to the corresponding target edge virtual network; the virtual network information and the MAC address information corresponding to the target cloud server are announced to the first edge network device, so as to send the triple information corresponding to the target cloud server to the second edge network device using the first edge network device; the triple information is sent to other first network devices having a neighbor relationship with the second edge network device using the second edge network device, and the triple information is saved to the identity mapping table local to the other first network devices having a neighbor relationship with the second edge network device; wherein the first edge network device is a first network device corresponding to the target edge server, the second edge network device is a second network device having a neighbor relationship with the first edge network device, and the triple information includes the virtual network information corresponding to the target cloud server, the MAC address information corresponding to the target cloud server, and the first address of the first edge network device.

[0032] That is, when a user purchases a cloud server in a virtual network through a cloud computing console, the cloud computing platform will create and start a cloud server on an edge server according to a corresponding scheduling strategy, and connect to the corresponding virtual network. After creation, the virtual network information (Network ID) and MAC address information of the cloud server need to be announced to the edge leaf switch, the edge leaf switch records the information, and announces the Network ID, MAC address and Locator ID of the leaf switch as a triple to all neighbor spine switches. After receiving the triple information, the neighbor spine switch records it in its local mapping table, and also announces the triple information to other neighbor leaf switches. After receiving the triple, the other neighbor leaf switches first judge whether there is a cloud server with the same Network ID in the edge server under the leaf switch, if there is, record the triple in the local virtual network identifier mapping table, if not, ignore the announcement. The identifier mapping table is used to record the mapping relationship between the MAC address of the cloud server in the virtual network inside the cloud data center and the first address of the first network device corresponding to the cloud server. By generating the full network topology information and complete routing table in advance, the session establishment delay is effectively reduced.

[0033] Step S13, based on the full network topology information and the pre-acquired identifier mapping table, determine the topology relationship between the corresponding source server and the destination server based on the target message, and use the first network device and the second network device to forward the target message to the destination server based on the topology relationship.

[0034] In the embodiment, the topology relationship between the source server and the destination server is determined based on the target message by using the full network topology information and the pre-acquired identifier mapping table, including: sending the target message by the source server to the first network device corresponding to the source server; the target message is a message encapsulated in VLAN format; processing the target message by the first network device corresponding to the source server to extract the second address in the target message from the outer Ethernet header of the target message; determining the first network device corresponding to the destination server of the target message based on the full network topology information and the pre-acquired identifier mapping table to determine the topology relationship between the source server and the destination server. The topology relationship includes a first topology relationship, a second topology relationship and a third topology relationship; the first topology relationship is that the source server and the destination server are in the same edge server, the second topology relationship is that the source server and the destination server are in different edge servers of the same first network device; the third topology relationship is that the source server and the destination server are in different edge servers of different first network devices. The message header format of the target message is as shown in Figure 5 The message header is composed of three parts, namely Destination, Source and Network ID. The Destination identifies the destination information of the target message, the Source identifies the source of the message, and the Network ID identifies the virtual network information carried by the message.

[0035] That is, in the embodiment, when two cloud servers in the virtual network communicate, there are three cases: the source server and the destination server are in the same edge server, the source server and the destination server are in different edge servers of the same leaf switch, and the source server and the destination server are in different edge servers under different leaf switches. The three cases correspond to the first topology relationship, the second topology relationship and the third topology relationship respectively.

[0036] In a specific embodiment, if the topological relationship between the source server and the destination server is a first topological relationship, the target message is forwarded directly within the edge servers corresponding to the source server and the destination server. If the topological relationship between the source server and the destination server is a second topological relationship, the target message is forwarded to the edge server corresponding to the destination server using the first network device corresponding to the source server, and the outer encapsulation of the target message is removed by the edge server corresponding to the destination server, and the obtained inner target message is forwarded to the destination server. If the topological relationship between the source server and the destination server is a third topological relationship, the inner message of the target message is extracted by the first network device corresponding to the source server, and the inner message is encapsulated based on a first preset message encapsulation format to obtain a new target message. The new target message is forwarded hop by hop to the first network device corresponding to the destination server, and the target message is re-encapsulated based on a second preset message encapsulation format by the first network device corresponding to the destination server, and the re-encapsulated target message is sent to the edge server corresponding to the destination server, and the re-encapsulated target message is sent to the destination server by the edge server corresponding to the destination server.

[0037] That is to say, if the source server and the destination server are both located inside the same edge server, the communication between the two can be achieved directly inside the edge server. If the source server and the destination server are inside different edge servers of the same leaf switch, the message sent by the source server will be added with an outer header by the virtual switch in the edge server and sent to the leaf switch. If the leaf switch determines that its destination is located in other local edge servers, it will directly forward the message to the edge server corresponding to the destination server. After receiving the message, the edge server corresponding to the destination server removes the outer header encapsulation and hands the inner message to the destination server. At this point, the sending and receiving process is completed. In a specific embodiment, in order to maximize the solution to compatibility issues, the message sent from the edge server to the leaf switch is encapsulated in the traditional VLAN format, that is, before the message of the virtual network is sent to the leaf switch, the virtual switch on the edge physical server where it is located adds an IEEE802.1q outer header to the Ethernet message header, such as Figure 6 As shown in the figure, the VLAN ID in the VLAN header is the Edge ID (virtual network ID) assigned by the leaf switch, while the outer Destination MAC and Source MAC addresses are directly copied from the inner MAC addresses. When the packet reaches the leaf switch, it extracts the VLAN ID from the outer Ethernet header, interprets it as the Edge ID, and queries the global Network ID of the virtual network corresponding to the Edge ID.

[0038] In this embodiment, if the topological relationship between the source server and the destination server is the third topological relationship, i.e., the source server and the destination server are located in different edge servers under different leaf switches, the message sent by the source server will be added with an outer header as shown in Figure 6 by the virtual switch in the edge server and sent to the leaf switch. After the message reaches the leaf switch, the virtual network identifier mapping table (as shown in Table 2 below) is queried. When it is found that the destination of the message is located in a different leaf switch, the inner message content is extracted, and a message header as shown in Figure 7 is re-encapsulated in the outer layer based on the first preset message encapsulation format, i.e., the inner Ethernet frame structure in the original message from the edge server is extracted and encapsulated in the frame structure of the new cloud network data center. The Network ID is a global virtual network identifier, the Destination is the first address of the first network device corresponding to the destination server, and the Source is the first address of the first network device corresponding to the source server. Through hop-by-hop forwarding, the target message is finally sent to the leaf switch corresponding to the destination server. After the leaf switch corresponding to the destination server receives the message, the outer encapsulation header is removed, and an outer header as shown in Figure 6 is re-attached based on the second preset message encapsulation format, wherein the VLAN field is the Edge ID of the virtual network under the leaf switch, and the MAC of the source server and the destination server is directly copied from the inner header. After encapsulation, the target message is sent to the edge server corresponding to the destination server, and the virtual switch in the edge server sends the target message to the final destination cloud server, and the message transmission process is completed.

[0039] Table 2 Virtual network identifier mapping table format

[0040]

[0041] It should be noted that, Figure 7 the high 48 bits of the Destination field in the outer header of the message in the embodiment are the Locator ID of the leaf switch corresponding to the destination server, and the low 12 bits are all 0. The high 48 bits of the Source field are the Locator ID of the leaf switch corresponding to the source server, and the low 12 bits are all 0. In the subsequent protocol expansion, the low 12 bits can be considered.

[0042] It should be noted that, by adopting the direct encapsulation and forwarding mechanism based on the data link layer, the embodiment avoids the complex flow process of the traditional tunnel virtualization technology which needs to be frequently sent to the network layer and the transport layer for processing, significantly reduces the CPU processing overhead of the edge server and the intermediate network device, reduces the forwarding delay, and improves the communication performance. As shown in Figure 8 Figure 8 a is a network protocol stack schematic diagram of a traditional network virtualization implementation method based on a transport layer tunnel mechanism, Figure 8 b is a network protocol stack schematic diagram of the network virtualization implementation method provided by the embodiment. By comparison, it can be known that, compared with the traditional network virtualization implementation method based on the transport layer tunnel mechanism, the network virtualization implementation method disclosed by the embodiment has great optimization at the protocol stack processing level of the edge server and the intermediate network device.

[0043] It can be understood that, based on the topology relationship, the first network device and the second network device are used to forward the target message to the destination server, which comprises: based on the topology relationship and a preset load balancing strategy, the first network device and the second network device are used to forward the target message to the destination server. Specifically, in order to simplify the traffic scheduling strategy of the network and reduce the probability of network congestion, for the message forwarding of equivalent routes, a full-flow load balancing strategy is adopted, that is, if there are N equivalent routes to a certain destination, the traffic is evenly sent to the next hop corresponding to the N equivalent routes. When all nodes adopt this balanced forwarding strategy, the traffic of the entire network will tend to be balanced, and the probability of congestion will be greatly reduced.

[0044] It can be seen that, in the cloud computing management platform in the present application, the device types of each target device based on the cloud data center are used to allocate corresponding target addresses to each target device, the target address includes a first address used to identify the position information of the target device and a second address used to identify the edge virtual network information of the target device, thereby redefining the network address structure inside the cloud data center, supporting fine management and efficient addressing of large-scale virtual networks. In addition, the cloud computing management platform announces the target address of each first network device to all network devices based on a preset routing exchange mechanism to determine the whole network topology information of the cloud data center, realizes the rapid synchronization and automatic routing generation of the whole network topology, and avoids the uncertainty and slow convergence problem caused by MAC learning in the traditional Ethernet. Then the cloud computing management platform can determine the topology relationship between the source server and the destination server based on the target message by using the whole network topology information and the identification mapping table, and forward the target message to the destination server based on the topology relationship. Through the above process, the network virtualization function is originally integrated into the network protocol design, the construction and operation process of the virtual network is comprehensively optimized, and the overall transmission efficiency and resource utilization rate of the cloud data center network are improved. ​

[0045] Referring to Figure 9 As shown in the drawings, the present application discloses a network virtualization implementation device for cloud data center, applied to a cloud computing management platform, comprising: An address allocation module 11 is configured to allocate a corresponding target address to each target device based on the device type of each target device in the cloud data center; the device type includes a network device and an edge virtual network; the target address includes a first address for identifying the location information of the target device and a second address for identifying the edge virtual network information of the target device; A topology information determination module 12 is configured to announce the target address of each first network device to all network devices based on a preset routing switching mechanism to determine the full-network topology information of the cloud data center; the network devices include the first network devices and second network devices; the first network devices are leaf switches, and the second network devices are spine switches; A packet forwarding module 13 is configured to determine the topology relationship between the corresponding source server and destination server based on a target packet by using the full-network topology information and a pre-acquired identifier mapping table, and to forward the target packet to the destination server by using the first network devices and the second network devices based on the topology relationship.

[0046] In a specific embodiment, the topology information determination module 12 can specifically include: A device determination unit is configured to determine all target second network devices corresponding to each first network device; the target second network devices are second network devices connected to the first network devices; A neighbor relationship establishment unit is configured to establish a neighbor relationship between each first network device and the corresponding target second network device; A routing information recording unit is configured to announce the first address of the first network device to the corresponding target second network device, to record the routing information of the first network device to the local routing table of the target second network device by using the target second network device, and to forward the routing information to other first network devices having a neighbor relationship with the target second network device, and to record the routing information in the local routing table of the other first network devices; A full-network topology information determination unit is configured to determine the full-network topology information of the cloud data center based on all the routing information.

[0047] In a specific embodiment, the device can further include: An edge server determination module is configured to determine a target edge server based on a cloud server startup request by using a preset server scheduling strategy if the cloud server startup request is acquired. a cloud server mounting module, configured to create and start a target cloud server corresponding to the cloud server starting request on the target edge server, and mount the target cloud server to a target edge virtual network corresponding thereto; a triple information sending module, configured to notify virtual network information and MAC address information corresponding to the target cloud server to a first edge network device, so as to send triple information corresponding to the target cloud server to a second edge network device by using the first edge network device; a triple information saving module, configured to send the triple information to other first network devices in a neighbor relationship with the second edge network device by using the second edge network device, and save the triple information to the identity mapping table local to the other first network devices in the neighbor relationship with the second edge network device; The first edge network device is the first network device corresponding to the target edge server, the second edge network device is the second network device in a neighbor relationship with the first edge network device, and the triple information includes virtual network information corresponding to the target cloud server, MAC address information corresponding to the target cloud server, and the first address of the first edge network device.

[0048] In a specific embodiment, the packet forwarding module 13 can specifically include: a first packet sending unit, configured to send a target packet to the first network device corresponding to a source server by using the source server; the target packet is a packet encapsulated in a VLAN format; an address extracting unit, configured to process the target packet by using the first network device corresponding to the source server, so as to extract a second address in the target packet from an outer Ethernet header of the target packet; a topology relationship determining unit, configured to determine the first network device corresponding to a destination server of the target packet based on the whole-network topology information and a pre-acquired identity mapping table, so as to determine a topology relationship between the source server and the destination server; The topology relationship includes a first topology relationship, a second topology relationship, and a third topology relationship; the first topology relationship is that the source server and the destination server are located in a same edge server, the second topology relationship is that the source server and the destination server are located in different edge servers of a same first network device, and the third topology relationship is that the source server and the destination server are located in different edge servers of different first network devices.

[0049] In a specific implementation, the packet forwarding module 13 can specifically include: The first packet forwarding unit is configured to, if the topological relationship between the source server and the destination server is a first topological relationship, directly perform the forwarding of the target packet within the edge server corresponding to the source server and the destination server. The second packet forwarding unit is configured to, if the topological relationship between the source server and the destination server is a second topological relationship, forward the target packet to the edge server corresponding to the destination server by using the first network device corresponding to the source server, so as to remove the outer encapsulation of the target packet by using the edge server corresponding to the destination server, and forward the obtained inner target packet to the destination server. The third packet forwarding unit is configured to, if the topological relationship between the source server and the destination server is a third topological relationship, extract the inner packet of the target packet by using the first network device corresponding to the source server, encapsulate the inner packet based on a first preset packet encapsulation format to obtain a new target packet, and forward the new target packet hop by hop to the first network device corresponding to the destination server, so as to re-encapsulate the target packet based on a second preset packet encapsulation format by using the first network device corresponding to the destination server, and send the re-encapsulated target packet to the edge server corresponding to the destination server, so as to send the re-encapsulated target packet to the destination server by using the edge server corresponding to the destination server.

[0050] In a specific implementation, the packet forwarding module 13 can specifically include: The fourth packet forwarding unit is configured to forward the target packet to the destination server by using the first network device and the second network device based on the topological relationship and a preset load balancing strategy.

[0051] Further, the application also discloses an electronic device, Figure 10 The electronic device 20 is shown in the structure diagram according to an exemplary embodiment, and the content in the diagram cannot be considered as any limitation on the use range of the application.

[0052] Figure 10A structural schematic diagram of an electronic device 20 is provided in the embodiments of the present application. The electronic device 20 can specifically include at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25 and a communication bus 26. The memory 22 is configured to store a computer program, and the processor 21 is configured to load and execute the computer program to implement the related steps in the cloud data center oriented network virtualization implementation method disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in the embodiments of the present application can be specifically an electronic computer.

[0053] In the embodiments of the present application, the power supply 23 is configured to provide working voltage for each hardware device on the electronic device 20; the communication interface 24 is capable of creating a data transmission channel between the electronic device 20 and external devices, and the communication protocol followed by the communication interface 24 can be any communication protocol applicable to the technical solution of the present application, which is not specifically limited here; the input / output interface 25 is configured to obtain external input data or output data to the outside, and the specific interface type can be selected according to the specific application needs, which is not specifically limited here.

[0054] In addition, the memory 22 as a carrier for resource storage can be a read-only memory, a random access memory, a magnetic or optical disk, etc., and the resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage mode can be temporary storage or permanent storage.

[0055] The operating system 221 is configured to manage and control each hardware device on the electronic device 20 and the computer program 222, and can be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program capable of completing the cloud data center oriented network virtualization implementation method executed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 can further include a computer program capable of completing other specific work.

[0056] Further, the present application further discloses a computer readable storage medium for storing a computer program; wherein the computer program is executed by a processor to implement the cloud data center oriented network virtualization implementation method disclosed in the foregoing embodiments. For the specific steps of the method, refer to the corresponding contents disclosed in the foregoing embodiments, which will not be repeated here.

[0057] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. For the same or similar parts between the embodiments, refer to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts refer to the method part.

[0058] Those skilled in the art will further appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or any combination thereof. To clearly illustrate the interchangeability of hardware and software, various components have been described above generally in terms of their functionality, without referring to the details of their implementation. Whether such functionality is implemented in hardware or software depends on the particular application and design constraints imposed on the overall system. Skilled persons can implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

[0059] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in random access memory (RAM), flash memory, read-only memory (ROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0060] Finally, it should be noted that the terms "first", "second", and the like, herein do not denote any order, quantity, combination, or importance, but rather are used to distinguish one element from another, and are not necessarily intended to denote the priority of one element over another. Also, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0061] The above has introduced the technical solutions provided by the present application in detail, and the principles and implementation manners of the present application have been described by applying specific examples; the above example descriptions are only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range can be changed; in view of the above, the content of the present description should not be understood as limiting the present application.

Claims

1. A network virtualization implementation method for a cloud data center, characterized in that: Applied to cloud computing management platforms, including: Allocating a corresponding target address to each target device based on the device type of each target device in the cloud data center; the device type includes a network device and an edge virtual network; the target address includes a first address for identifying location information of the target device and a second address for identifying edge virtual network information of the target device; Based on a preset routing exchange mechanism, the target address of each first network device is announced to all the network devices to determine the entire network topology information of the cloud data center; the network devices include the first network device and the second network device; the first network device is a leaf switch, and the second network device is a spine switch; Utilizing the entire network topology information and the pre-acquired identification mapping table, the topological relationship between the corresponding source server and the destination server is determined based on the target message, so as to forward the target message to the destination server based on the topological relationship using the first network device and the second network device.

2. The network virtualization implementation method for cloud data centers according to claim 1, characterized in that: The declaring the target address of each first network device to all the network devices based on a preset routing exchange mechanism to determine the entire network topology information of the cloud data center includes: Determine all target second network devices corresponding to each first network device; the target second network device is a second network device connected to the first network device; Establishing a neighbor relationship between each of the first network devices and the corresponding target second network device; announcing the first address of the first network device to the corresponding target second network device, so that the target second network device records the routing information of the first network device in a local routing table of the target second network device, and forwards the routing information to other first network devices that have a neighbor relationship with the target second network device, and records the routing information in the local routing tables of the other first network devices; The entire network topology information of the cloud data center is determined based on all the routing information.

3. The network virtualization implementation method for cloud data centers according to claim 1, characterized in that: The identification mapping table is used to record the mapping relationship between the MAC address of the cloud server in the virtual network within the current cloud data center and the first address of the first network device corresponding to the cloud server.

4. The method for implementing network virtualization for a cloud data center according to claim 3, wherein: Also includes: If a cloud server startup request is obtained, a target edge server is determined based on the cloud server startup request using a preset server scheduling strategy; Creating and starting a target cloud server corresponding to the cloud server startup request on the target edge server, and mounting the target cloud server to the corresponding target edge virtual network; Notify the virtual network information and MAC address information corresponding to the target cloud server to the first edge network device, so that the first edge network device can send the triplet information corresponding to the target cloud server to the second edge network device; Using the second edge network device to send the triplet information to the other first network devices that have a neighbor relationship with the second edge network device, and saving the triplet information to the local identifier mapping table of the other first network devices that have a neighbor relationship with the second edge network device; Among them, the first edge network device is the first network device corresponding to the target edge server, the second edge network device is the second network device that has a neighbor relationship with the first edge network device, and the triplet information includes the virtual network information corresponding to the target cloud server, the MAC address information corresponding to the target cloud server, and the first address of the first edge network device.

5. The method for implementing network virtualization for a cloud data center according to claim 1, wherein: The determining of the topological relationship between the corresponding source server and the destination server based on the target message by using the entire network topology information and the pre-acquired identification mapping table includes: Using a source server to send a target message to the first network device corresponding to the source server; the target message is a message encapsulated in a VLAN format; Processing the target message using the first network device corresponding to the source server to extract the second address in the target message from an outer Ethernet header of the target message; Determining the first network device corresponding to the destination server of the target message based on the entire network topology information and the pre-acquired identifier mapping table, so as to determine the topological relationship between the source server and the destination server; Among them, the topological relationship includes a first topological relationship, a second topological relationship and a third topological relationship; the first topological relationship is that the source server and the destination server are located inside the same edge server, the second topological relationship is that the source server and the destination server are located inside different edge servers of the same first network device; the third topological relationship is that the source server and the destination server are located inside different edge servers of different first network devices.

6. The method for implementing network virtualization for a cloud data center according to claim 5, wherein: The forwarding the target message to the destination server by using the first network device and the second network device based on the topological relationship includes: If the topological relationship between the source server and the destination server is the first topological relationship, the target message is directly forwarded within the edge servers corresponding to the source server and the destination server; If the topological relationship between the source server and the destination server is the second topological relationship, forwarding the target message to the edge server corresponding to the destination server by using the first network device corresponding to the source server, removing the outer encapsulation of the target message by using the edge server corresponding to the destination server, and forwarding the obtained inner target message to the destination server; If the topological relationship between the source server and the destination server is the third topological relationship, the inner layer message of the target message is taken out by using the first network device corresponding to the source server, and the inner layer message is encapsulated based on the first preset message encapsulation format to obtain the new target message, and the new target message is forwarded hop by hop to the first network device corresponding to the destination server, so that the first network device corresponding to the destination server is used to re-encapsulate the target message based on the second preset message encapsulation format, and the re-encapsulated target message is sent to the edge server corresponding to the destination server, so that the edge server corresponding to the destination server is used to send the re-encapsulated target message to the destination server.

7. The method for implementing network virtualization for a cloud data center according to any one of claims 1 to 6, characterized in that: The forwarding the target message to the destination server by using the first network device and the second network device based on the topological relationship includes: The target message is forwarded to the destination server using the first network device and the second network device based on the topology relationship and a preset load balancing strategy.

8. A network virtualization implementation device for a cloud data center, characterized in that: Applied to cloud computing management platforms, including: An address allocation module is configured to allocate a corresponding target address to each target device based on a device type of each target device in the cloud data center; the device type includes a network device and an edge virtual network; the target address includes a first address for identifying location information of the target device and a second address for identifying edge virtual network information of the target device; a topology information determination module, configured to announce the target address of each first network device to all the network devices based on a preset routing exchange mechanism to determine the entire network topology information of the cloud data center; the network devices include the first network device and the second network device; the first network device is a leaf switch, and the second network device is a spine switch; A message forwarding module is used to use the entire network topology information and a pre-acquired identification mapping table to determine the topological relationship between the corresponding source server and the destination server based on the target message, so as to forward the target message to the destination server based on the topological relationship using the first network device and the second network device.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the network virtualization implementation method for a cloud data center according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that Used to store a computer program, wherein when the computer program is executed by a processor, it implements the network virtualization implementation method for a cloud data center according to any one of claims 1 to 7.

Images