CN111225031B - Cloud data center virtual bottom layer network architecture and data transmission method thereof - Google Patents
Cloud data center virtual bottom layer network architecture and data transmission method thereof Download PDFInfo
- Publication number
- CN111225031B CN111225031B CN201911303921.4A CN201911303921A CN111225031B CN 111225031 B CN111225031 B CN 111225031B CN 201911303921 A CN201911303921 A CN 201911303921A CN 111225031 B CN111225031 B CN 111225031B
- Authority
- CN
- China
- Prior art keywords
- packet
- virtual node
- virtual
- destination
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/46—Interconnection of networks
- H04L12/4633—Interconnection of networks using encapsulation techniques, e.g. tunneling
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/46—Interconnection of networks
- H04L12/4641—Virtual LANs, VLANs, e.g. virtual private networks [VPN]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
- H04L47/2425—Traffic characterised by specific attributes, e.g. priority or QoS for supporting services specification, e.g. SLA
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L61/00—Network arrangements, protocols or services for addressing or naming
- H04L61/09—Mapping addresses
- H04L61/10—Mapping addresses of different types
- H04L61/103—Mapping addresses of different types across network layers, e.g. resolution of network layer into physical layer addresses or address resolution protocol [ARP]
Abstract
The application discloses virtual underlying network architecture of cloud data center, leaf ridge topological network that includes a plurality of leaf switches and a plurality of spine switches formation, wherein leaf switch includes website analysis unit and route selection unit, website analysis unit is used for analyzing this package after receiving the package that physical server sent, obtain the virtual node of source and/or the virtual node's of purpose characteristic information that contains in this package, route selection unit is used for confirming the corresponding spine switch of the business quality that reflects with characteristic information, and then confirms the network transmission route of package, so that leaf switch sends the package to the virtual node of purpose according to this network transmission route. The framework has the capability of analyzing the tunnel inner layer message, different forwarding paths are adopted for virtual node flows of different service applications in the underlying network tunnel by utilizing an analysis result, and the interference of the flow with low quality demand on the flow with high quality demand is avoided.
Description
Technical Field
The application relates to the technical field of network data transmission optimization, in particular to a cloud data center virtual underlying network architecture and a data transmission method based on the cloud data center virtual underlying network architecture.
Background
The Network architecture of modern cloud computing center is generally composed of an underlying Network (Underlay Network) and an Overlay Network (Overlay Network). The bottom layer network is a physical network constructed by traditional switches/routers and is used as a bottom layer bearing network for interconnection; the overlay network is a virtual network built by virtual nodes and virtual switching mechanisms in the physical server and is a virtual network built by applying a tunnel technology on the basis of interconnection and intercommunication of an underlying network.
Within different physical servers, communication between the virtual nodes is completed through virtual communication tunnels established among different physical servers of the cloud computing center, and the substantial communication of the virtual communication tunnels is performed through underlying network communication equipment.
When the source sends a message at the source physical switch, the source virtual switch searches the network address of the destination physical server where the destination virtual node is located. If the destination physical server and the source physical server are not the same physical server, the source virtual switch performs tunnel encapsulation on the message sent by the source virtual node, namely, a layer of packet header of an outer layer tunnel is additionally arranged in front of the original message. This outer-layer packet header includes the network address of the destination physical server and the network address of the source physical server. And then the source virtual switch sends the encapsulated tunnel packet to underlying network communication equipment connected with the network cable of the source physical server for forwarding. The communication equipment of the underlying network can forward the tunnel packet to a destination physical server according to the destination address of the outer packet head of the tunnel packet. When the destination virtual switch in the destination physical server receives the tunnel packet, it will take off the outer packet header of the encapsulated tunnel, and then forward it to the destination virtual node according to the destination address in the original message.
Since each virtual node within a physical server may host different business applications, some may be core big data applications, some may be AI smart learning applications, and some may be traditional email or ftp applications. Different applications may require different Quality of Service (QoS), for example, different applications have different requirements on packet loss rate and end-to-end delay.
However, the underlying network communication device forwards the packet according to the outer packet header of the tunnel, but does not detect the characteristics of the virtual nodes inside the tunnel, so that different network service qualities cannot be provided for applications of different virtual nodes, and thus the communication quality between end to end between virtual nodes of different physical servers cannot be guaranteed, and the service quality of service applications in the cloud data center is finally affected.
Disclosure of Invention
Object of the application
Based on this, in order to avoid the interference of the traffic with low quality demand on the traffic with high quality demand, further avoid packet loss and delay of the traffic with high quality demand, implement reasonable allocation of network resources for service application, and meanwhile, in order to perform differentiated control on the traffic with different quality of service according to the occurrence of congestion conditions to alleviate network congestion, directly prevent the occurrence of network congestion, and in order to increase the grasp on the status of network transmission process, the following technical solutions are disclosed in the present application.
(II) technical scheme
In one aspect, the present application provides a cloud data center virtual underlying network architecture, which is characterized by comprising: a leaf-spine topology network formed by a plurality of leaf switches and a plurality of spine switches; wherein the content of the first and second substances,
the leaf switch includes:
the network address analyzing unit is used for analyzing the packet after receiving the packet sent by the leaf switch to obtain the characteristic information of the source virtual node and/or the destination virtual node contained in the packet;
and the path selection unit is used for determining a ridge switch corresponding to the service quality reflected by the characteristic information and further determining a network transmission path of the packet, so that the leaf switch sends the packet to a destination virtual node according to the network transmission path.
In a possible implementation manner, the characteristic information of the source virtual node includes a web address of the source virtual node and/or characteristic information of other virtual nodes except the source virtual node; the characteristic information of the target virtual node comprises a website of the target virtual node and/or characteristic information of other virtual nodes except the target virtual node.
In one possible implementation, the architecture further includes:
the system comprises an identifier adding unit, a service quality judging unit and a service quality judging unit, wherein the identifier adding unit is used for adding an identifier capable of indicating a communication quality change condition to a packet which is sent by a source leaf switch and has a service quality lower than a set service quality when the communication quality of the source leaf switch is changed;
the identification feedback unit is used for identifying the identification contained in the packet after the destination virtual node receives the packet and sending a flow regulation instruction corresponding to the communication quality change condition indicated by the identification to the flow regulation unit;
and the flow regulating unit is used for receiving the flow regulating instruction and regulating the flow rate of the source virtual node of the corresponding service quality according to the flow regulating instruction.
In a possible embodiment, the website parsing unit is further configured to parse whether a destination virtual node included in the packet is a virtual destination virtual node; and the number of the first and second electrodes,
the leaf switch further comprises:
the node selection unit is used for acquiring each real virtual node corresponding to the virtual target virtual node of the packet and selecting the real target virtual node from each real virtual node;
and the path modifying unit is used for modifying the virtual destination virtual node of the packet into the selected real destination virtual node so as to ensure that the path selecting unit determines the network transmission path of the selected real destination virtual node.
In one possible embodiment, the spine switch comprises a first parameter adding unit for adding a first in-band network telemetry parameter of a packet into the packet before forwarding the packet to a leaf switch; and/or the presence of a gas in the gas,
the leaf switch comprises a second parameter adding unit, which is used for adding a second in-band network telemetry parameter of the packet into the packet before forwarding the packet to a destination physical server.
On the other hand, the application also provides a data transmission method based on the virtual underlying network architecture of the cloud data center, the method utilizes a leaf-spine topology network formed by a plurality of leaf switches and a plurality of spine switches to perform data transmission, and the method comprises the following steps:
analyzing the packet after receiving the packet sent by the leaf switch to obtain the characteristic information of the source virtual node and/or the destination virtual node contained in the packet;
and determining a spine switch corresponding to the service quality reflected by the characteristic information, and further determining a network transmission path of the packet, so that the leaf switch sends the packet to a destination virtual node according to the network transmission path.
In a possible implementation manner, the characteristic information of the source virtual node includes a web address of the source virtual node and/or characteristic information of other virtual nodes of the source server except the source virtual node; the characteristic information of the target virtual node comprises a website of the target virtual node and/or characteristic information of other virtual nodes except the target virtual node.
In one possible embodiment, the method further comprises:
when the communication quality of a source leaf switch changes, adding an identifier capable of indicating the change condition of the communication quality to a packet which is sent by the source leaf switch and has the service quality lower than the set service quality;
after receiving a packet, a destination virtual node identifies the identifier contained in the packet and sends a flow regulation instruction corresponding to the communication quality change condition indicated by the identifier to a flow regulation unit;
and receiving the flow regulating instruction, and regulating the flow rate of the source virtual node of the corresponding service quality according to the flow regulating instruction.
In a possible implementation manner, when the packet is parsed, whether a destination virtual node included in the packet is a virtual destination virtual node is also parsed; and the number of the first and second electrodes,
the method further comprises the following steps:
when the target virtual node is a virtual target virtual node, acquiring each real target virtual node corresponding to the packaged virtual target virtual node, and selecting one real target virtual node from each real target virtual node;
and modifying the virtual destination virtual node of the packet into the selected real destination virtual node so that the path selection unit determines the network transmission path of the selected real destination virtual node.
In one possible embodiment, the method further comprises:
adding a first in-band network telemetry parameter of a packet into the packet prior to forwarding the packet to a leaf switch; and/or the presence of a gas in the gas,
the second in-band network telemetry parameter of the packet is added to the packet before forwarding the packet to the destination physical server.
(III) advantageous effects
The utility model discloses a cloud data center virtual bottom layer network framework and data transmission method thereof, possess the ability of analytic tunnel inlayer message, utilize the analytic result to adopt different retransmission route to the virtual node flow of different business applications in the bottom layer network tunnel, make and not produce the influence between the business that has different requirements to network service quality, avoid the flow of low-quality demand to the interference that the flow of high-quality demand produced, and then avoid high-quality demand business to take place packet loss and time delay, guaranteed the communication quality between the end-to-end between the virtual node of different physical servers, make the realization network resource rational distribution of business application.
By setting the congestion alarm mark, congestion control of different degrees is provided for different service application virtual node flows in an underlying network tunnel, when the underlying network congestion occurs, congestion control is firstly carried out on low-priority flows, the high-priority flows are ensured to pass through without obstruction, and the influence of network congestion on high-priority service application data is avoided to the maximum extent.
By carrying out node selection and path modification, the data to be transmitted is sent to real target virtual nodes positioned on different physical servers, the situation that only one or too few real target virtual nodes cannot process service requests sent by a plurality of concurrent source virtual nodes at the same time is avoided, load balancing service is provided for virtual node flow in an underlying network tunnel, the situation that the virtual nodes are overloaded is prevented, and optimization of service processing is realized.
The method and the device realize the performance measurement of accurate fineness provided for the flow of the virtual nodes applied to different services in the tunnel of the underlying network, increase the grasp on the network transmission process condition, facilitate the analysis and optimization of the running performance of different services of the subdivided cloud data center, and facilitate the subsequent background analysis and strategy response aiming at optimization.
Drawings
The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining and illustrating the present application and should not be construed as limiting the scope of the present application.
Fig. 1 is a block diagram of a cloud data center virtual underlying network architecture in terms of virtual traffic slicing, which is disclosed in the present application.
Fig. 2 is a block diagram of a cloud data center virtual underlying network architecture in terms of virtual traffic congestion control, which is disclosed in the present application.
Fig. 3 is a block diagram of a cloud data center virtual underlying network architecture in terms of traffic load balancing.
Fig. 4 is a schematic flow chart of a data transmission method based on a cloud data center virtual underlying network architecture in terms of virtual traffic service slicing, disclosed by the application.
Detailed Description
In order to make the implementation objects, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described in more detail below with reference to the drawings in the embodiments of the present application.
Embodiments of a cloud data center virtual underlay network architecture disclosed herein are described in detail below with reference to fig. 1-3. The embodiment is mainly suitable for the fields of programmable switches/routers, SDNs (Software Defined networks) and the like, and can perform virtual traffic service slicing optimization on communication connections between virtual nodes in an underlying Network of cloud computing, for scenes such as data centers and campus networks.
As shown in fig. 1, the virtual underlying network architecture disclosed in this embodiment includes: a plurality of leaf switches and a plurality of spine switches that together form a leaf-spine topology network.
In a leaf-spine (leaf-spine) topology network, a switch deployed in a spine layer is called a spine switch and used for implementing a routing function, a switch deployed in a leaf layer is called a leaf switch, and the leaf switch is an access switch and is usually connected with a server, a storage device, a firewall, a boundary router and other terminal devices and used for aggregating traffic of a user. The spine switch is connected with each leaf switch of the leaf layer, the terminal equipment sends data to the corresponding leaf switch, and the leaf switches are communicated with the other leaf switch through the spine switch, so that data communication from the source physical server to the destination physical server is realized.
In the network architecture diagram shown in fig. 1, any one of the four spine switches 101,102,103,104 is connected to any one of the four leaf switches 111,112,113,114, and the four leaf switches are connected to four physical servers 121,122,123,124 one by one.
In the virtual underlying network architecture disclosed in this embodiment, part or all of the leaf switches include a website address resolution unit and a path selection unit.
The network address analyzing unit is used for analyzing the packet after receiving the packet sent by the leaf switch to obtain the characteristic information of the source virtual node and/or the destination virtual node contained in the packet.
Virtual nodes, also called virtual machines or containers. Specifically, in fig. 1, when a virtual node (i.e., a source virtual node) of a source physical server 121 needs to send a packet 191 obtained by tunnel encapsulation of service information to a virtual node (i.e., a destination virtual node) of a destination physical server 123, the source physical server 121 first sends the packet 191 to a leaf switch 111, and the leaf switch 111 first analyzes the packet 191 through a network address analysis unit to obtain feature information of the source virtual node included in a tunnel inner layer message, or obtain feature information of the destination virtual node included in the tunnel inner layer message, or obtain feature information of both the source virtual node and the destination virtual node.
In a cloud data center network composed of a plurality of spine switches and leaf switches, a tunnel sent out from a physical server may include traffic of a plurality of different virtual machine node service applications, the service quality of the service applications is not completely the same, and the characteristic information can reflect the corresponding service quality of data transmitted by the virtual nodes. The service quality may be a delay, a packet loss rate, a bandwidth size, stability, and the like. For example: for high-frequency transaction, virtual space-time (VR) and other types of services, the method is sensitive to time delay and packet loss rate; for stream services such as online video and the like, the method is sensitive to time delay and insensitive to packet loss rate; for big data processing services, the method is insensitive to time delay and sensitive to packet loss rate; for background services such as P2P download and FTP transmission, it is not sensitive to delay and packet loss rate. The service quality corresponding to the high-frequency transaction service is a high service quality, the service quality corresponding to the big data processing service is a medium service quality, and the service quality corresponding to the background service is a low service quality.
In one embodiment, the source virtual node characteristic information obtained by the website address parsing unit may include only the website address (e.g., IP address) of the source virtual node, or only the characteristic information (e.g., communication session port number) of the sessions of other virtual nodes besides the source virtual node, or both the website address and the characteristic information of the other virtual nodes. The destination virtual node feature information obtained by the website analyzing unit may also include only the website of the destination virtual node, or only the feature information of other virtual nodes besides the destination virtual node, or both the website and the feature information of other virtual nodes. According to different application scenes, the website analysis unit can correspondingly adjust what kind of characteristic information is specifically acquired.
Specifically, in fig. 1, when the leaf switch 111 receives the tunnel packet sent by the physical server 121, the leaf switch 111 may extract feature information according to the IP subnet of the tunnel inner layer packet virtual node, for example, extract the category (class a, class B, class C, etc.) of the network address as the feature information, and for example, extract N-bit numbers in the network address as the feature information. The leaf switch 111 may also determine the qos requirements of the virtual node traffic according to the characteristics of other virtual nodes, such as IP port number, IP protocol properties (TCP, UDP, FTP), and so on.
The path selection unit is used for determining a ridge switch corresponding to the service quality requirement reflected by the characteristic information, and further determining a network transmission path of the packet, so that the leaf switch sends the packet to the destination virtual node according to the network transmission path.
Specifically, after the website analyzing unit obtains the feature information, the path selecting unit selects a suitable spine switch from spine switches in the leaf-spine topology network to forward, and then sends the packet to the selected spine switch. The selection criteria of the spine switch is to select the spine switch corresponding to the service quality of the virtual node (source virtual node, destination virtual node, or both). The correspondence between the qos and the spine switches may be predetermined, for example, a virtual node that transmits a high qos packet corresponds to the spine switch 101, a virtual node that transmits a medium qos packet corresponds to the spine switch 102, and a virtual node that transmits a low qos packet corresponds to the spine switch 104.
It can be understood that the virtual nodes with the same level of service quality may simultaneously correspond to a plurality of spine switches, and select one spine switch when actually determining the network transmission path of the packet. For example, in fig. 1, both the spine switches 101 and 103 are switches for forwarding high-quality of service data, but the spine switch 101 is selected for forwarding in the above-mentioned forwarding scheme.
In addition, when selecting a plurality of spine switches corresponding to the same level of service quality, the selection may be performed according to the data transceiving capability of the spine switches, for example, the spine switch with the largest bandwidth may be selected from the plurality of spine switches corresponding to the high level of service quality according to the requirement for packet forwarding.
Since the source physical server (physical server 121) and the destination physical server (physical server 123) are determined, and the source leaf switch (leaf switch 111) and the destination leaf switch (leaf switch 113) are also determined, when the corresponding spine switch is determined, it is equivalent to determining the network data transmission path of the packet. For example, the transmission path of the high quality of service packets (traffic) at the leaf switch end may be leaf switch 111-spine switch 101-leaf switch 113 (as shown by the arrow in the figure), and similarly, the medium quality of service packets may be forwarded through spine switch 102, and the low quality of service packets may be forwarded through spine switch 104. Through the transmission path, the packet is transmitted according to the source virtual node of the physical server 121, namely the leaf switch 111, the corresponding spine switch, the leaf switch 113, and the destination virtual node of the destination physical server, and finally the transmission process of the data packet is completed.
In the virtual underlying network architecture, the communication equipment of the virtual underlying network has the capability of analyzing the tunnel inner layer message except for the 2-layer and 3-layer packet forwarding functions of the traditional network, different forwarding paths are adopted for the virtual node flows of different service applications in the underlying network tunnel by utilizing the analysis result, so that the service with different requirements on the network service quality is not influenced, the interference of the flow with low quality requirements on the flow with high quality requirements is avoided, the packet loss and the time delay of the high quality requirement service are avoided, the communication quality between the virtual nodes of different physical servers from end to end is ensured, and the network resources are reasonably distributed for the realization of the service application.
When the same leaf switch sends data packets to different destination physical servers through a plurality of transmission paths simultaneously, when the forwarding data rate of the output port of the leaf switch is low, the storage space of the port is occupied, or when the bandwidth of the output port of the leaf switch is greater than the bandwidth of a link, so that a bandwidth bottleneck is formed, or when the performance of a processor cannot meet the requirement of the link, the data packets may have a network congestion phenomenon in the transmission process, so that the transmission of service application flow is affected. Thus, as shown in fig. 2, in one embodiment, the virtual underlying network architecture further comprises: the system comprises an identifier adding unit, an identifier identification feedback unit and a flow regulating unit, so as to realize virtual flow congestion control and preferentially ensure normal transmission of high-class service application flow.
The identification adding unit is used for adding an identification capable of indicating the communication quality change condition to a packet which is sent by the source leaf switch and has the service quality lower than the set service quality when the communication quality of the source leaf switch changes.
Specifically, in the network architecture diagram shown in fig. 2, the leaf-spine topology network is the same as that in fig. 1, and is four spine switches 201,202,203,204 and four leaf switches 211,212,213,214, and there are four physical servers 221,222,223,224, where the spine switch 201 corresponds to high quality of service, the spine switch 202 corresponds to medium quality of service, the spine switch 204 corresponds to low quality of service, and the corresponding spine switch is selected by the website resolution unit and the path selection unit for data forwarding (as shown by the arrow in the figure). Only one of which is shown in fig. 2 for each physical server.
The change in communication quality includes a network transport congestion condition. For example, if the source leaf switch 211 is congested during the process of sending packets with different service qualities to the destination leaf switch 213 through the spine switch 201,202,204, the identifier adding unit determines that the condition belongs to a condition that the communication quality of the source leaf switch changes, and therefore adds a congestion alarm identifier to a part of packets sent by the source leaf switch 211, and indicates that the network congestion occurs in the source leaf switch 211 that sends the part of packets by using the congestion alarm identifier.
Since the high-quality service has the highest requirement on the network transmission quality, the data packet with the high-quality service needs to be sent preferentially when the network is congested, and therefore the data packet added with the congestion alarm flag is a non-high-quality service data packet. The specific basis for adding the identifier is the above-mentioned set service quality, if the set service quality is a high service quality, the congestion alarm identifier is added to the packets with medium and low service qualities (the source virtual node traffic with medium and low service qualities), and the identifier is not added to the packets with high service quality. Therefore, in addition to indicating network congestion by means of the congestion alarm indicator, the congestion alarm indicator is also used to indicate what level of quality of service packets needs to be restricted and adjusted in flow rate.
The identifier adding unit can be configured in part or all of the leaf switches, or can be independent of each leaf switch but in communication connection with each leaf switch, so that the addition of the congestion alarm identifier is directly performed when each leaf switch judges that the leaf switch is congested as a source leaf switch. The tag addition units in fig. 2 are individually configured in all leaf switches.
It is understood that the identification adding unit may be disposed in part or all of the spine switches, or may be independent of, but communicatively coupled to, the spine switches. When the spine switch judges that a certain source leaf switch is congested through the manner that the flow rate sent by the certain source leaf switch changes and the like, the identification adding unit of the spine switch can judge whether the service quality of the packet forwarded by the spine switch belongs to a service application lower than the set service quality, if so, the identification adding unit of the spine switch adds a congestion alarm identification to the packet before the spine switch forwards the packet, but does not add the identification to the packet not lower than the set service quality.
Thereafter, as shown in fig. 2, the packet 291 from the source leaf switch 211 with the congestion alarm id added thereto will reach the destination leaf switch 213 and then reach the corresponding destination virtual node of the destination physical server 223 under the forwarding of the spine switches 202 and 204.
The identifier identification feedback unit is used for identifying the identifier which can indicate the communication quality change condition and is contained in the packet after the destination virtual node receives the packet, and sending a flow regulation instruction corresponding to the communication quality change condition indicated by the identifier to the flow regulation unit.
The identifier recognition feedback unit and the traffic regulation unit may be configured in each physical server, for example, configured in each virtual node of the physical server. Specifically, the identifier identification feedback unit of the destination virtual node identifies congestion alarm identifiers contained in packets with medium and low service quality, and sends a flow regulation instruction for reducing medium and low service application flow to the flow regulation unit of the source virtual node of the source physical server.
And the flow regulating unit is used for receiving the flow regulating instruction and regulating the flow rate of the source virtual node corresponding to the service quality according to the flow regulating instruction.
Specifically, after receiving a traffic adjustment instruction for reducing the medium and low service application traffic from the destination virtual node, the traffic adjustment unit of the source virtual node decreases the service application traffic rate sent by the corresponding medium and low service quality source virtual node, thereby avoiding further network congestion and avoiding further occurrence of packet loss or excessive packet delay.
It should be noted that the identifier adding unit may be configured in the leaf switch and the spine switch at the same time, if the identifier adding unit of the source leaf switch 211 finds that the source leaf switch 211 is congested and adds the congestion alarm identifier according to the set service quality packet of the medium and low service quality packets, but the identifier adding unit of the spine switch 204 responsible for the low service quality finds that the spine switch 204 is also congested, the spine switch 204 may select to directly discard the low service quality virtual node traffic with the congestion alarm identifier, that is, not forward, so as to avoid further congestion of the spine switch 204. Specifically, the identifier adding unit of the spine switch 204 may control the spine switch 204 not to forward the packet to the destination leaf switch 213 when the spine switch 204 is congested and it is recognized that the packet has been added with the congestion alarm identifier.
It should be further noted that, the change of the communication quality of the source leaf switch also includes a situation that the network communication quality is improved, for example, the network congestion is improved or the congestion is solved, at this time, the identifier adding unit adds the rate recovery identifier to the corresponding packet, and the identifier identifying and feeding back unit sends the flow rate adjusting instruction for increasing the flow rate of the corresponding service quality virtual node to the corresponding flow rate adjusting unit after receiving the packet with the rate recovery identifier, so as to recover or even increase the flow rate of the corresponding service quality virtual node. The overall flow of increasing the flow rate is substantially the same as the overall flow of decreasing the flow rate when the network is congested, and refer to the description of the flow rate decreasing process.
By setting the congestion alarm mark, congestion control of different degrees is provided for different service application virtual node flows in an underlying network tunnel, when the underlying network congestion occurs, congestion control is firstly carried out on low-priority flows, the high-priority flows are ensured to pass through without obstruction, and the influence of network congestion on high-priority service application data is avoided to the maximum extent.
In the network data transmission process, if a certain physical server receives data sent by a plurality of other physical servers at the same time, or a certain physical server receives a large amount of data sent by one other physical server, the physical server receiving the data as a guided destination physical server may cause network congestion due to an excessive load, which affects the transmission of service application data. Therefore, in one embodiment, the address resolution unit is further configured to resolve whether the destination virtual node included in the packet is a virtual destination virtual node.
The website parsing unit parses the feature information of the message source virtual node in the tunnel inner layer to realize the traffic service slicing function, and also parses the type of the message destination virtual node in the tunnel inner layer. The types of the target virtual nodes comprise virtual target virtual nodes and real target virtual nodes, wherein the virtual target virtual nodes are 'front' IP addresses and are virtual IP addresses; the real destination virtual node is a 'post' IP address and is a real IP address.
And, the leaf switch further includes a node selection unit and a path modification unit.
In the network architecture diagram shown in fig. 3, the leaf-spine topology network is the same as that in fig. 2 (a part of the network transmission path is omitted), and is all four spine switches 301,302,303,304 and four leaf switches 311,312,313,314, and there are four physical servers 321,322,323,324, where the spine switch 301 corresponds to high quality of service, the spine switch 302 corresponds to medium quality of service, and the spine switch 304 corresponds to low quality of service. Each physical server includes a real virtual node as a destination virtual node of a source virtual node, and each real destination virtual node corresponding to the same virtual destination virtual node may be located in a plurality of different physical servers. The identification addition unit, the identification recognition feedback unit, and the flow rate adjustment unit are not shown in fig. 3.
The node selection unit is used for acquiring each real virtual node corresponding to the virtual target virtual node of the packet and selecting the real target virtual node from each real virtual node.
Each virtual destination virtual node is typically configured with one or more real destination virtual nodes having a corresponding relationship with the virtual destination virtual node, and the one or more real destination virtual nodes may be located in a plurality of different physical servers. If the target virtual node analyzed by the website analyzing unit is virtual, the real target virtual node corresponding to the virtual target virtual node needs to be determined by the node selecting unit.
If there is a need to adjust the load to balance the traffic, then in order to simultaneously handle multiple concurrent flows to the virtual destination virtual node, it is necessary to select one of the multiple real destination virtual nodes corresponding to the virtual destination virtual node as a real destination to which the traffic is directed for each concurrent flow. Meanwhile, different flows can enable the node selection unit to select different real destination virtual nodes from the plurality of real destination virtual nodes as real destinations to which the flows are guided, and the flows are guided to the plurality of different real destination virtual nodes.
If there is no need to adjust the load to balance it, then 1 virtual destination virtual node may only configure 1 real destination virtual node.
There are various ways to select the real destination virtual node, for example, if one virtual destination virtual node corresponds to N real destination virtual nodes, a hash algorithm or other ways may be used to calculate one of the N values of the virtual destination virtual node, and then the real destination virtual node corresponding to the calculated value is determined according to the predetermined corresponding relationship between the N values and the N real destination virtual nodes, so as to know the destination physical server to which the packet should be sent and the real destination virtual node thereof.
The path modifying unit is used for modifying the virtual destination virtual node of the tunnel packet inner layer message into the real destination virtual node selected by the node selecting unit, and placing the modified inner layer message into the tunnel between the source leaf switch and the physical server where the selected real destination virtual node is located, so that the path selecting unit determines the network transmission path of the selected real destination virtual node.
For the service traffic continuously sent from the source virtual node to the virtual destination virtual node, in order to achieve the effect of load balancing, the real destination virtual nodes corresponding to each traffic are not completely the same, so as to avoid that 1 real destination virtual node cannot process the service requests sent by multiple concurrent source virtual nodes.
Fig. 3 shows the transmission of two packets 391 and 392 that are adjacent to each other in the sending order in the high quality of service traffic under load balancing requirements. Before the packet 391 is sent to the spine switch 301 responsible for high service quality, the node selection unit of the leaf switch 311 acquires the virtual destination virtual node from the packet 391, and calculates, by using a hash algorithm, that the corresponding real destination virtual node is located in the physical server 322, and then the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 391 to change the virtual destination virtual node into the corresponding real destination virtual node located in the physical server 322, so that the packet 391 becomes the packet 391A. Leaf switch 311 then sends packet 391A through spine switch 301 to leaf switch 312, eventually reaching the real destination virtual node at physical server 322.
If the node selection unit of the leaf switch 311 uses the hash algorithm to calculate that the corresponding real destination virtual node is located in the physical server 323, the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 391 to make it become the corresponding real destination virtual node located in the physical server 323, and at this time, the packet 391 becomes the packet 391B. Leaf switch 311 then sends packet 391B through spine switch 301 to leaf switch 313, eventually reaching the real destination virtual node at physical server 323.
For the next packet 392 that needs to be sent after the source physical server 321 sends the packet 391, the node selection unit of the leaf switch 311 calculates, by using a hash algorithm, that the corresponding real destination virtual node is located in the physical server 324, and the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 392, so that the virtual destination virtual node becomes the corresponding real destination virtual node located in the physical server 324, and at this time, the packet 392 becomes the packet 392A. Leaf switch 311 then sends packet 392A through spine switch 301 to leaf switch 314, eventually reaching the real destination virtual node at physical server 324. Therefore, the shunting and load balancing of high-quality service flow are realized.
In addition, the process of implementing load balancing for medium and low grade traffic quality flows and the same principle of high grade traffic quality flows are not described herein again. By carrying out node selection and path modification, the data to be transmitted is sent to real target virtual nodes positioned on different physical servers, the situation that only one or too few real target virtual nodes cannot process service requests sent by a plurality of concurrent source virtual nodes at the same time is avoided, load balancing service is provided for virtual node flow in an underlying network tunnel, the occurrence of too heavy load of the virtual nodes is prevented, and optimization of service processing is realized.
To implement the virtual traffic performance measurement function, in one embodiment, some or all of the spine switches within the fabric include a first parameter adding unit for adding a first in-band network telemetry parameter of a packet into the packet before forwarding the packet to a leaf switch.
In the architecture shown in fig. 1, after the spine switch 104 responsible for forwarding low quality of service traffic receives the packet 491 sent by the source leaf switch 111, the first parameter adding unit of the spine switch 104 writes the in-band network telemetry parameter generated by the packet 491 in the transmission path from the source leaf switch 111 to the spine switch 104, that is, the first in-band network telemetry parameter, into the inner layer message of the tunnel packet, so as to form a packet 492. The in-band Network Telemetry parameters include in-band Network Telemetry MetaData (INT _ MD), and the INT _ MD may include switch information (e.g., a switch ID), message entry information (e.g., an ingress port number and time when a message enters), message sending information (e.g., an egress port number, a time spent in a device, a link utilization rate of an egress port), buffer information (e.g., queue occupancy information, a queue congestion status), and the like.
The leaf switch comprises a second parameter adding unit for adding a second in-band network telemetry parameter of the packet into the packet before forwarding the packet to the destination physical server.
Specifically, after the spine switch 104 forwards the packet 492 containing the first in-band network telemetry parameter to the destination leaf switch 113, the second parameter adding unit of the destination leaf switch 113 writes the in-band network telemetry parameter generated by the packet 492 in the transmission path from the spine switch 104 to the destination leaf switch 113, i.e. the second in-band network telemetry parameter, into the inner layer packet of the tunnel packet.
It is understood that for the source leaf switch 111 in fig. 1, the medium and high class quality of service traffic packets sent to the destination leaf switch 113 through the spine switch 102 and the spine switch 101 may be added to the corresponding INT _ MD as well.
When the virtual node flows with different grades of service quality reach different spine switches, the spine switches add the first INT _ MD of the virtual node flows with different grades of service quality into the inner layer message of the tunnel packet, and then forward the tunnel packet to the corresponding leaf switches. When the destination leaf switch receives the virtual node traffic of different grades of service quality from different spine switches, the destination leaf switch can also add the second INT _ MD of the virtual node traffic of different grades of service quality in the destination leaf switch into the first INT _ MD generated from the spine switch, thereby measuring the performance of the virtual node traffic of different grades of service quality in the original tunnel from end to end, namely from the source leaf switch to the destination leaf switch.
By setting the parameter adding unit, the accurate and fine performance measurement of the application virtual node flow of different services in the underlying network tunnel is realized, the mastering of the network transmission process condition is increased, the analysis and optimization of the running performance of different services of the cloud data center are conveniently subdivided, and the subsequent background analysis and strategy response aiming at optimization are facilitated.
Embodiments of a data transmission method based on a cloud data center virtual underlying network architecture disclosed in the present application are described in detail below with reference to fig. 1 to 4. The present embodiment is a method for implementing the foregoing cloud data center virtual underlay network architecture embodiment. The method utilizes a leaf-spine topology network formed by a plurality of leaf switches and a plurality of spine switches shown in fig. 1 for data transmission.
As shown in fig. 1 and 4, the method disclosed in this embodiment includes the following steps:
step 100, after receiving a packet sent by a leaf switch, analyzing the packet to obtain feature information of a source virtual node and/or a destination virtual node included in the packet.
Specifically, the leaf switch includes a website analysis unit, the source physical server 121 first sends the packet 191 to the leaf switch 111, and the leaf switch 111 first analyzes the packet 191 through the website analysis unit to obtain the feature information of the source virtual node included in the tunnel inner layer message, or obtain the feature information of the destination virtual node included in the tunnel inner layer message, or obtain the feature information of the source virtual node and the destination virtual node at the same time.
Step 200, determining the spine switch corresponding to the service quality reflected by the characteristic information, and further determining the network transmission path of the packet, so that the leaf switch sends the packet to the destination virtual node according to the network transmission path.
Specifically, the leaf switch includes a path selection unit, and after the website analysis unit obtains the feature information, the path selection unit selects a suitable spine switch from spine switches in the leaf spine topology network to forward, and then sends the packet to the selected spine switch. The selection criteria of the spine switch is to select the spine switch corresponding to the service quality of the virtual node (source virtual node, destination virtual node, or both).
By analyzing the tunnel inner layer message, different forwarding paths are adopted for the virtual node flows of different service applications in the underlying network tunnel by utilizing the analysis result, so that the services with different requirements on the network service quality are not influenced, the interference of the flows with low quality requirements on the flows with high quality requirements is avoided, the packet loss and the time delay of the services with high quality requirements are avoided, the communication quality between end to end among the virtual nodes of different physical servers is ensured, and the reasonable distribution of network resources for the service applications is realized.
In one embodiment, the characteristic information of the source virtual node includes a web address of the source virtual node and/or characteristic information of other virtual nodes except the source virtual node. The characteristic information of the destination virtual node comprises a website address of the destination virtual node and/or characteristic information of other virtual nodes except the destination virtual node.
In one embodiment, the method further comprises steps a1 to A3.
Step A1, when the communication quality of the source leaf switch changes, adding an identifier capable of indicating the change situation of the communication quality to the packet with the service quality lower than the set service quality sent by the source leaf switch.
Specifically, the virtual underlying network architecture further includes an identifier adding unit, as shown in fig. 2, when the source leaf switch 211 is congested during the process of sending packets with different service qualities to the destination leaf switch 213 through the spine switch 201,202,204, the identifier adding unit determines that the congestion occurs in the process of changing the communication quality of the source leaf switch, and therefore adds a congestion alarm identifier to a part of packets (e.g., packet 291) sent by the source leaf switch 211, and indicates that the network congestion occurs in the source leaf switch 211 that sends the part of packets by using the congestion alarm identifier.
Step a2, after the destination virtual node receives the packet, it identifies the identifier contained in the packet, and sends a traffic regulation instruction corresponding to the communication quality change condition indicated by the identifier to the traffic regulation unit.
Specifically, the virtual underlying network architecture further comprises an identification feedback unit, both the identification feedback unit and the flow regulation unit can be configured in each physical server, and the identification feedback unit identifies congestion alarm identifications contained in packets with medium and low service quality and sends flow regulation instructions for reducing medium and low service application flow to the flow regulation unit of the source virtual node of the source physical server.
And step A3, receiving the flow regulation instruction, and regulating the flow rate of the source virtual node corresponding to the service quality according to the flow regulation instruction.
Specifically, the virtual underlying network architecture further includes a flow adjusting unit, and after receiving a flow adjusting instruction for reducing medium and low service application flows from the destination virtual node, the flow adjusting unit of the source virtual node reduces the service application flow rate sent from the corresponding medium and low service quality source virtual node, thereby avoiding further network congestion and avoiding further occurrence of packet loss or excessive packet delay.
By setting the congestion alarm mark, congestion control of different degrees is provided for different service application virtual node flows in an underlying network tunnel, when the underlying network congestion occurs, congestion control is firstly carried out on low-priority flows, the high-priority flows are ensured to pass through without obstruction, and the influence of network congestion on high-priority service application data is avoided to the maximum extent.
In one embodiment, when parsing the packet, it is further parsed whether a destination virtual node included in the packet is a virtual destination virtual node. That is, the address resolution unit is further configured to resolve whether the destination virtual node included in the packet is a virtual destination virtual node.
And, the method further includes steps B1 and B2.
Step B1, when the destination virtual node is a virtual destination virtual node, obtaining each real destination virtual node corresponding to the virtual destination virtual node of the packet, and selecting a real destination virtual node from each real destination virtual node.
Specifically, if the target virtual node analyzed by the website analysis unit is virtual, the node selection unit needs to be used to determine a real target virtual node corresponding to the virtual target virtual node. In order to avoid the overload of the virtual node, one of the real destination virtual nodes corresponding to the virtual destination virtual node needs to be selected as a real destination for traffic guidance. The corresponding plurality of real destination virtual nodes are located within a plurality of different physical servers.
Step B2, the virtual destination virtual node of the packet is modified into the selected real destination virtual node, so that the path selection unit determines the network transmission path of the selected real destination virtual node.
Specifically, as shown in fig. 3, fig. 3 shows a transmission process of two packets 391 and 392 adjacent to each other in the sending order in the high quality of service traffic under the load balancing requirement. Before the packet 391 is sent to the spine switch 301 responsible for high service quality, the node selection unit of the leaf switch 311 acquires the virtual destination virtual node from the packet 391, and calculates, by using a hash algorithm, that the corresponding real destination virtual node is located in the physical server 322, and then the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 391 to change the virtual destination virtual node into the corresponding real destination virtual node located in the physical server 322, so that the packet 391 becomes the packet 391A. Leaf switch 311 then sends packet 391A through spine switch 301 to leaf switch 312, eventually reaching the real destination virtual node at physical server 322.
If the node selection unit of the leaf switch 311 uses the hash algorithm to calculate that the corresponding real destination virtual node is located in the physical server 323, the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 391 to make it become the corresponding real destination virtual node located in the physical server 323, and at this time, the packet 391 becomes the packet 391B. Leaf switch 311 then sends packet 391B through spine switch 301 to leaf switch 313, eventually reaching the real destination virtual node at physical server 323.
For the next packet 392 that needs to be sent after the source physical server 321 sends the packet 391, the node selection unit of the leaf switch 311 calculates, by using a hash algorithm, that the corresponding real destination virtual node is located in the physical server 324, and the path modification unit of the leaf switch 311 correspondingly modifies the virtual destination virtual node of the packet 392, so that the virtual destination virtual node becomes the corresponding real destination virtual node located in the physical server 324, and at this time, the packet 392 becomes the packet 392A. Leaf switch 311 then sends packet 392A through spine switch 301 to leaf switch 314, eventually reaching the real destination virtual node at physical server 324. Therefore, the shunting and load balancing of high-quality service flow are realized.
By carrying out node selection and path modification, the data to be transmitted is sent to real target virtual nodes positioned on different physical servers, the situation that only one or too few real target virtual nodes cannot process service requests sent by a plurality of concurrent source virtual nodes at the same time is avoided, load balancing service is provided for virtual node flow in an underlying network tunnel, the situation that the virtual nodes are overloaded is prevented, and optimization of service processing is realized.
In one embodiment, the method further comprises steps C1 and/or C2.
Step C1, a first in-band network telemetry parameter of the packet is added to the packet before forwarding the packet to the leaf switch.
In step C2, a second in-band network telemetry parameter of the packet is added to the packet before forwarding the packet to the destination physical server.
Specifically, as shown in fig. 1, the first parameter adding unit of the spine switch 104 writes the first in-band network telemetry parameter generated by the packet 491 in the transmission path from the source leaf switch 111 to the spine switch 104 into the inner layer message of the tunnel packet to form a packet 492. After spine switch 104 forwards packet 492 containing the first in-band network telemetry parameter to destination leaf switch 113, the second parameter adding unit of destination leaf switch 113 writes the second in-band network telemetry parameter generated by packet 492 in the transmission path from spine switch 104 to destination leaf switch 113 into the inner layer packet of the tunnel packet.
Through the steps C1 and/or C2, the method and the device realize the performance measurement of accurate fineness provided for the flow of the virtual nodes applied to different services in the underlying network tunnel, increase the grasp on the network transmission process condition, facilitate the analysis and optimization of the running performance of different services of the cloud data center, and facilitate the subsequent background analysis and strategy response aiming at optimization.
The division of the modules and units herein is only one division of logical functions, and other divisions may be possible in actual implementation, for example, a plurality of modules and/or units may be combined or integrated in another system. The modules and units described as separate parts may be physically separated or not. The components displayed as cells may or may not be physical cells, and may be located in a specific place or distributed in grid cells. Therefore, some or all of the units can be selected according to actual needs to implement the scheme of the embodiment.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (8)
1. A cloud data center virtual underlay network architecture, comprising: a leaf-spine topology network formed by a plurality of leaf switches and a plurality of spine switches; wherein the content of the first and second substances,
the leaf switch includes:
the network address analyzing unit is used for analyzing the packet after receiving the packet sent by the leaf switch to obtain the characteristic information of the source virtual node and/or the destination virtual node contained in the packet;
a path selection unit, configured to determine a spine switch corresponding to the service quality reflected by the feature information, and further determine a network transmission path of the packet, so that the leaf switch sends the packet to a destination virtual node according to the network transmission path; in addition, the architecture further comprises:
the system comprises an identifier adding unit, a service quality judging unit and a service quality judging unit, wherein the identifier adding unit is used for adding an identifier capable of indicating a communication quality change condition to a packet which is sent by a source leaf switch and has a service quality lower than a set service quality when the communication quality of the source leaf switch is changed;
the identification feedback unit is used for identifying the identification contained in the packet after the destination virtual node receives the packet and sending a flow regulation instruction corresponding to the communication quality change condition indicated by the identification to the flow regulation unit;
and the flow regulating unit is used for receiving the flow regulating instruction and regulating the flow rate of the source virtual node of the corresponding service quality according to the flow regulating instruction.
2. The architecture according to claim 1, characterized in that the characteristic information of the source virtual node comprises a web address of the source virtual node and/or characteristic information of other virtual nodes than the source virtual node; the characteristic information of the target virtual node comprises a website of the target virtual node and/or characteristic information of other virtual nodes except the target virtual node.
3. The architecture according to any of claims 1-2, wherein the address resolution unit is further configured to resolve whether a destination virtual node included in the packet is a virtual destination virtual node; and the number of the first and second electrodes,
the leaf switch further comprises:
the node selection unit is used for acquiring each real virtual node corresponding to the virtual target virtual node of the packet and selecting the real target virtual node from each real virtual node;
and the path modifying unit is used for modifying the virtual destination virtual node of the packet into the selected real destination virtual node so as to ensure that the path selecting unit determines the network transmission path of the selected real destination virtual node.
4. The architecture of any of claims 1-2, wherein the spine switch comprises a first parameter adding unit to add a first in-band network telemetry parameter of a packet into the packet prior to forwarding the packet to a leaf switch; and/or the presence of a gas in the gas,
the leaf switch comprises a second parameter adding unit, which is used for adding a second in-band network telemetry parameter of the packet into the packet before forwarding the packet to a destination physical server.
5. A data transmission method based on a cloud data center virtual bottom network architecture is characterized in that the method utilizes a leaf-ridge topology network formed by a plurality of leaf switches and a plurality of ridge switches to carry out data transmission, and the method comprises the following steps:
analyzing the packet after receiving the packet sent by the leaf switch to obtain the characteristic information of the source virtual node and/or the destination virtual node contained in the packet;
determining a spine switch corresponding to the service quality reflected by the characteristic information, and further determining a network transmission path of the packet, so that the leaf switch sends the packet to a destination virtual node according to the network transmission path; in addition, the method further comprises:
when the communication quality of a source leaf switch changes, adding an identifier capable of indicating the change condition of the communication quality to a packet which is sent by the source leaf switch and has the service quality lower than the set service quality;
after receiving a packet, a destination virtual node identifies the identifier contained in the packet and sends a flow regulation instruction corresponding to the communication quality change condition indicated by the identifier to a flow regulation unit;
and receiving the flow regulating instruction, and regulating the flow rate of the source virtual node of the corresponding service quality according to the flow regulating instruction.
6. The method according to claim 5, wherein the characteristic information of the source virtual node comprises a web address of the source virtual node and/or characteristic information of other virtual nodes except the source virtual node; the characteristic information of the target virtual node comprises a website of the target virtual node and/or characteristic information of other virtual nodes except the target virtual node.
7. The method according to any of claims 5-6, wherein in said parsing the packet, it is further parsed whether a destination virtual node included in the packet is a virtual destination virtual node; and the number of the first and second electrodes,
the method further comprises the following steps:
when the target virtual node is a virtual target virtual node, acquiring each real target virtual node corresponding to the packaged virtual target virtual node, and selecting one real target virtual node from each real target virtual node;
and modifying the virtual destination virtual node of the packet into the selected real destination virtual node so that the path selection unit determines the network transmission path of the selected real destination virtual node.
8. The method of any one of claims 5-6, further comprising:
adding a first in-band network telemetry parameter of a packet into the packet prior to forwarding the packet to a leaf switch; and/or the presence of a gas in the gas,
the second in-band network telemetry parameter of the packet is added to the packet before forwarding the packet to the destination physical server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911303921.4A CN111225031B (en) | 2019-12-17 | 2019-12-17 | Cloud data center virtual bottom layer network architecture and data transmission method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911303921.4A CN111225031B (en) | 2019-12-17 | 2019-12-17 | Cloud data center virtual bottom layer network architecture and data transmission method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111225031A CN111225031A (en) | 2020-06-02 |
| CN111225031B true CN111225031B (en) | 2020-12-18 |
Family
ID=70830160
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911303921.4A Active CN111225031B (en) | 2019-12-17 | 2019-12-17 | Cloud data center virtual bottom layer network architecture and data transmission method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111225031B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11444855B2 (en) * | 2020-07-07 | 2022-09-13 | Juniper Networks, Inc. | System and method for determining a data flow path in an overlay network |
| CN112003776B (en) * | 2020-08-12 | 2022-05-20 | 广东省新一代通信与网络创新研究院 | Message processing method and system |
| CN112787925B (en) * | 2020-10-12 | 2022-07-19 | 中兴通讯股份有限公司 | Congestion information collection method, optimal path determination method and network switch |
| CN112994987B (en) * | 2021-05-07 | 2021-09-14 | 阿里云计算有限公司 | Cloud network, measurement system, method, device and storage medium for cloud network |
| CN113630346B (en) * | 2021-09-14 | 2023-08-04 | 北京百度网讯科技有限公司 | Distributed network system |
| CN114978980B (en) * | 2022-04-08 | 2024-01-19 | 新奥特（北京）视频技术有限公司 | IP signal cross point scheduling device and method |
| CN115225589A (en) * | 2022-07-17 | 2022-10-21 | 奕德(广州)科技有限公司 | CrossPoint switching method based on virtual packet switching |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109547342A (en) * | 2018-11-30 | 2019-03-29 | 新华三技术有限公司 | A kind of delivery method and network node of policybased routing |
| CN109640348A (en) * | 2019-01-08 | 2019-04-16 | 中国联合网络通信集团有限公司 | The multi-service MEC network architecture, the processing method and processing device of multi-service data flow |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10277466B2 (en) * | 2016-11-10 | 2019-04-30 | Cisco Technology, Inc. | Network aware dynamic orchestration method and system for multicast services |
-
2019
- 2019-12-17 CN CN201911303921.4A patent/CN111225031B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109547342A (en) * | 2018-11-30 | 2019-03-29 | 新华三技术有限公司 | A kind of delivery method and network node of policybased routing |
| CN109640348A (en) * | 2019-01-08 | 2019-04-16 | 中国联合网络通信集团有限公司 | The multi-service MEC network architecture, the processing method and processing device of multi-service data flow |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111225031A (en) | 2020-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111225031B (en) | Cloud data center virtual bottom layer network architecture and data transmission method thereof | |
| CN107005439B (en) | Passive performance measurement for online service chaining | |
| EP4054126A1 (en) | Transmission quality test method and apparatus, system, and storage medium | |
| US7525919B2 (en) | Packet communication method with increased traffic engineering efficiency | |
| CN111771359B (en) | Method and system for connecting communication networks | |
| US20140207968A1 (en) | Server Load Balancer Traffic Steering | |
| US10153964B2 (en) | Network routing using dynamic virtual paths in an overlay network | |
| CN107864061A (en) | A kind of method of virtual machine port speed constraint and mirror image in private clound | |
| CN111953604A (en) | Method and device for providing service for service flow | |
| US20220278904A1 (en) | Method, Apparatus, and System for Sending Packet and Receiving Packet to Perform OAM | |
| JP2012075173A (en) | Systems and methods for dynamic mode-driven link management | |
| WO2021082575A1 (en) | Packet forwarding method, apparatus, storage medium, and system | |
| US10009282B2 (en) | Self-protecting computer network router with queue resource manager | |
| US7027410B2 (en) | Method and apparatus for maintaining consistent per-hop forwarding behavior in a network using network-wide per-hop behavior definitions | |
| US10623278B2 (en) | Reactive mechanism for in-situ operation, administration, and maintenance traffic | |
| Rezende et al. | An SDN-based framework for routing multi-streams transport traffic over multipath networks | |
| CA2354360A1 (en) | Policy enforcing switch | |
| CN108965121A (en) | Transmit method, host and the interchanger of data | |
| CN115412482B (en) | Calculation force routing method and device, electronic equipment and storage medium | |
| US20200304399A1 (en) | Method and system for interfacing communication networks | |
| CN113872824B (en) | CDN network quality detection method, system, server and storage medium | |
| Gonzalez-Diaz et al. | Stateless flow-zone switching using software-defined addressing | |
| CN112671662B (en) | Data stream acceleration method, electronic device and storage medium | |
| CN114157595A (en) | Communication system, data processing method and related equipment | |
| Kushwaha et al. | IPv6 flow-label based application aware routing in SDNs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |