CN116436730A - Virtual network mapping method and mapping system based on cloud computing service - Google Patents
Virtual network mapping method and mapping system based on cloud computing service Download PDFInfo
- Publication number
- CN116436730A CN116436730A CN202310694407.8A CN202310694407A CN116436730A CN 116436730 A CN116436730 A CN 116436730A CN 202310694407 A CN202310694407 A CN 202310694407A CN 116436730 A CN116436730 A CN 116436730A
- Authority
- CN
- China
- Prior art keywords
- virtual
- network
- snowflake
- virtual network
- level
- 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.)
- Granted
Links
- 238000013507 mapping Methods 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 43
- 241000533950 Leucojum Species 0.000 claims abstract description 170
- 230000008569 process Effects 0.000 claims description 13
- 238000010276 construction Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 101100134058 Caenorhabditis elegans nth-1 gene Proteins 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 6
- 230000002093 peripheral effect Effects 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 abstract description 4
- 239000010410 layer Substances 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 235000008694 Humulus lupulus Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 239000012792 core layer Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- 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
- H04L41/12—Discovery or management of network topologies
Abstract
The invention relates to a virtual network mapping method and a mapping system of cloud computing service, which aim at the personalized demand of a virtual network of a user terminal, an improved network topology structure is built for a bottom physical network, and a novel snowflake network structure with high expansibility, high operation efficiency and low failure rate is provided according to a Koch curve, so that the network system architecture can execute network resource expansion at any time according to the demand, and a routing mechanism between nodes can be realized in a shorter average path on the premise of ensuring a lower number proportion of a switch and a server, thereby improving the operation efficiency, reducing the risk of resource failure caused by link failure through the built multiple parallel paths, and providing a stable operation environment for building the personalized virtual network; and further, the resource attributes are used for similarity comparison, and snowflake structures with highest similarity are selected to be mapped onto the virtual network, so that the request acceptance rate of the virtual network is remarkably improved.
Description
Technical Field
The invention relates to the technical field of cloud computing services, in particular to a virtual network mapping method and a virtual network mapping system based on cloud computing services.
Background
Under the influence of the current internet architecture, cloud computing faces many challenges in terms of service diversity provision and the like, and network virtualization technology is an effective technical means for solving the problem. The network virtualization technology establishes a plurality of mutually independent virtual networks on the same underlying physical network, so that specific types of services and novel network technology deployment, such as IPTV, voIP and the like, are rapidly developed at low cost, special bandwidth and flow control protocols are adopted, and the requirements of clients on QoS sensitive services such as video, voice call and the like are met. Generating a virtual network request at a user end according to application requirements, and providing reasonable bottom network resource allocation for a virtual network is called virtual network mapping.
Most of the existing virtual network mapping algorithms focus on several aspects in a fault-free network environment: the method maximizes the receiving rate of the mapping request, the total income and the utilization rate of the underlying resource, minimizes the scheduling delay of the request, the total energy consumption of the virtual network and the like. The mapping method is based on the enhanced graph, such as the VineYard algorithm proposed by Chordhury et al, the RW-BFS algorithm proposed by Cheng et al and based on the ranking of topology-aware nodes, and the mapping algorithm proposed by Houidi et al and based on distributed cooperation, but the personalized service problem of the mapped virtual network is not considered by the mapping methods.
With the continuous emergence of new technologies and products, users' demands on the internet are also more personalized and diversified. Such as network services, network topology, real-time security, etc., specifically tailored to the application scenario. Therefore, a set of flexible and variable virtual network mapping mechanism with adjustable optimization targets is adopted, so that the mapped virtual network has different structural forms and performance characteristics, and the diversified requirements of users can be met. The core problem to be solved by the virtual network mapping is that: how to find a corresponding relation between the virtual network and the physical network, and map the nodes and links in the virtual network into the physical network according to the user demands, thereby providing satisfactory service for the user while guaranteeing the benefits of the infrastructure provider.
The network virtualization technology can realize a plurality of completely different virtual networks on one or more physical networks through mechanisms such as abstraction, distribution, isolation and the like, and the network topology on the logic is defined by software to meet the personalized requirements of different services on network resources, so that the flexible configuration and dynamic management of the network resources are realized. However, the architecture, networking mode and resource allocation of the existing information communication network have various drawbacks such as "static" and "stiff", and it has been difficult to meet the service requirements of random dynamics, complexity, variability, subjective performance and individuation of the virtualization technology. Therefore, how to provide personalized virtual network services with diversification is a problem to be solved by infrastructure and mapping function providers.
Disclosure of Invention
The invention aims to solve the problem that the conventional virtual network mapping technology lacks the capability of providing personalized services for users, and provides a virtual network mapping method and a virtual network mapping system based on cloud computing services. By performing improved layout on the traditional topological network structure, a modularized network group is formed, the requested personalized virtual network is matched with the corresponding module, and the bottom network resource which is suitable for the application environment is acquired.
In order to achieve the above object, the present invention provides a virtual network mapping method based on cloud computing service, which includes the following stages:
and (3) a network construction stage: establishing a snowflake type network topology structure for a bottom physical network in a cloud computing environment, wherein the network topology structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and the nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
virtual network request phase: the user terminal generates a virtual network request, wherein the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
virtual network mapping stage: and selecting a snowflake structure level consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each peer snowflake structure and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity onto the virtual network.
Further preferably, the nth stage snowflake structure generating process includes: and (3) disconnecting all virtual connections and real connections on the n-1-level snowflake structure, adding a 0-level snowflake structure at each breakpoint, and reestablishing two real connections at two ends of the 0-level snowflake structure and the original connection relationship to form the n-level snowflake structure, wherein the virtual connections are derived from the 0-level snowflake structure added in the n-1-level snowflake structure forming stage.
Further preferably, the 0 th-level snowflake structure comprises a switch and a plurality of cloud servers in a central position, all cloud servers are scattered on the periphery of the switch and are connected with the switch in an internal mode, and virtual connection is established between any two adjacent cloud servers.
Further preferably, the 0 th stage snowflake structure comprises a central switch arranged at a central position and peripheral switches distributed at the periphery of the central switch and establishing internal connection with the central switch, and virtual connection is established between any two adjacent peripheral switches.
Further preferably, the resource attribute related to the resource constraint condition of the virtual node includes: the node number, the node computing capacity, the node storage capacity, the node transmission capacity and the node distance; the resource attribute related to the resource constraint condition of the virtual link comprises: the number of links, the link bandwidth and the link delay.
Further preferably, the selection rule of the snowflake structure level is:
step 1-1) counting the number of virtual nodes in a virtual network, and selecting an mth-level snowflake structure with the number of servers being greater than and closest to the number of the virtual nodes;
step 1-2) counting the number of virtual links in the virtual network, comparing the number of real connections in the mth stage snowflake structure with the number of virtual links, if the number of real connections is not less than the number of virtual links, taking the mth stage as the final selected stage, otherwise, executing the step 1-3);
and 1-3) reselecting the m+1st stage snowflake structure, and continuing to execute the step 1-2) until the number of the real connections is not smaller than the level corresponding to the number of the virtual links.
Further preferably, the similarity calculation process of the peer snowflake structure and the virtual network about the resource attribute includes:
step 2-1), calculating the average value of all nodes and links in the current snowflake structure on each resource attribute, and collecting the average value of all the resource attributes to generate a node cluster center and a link cluster center;
step 2-2) calculating the similarity between each virtual node and the node clustering center and the similarity between each virtual link and the link clustering center respectively by using a cosine similarity algorithm;
step 2-3) setting up weight coefficients for the nodes and the links, carrying out weighted summation on the similarity of all the virtual nodes and the virtual links, and further taking the homogenized result as a similarity value of the virtual network and the current snowflake structure.
Further preferably, the method further comprises a constraint process of node distance:
defining a Boolean variable, setting a distance threshold for each virtual node, and when the distance between the virtual node and the node clustering center is not greater than the distance threshold, enabling the Boolean variable to be equal to 1, otherwise enabling the Boolean variable to be equal to 0; and taking the Boolean variable as a constraint parameter to participate in similarity product operation of the corresponding virtual node.
The invention also provides a virtual network mapping system based on the cloud computing service, which specifically comprises: the network construction unit and the virtual network mapping unit are arranged on the cloud computing data center, and the virtual network request unit is arranged on the user terminal;
a network construction unit: the method comprises the steps that a snowflake type network topological structure is established for a cloud computing bottom physical network, wherein the network topological structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and an nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
virtual network request unit: generating a virtual network request according to the application requirement of the user terminal, wherein the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
virtual network mapping unit: and selecting a snowflake structure level consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each peer snowflake structure and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity onto the virtual network.
The virtual network mapping method and the virtual network mapping system provided by the invention have the beneficial effects that:
aiming at the personalized demands of the virtual network of the user terminal, the invention establishes an improved network topology structure for the bottom physical network, and according to the Koch curve, a novel snowflake network structure with high expansibility, high operation efficiency and low failure rate is provided, so that the network system structure can execute network resource expansion at any time according to the demands, and can realize a routing mechanism among nodes in a shorter average path on the premise of ensuring a lower quantity ratio of a switch to a server, thereby improving the operation efficiency, and reducing the risk of resource failure caused by link failure through the established multiple parallel paths, thereby providing a stable operation environment for the establishment of the personalized virtual network.
The cloud computing network is provided with a plurality of modularized snowflake structures with different scales by adopting a modularized network structure established by recursively defined snowflake topological network, and each modularized snowflake structure has relatively independent data operation, storage and transmission capacity, so that the snowflake structure consistent with the requested virtual network scale can be arbitrarily selected in the network, the snowflake structure is taken as a mapping object of the virtual network, and is adapted to the required resource scale, unnecessary resource waste or excessive resource load is reduced, and a multi-port server is taken as a forwarding center, so that the bisection bandwidth of the network is greatly improved, and the deployment requirement of more and more application services of the cloud computing data center network is met.
The constraint on numerous resource attributes is established for the virtual nodes and the virtual links, similarity comparison is further carried out by utilizing the resource attributes, the comprehensive similarity degree between each snowflake structure to be selected and the virtual network of the user is obtained, then the snowflake structure with the highest similarity is selected from the snowflake structures to be mapped onto the virtual network, namely, the preference difference of different virtual nodes or virtual links on the resource performance is considered, and the mapping object closest to the virtual network is selected from the aspect of the overall resource performance, so that the request acceptance rate of the virtual network is obviously improved, the more virtual networks which are successfully mapped are obtained, and the average benefit of the bottom layer network is also increased.
Drawings
FIG. 1 is a flow chart of a virtual network mapping method based on cloud computing services provided by the invention;
FIG. 2 is a schematic view of a 0-level snowflake structure according to an embodiment of the present invention;
FIG. 3 is a schematic view of a level 1 snowflake structure according to an embodiment of the present invention;
FIG. 4 is a schematic view of a 2-stage snowflake structure according to an embodiment of the present invention;
FIG. 5 is a flow chart of snowflake structure level selection provided in one embodiment of the invention;
fig. 6 is an interaction schematic diagram of a virtual network mapping system based on cloud computing service provided by the invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings.
The data center is an underlying facility for carrying various information applications and services, and has the functions of calculation, storage and network. The data center network depends on a large number of high-speed switches, routers and physical circuit connection servers to realize the requirements of cooperative computation of equipment in the data center and internal and external data interaction, and the network topology structure determines the selection standard of the data center network equipment and the interconnection mode of the data center network equipment, so that the network structure is an important factor which must be considered in designing the cloud computing data center, and provides sufficient guarantee for high expansibility of the cloud computing data center and high utilization rate of resources. The traditional data center network carries mainly the application service of the C/S mode, generally adopts a tree structure comprising a core layer, a convergence layer and an access layer, and researches on the topological structure of the traditional data center with a three-layer structure find that the traditional topological structure and the data center facing the novel computing mode such as cloud computing have various defects, including small network scale, poor expansibility, limited available bandwidth, dispersed resources, low utilization rate, easiness in single-point failure and other problems, which are difficult to solve.
Therefore, the invention applies the novel snowflake topology structure to improve the cloud computing bottom physical network so as to enhance the network interconnection among the data center servers and improve the collaborative computing capability among the data center servers. And meanwhile, selecting a snowflake structure of a certain level, which is consistent with the scale of the virtual network defined by the user, from the network area, and selecting the snowflake structure which is closest to the virtual network in the same level as a mapping object by combining a similarity algorithm. As shown in fig. 1, the method for mapping a virtual network based on cloud computing service provided by the present invention specifically includes the following implementation stages:
and (3) a network construction stage: establishing a snowflake type network topology structure for a bottom physical network in a cloud computing environment, wherein the network topology structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and the nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
virtual network request phase: the user terminal generates a virtual network request, wherein the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
virtual network mapping stage: and selecting a snowflake structure level consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each peer snowflake structure and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity onto the virtual network.
The snowflake network structure adopts a recursion definition mode, and the n-level snowflake structure is formed by adding a plurality of 0-level snowflake structures on the n-1-level snowflake structure. Each 0-level snowflake structure consists of a switch at a central position and k servers, wherein k is generally 3-8, all cloud servers are scattered on the periphery of the switch and establish internal connection with the switch, and the 0-level snowflake structure is shown in fig. 2, wherein 3 servers are arranged. 3 servers are connected to 1 multiport switch. And adjusting theoretical positions of 3 servers, and adding virtual connections to any two adjacent servers to form 3 virtual connections. These 3 virtual connections are not physical links that actually exist, but are used to facilitate the construction process of the next-level snowflake structure.
Then, a 1-level snowflake structure is constructed, as shown in fig. 3, 3 virtual connections are disconnected on the basis of the original 0-level snowflake structure, and each time one virtual connection is disconnected, namely, a 0-level snowflake structure is added at the position of the virtual connection, a switch in the newly added 0-level snowflake structure is connected with two ends (two servers) of the original virtual connection, and relative to the virtual connection, the reestablished connection is called as a real connection, and the 1-level snowflake structure is formed after 3 0-level snowflake structures are added in an accumulated mode for a link with an actual connection relation. It should be noted that the real connection is not only a connection actually existing, but also a process of establishing a virtual connection to a real connection, and the connection has no state change from scratch, so that the connection link between the server and the switch in the 0-level snowflake structure is not defined as a real connection, and the connection link is not a state change although the connection link is an actually existing connection. In addition, unlike existing snowflake structures, it is: the invention does not do virtual connection default treatment for the newly added 0-level snowflake structure, and always keeps the same structure with the original 0-level snowflake structure at the bottom layer, so that each redefined level snowflake structure has a more standard modularized format, and the quantity of the n-1 th to 0-th level snowflake structures, the total quantity of switches and servers in the n-th level snowflake structure can be determined as long as the k value and the selected level n are known, thereby being convenient for realizing the modularized mapping process of the virtual network.
Further build 2-level snowflake structure, as shown in fig. 4, firstly, disconnect all real connections on the basis of original 1-level snowflake structure, add a 0-level snowflake structure between two ends of original real connections (i.e. 0-level snowflake structure and server shown in fig. 3), connect the switch in newly added 0-level snowflake structure with two ends of original virtual connections (switch and server of 0-level snowflake structure), then disconnect all virtual connections on 3 existing 0-level snowflake structures, continue to add 0-level snowflake structure, re-build real connections, and upgrade 3 0-level snowflake structures into 1-level snowflake structure.
And by analogy, on the n-1 level snowflake structure, all virtual connections and real connections are disconnected, a 0 level snowflake structure is added at each breakpoint, and then two real connections are reestablished at the two ends of the 0 level snowflake structure and the two ends of the original connection relationship to form the n level snowflake structure, so that the virtual connections are all derived from the 0 level snowflake structure added in the n-1 level snowflake structure forming stage.
The network structure is suitable for the personalized virtual network requirement in the big data environment, firstly, the structure is modularly processed in a mode of adding the fixed structure as far as possible, the modular connection of the structure is facilitated, the modular structure with a corresponding scale can be selected according to the requirement and mapped onto the virtual network, the servers in the module are connected with the switch and the server in the domain, and the data transmission delay is negligible; secondly, when the n-level snowflake structure is not fully expanded, the n+1-level snowflake structure is easier to continue to expand under the condition of not damaging the original link, and as can be seen from the change from 1 level to 2 level in the figure, the 0 level structure in the figure 3 is expanded and then is upgraded to 1 level, and the number of 0 level structures in the figure 4 is doubled, and the n-level snowflake structure comprises 3 1 level units and 6 0 level units, so that the n-level snowflake structure inevitably exists
An m-th stage snowflake structure capable of accommodating
0 th to n-1 th snowflake structures, supposing +.>
The constructed 10 th-level snowflake structure totally comprises 5110 0-9-level snowflake structure modules, the number of the modules can exponentially increase along with the increase of n values, the requirement of a cloud computing data center on a large-scale network is met, and a large selection space is provided for various heterogeneous virtual networks; again, by reliability verification, the shortest path between any 2 servers of the n-level snowflake structure will not exceed 2n+1 hops, and there are at least 2, at most +.>
The parallel paths ensure that the network has smaller network diameter and larger network bisection bandwidth among servers; in addition, for any level of module structure, the number ratio of the switch to the server is always kept
，/>
The larger the value is, the smaller the duty ratio of the switch is, so that the energy consumption can be effectively reduced, and meanwhile, the cost overhead of the switch can be reduced.
In another embodiment provided by the invention, the server in the 0 th-level snowflake structure is replaced by a switch, namely, the switch comprises a central switch arranged at a central position and peripheral switches distributed on the periphery of the central switch and connected with the central switch in an internal mode, and similarly, virtual connection is established between any two adjacent peripheral switches. Considering that when the snowflake level is continuously enlarged and the number of servers in the structure is continuously increased to reach the million levels, the number of the routed messages is greatly increased, and at the moment, in the first established bottom snowflake structure, the load of the servers is serious, therefore, the bottom servers are replaced by switches, for example, in the first established 2-level snowflake structure, switches with ten megaEthernet ports are arranged at 57 server positions in fig. 4, so that the information forwarding load intensity of the low-level snowflake structure is relieved, the message forwarding speed is improved, meanwhile, the servers of the upper layout are reserved, and the service quality of users is not affected by the conversion of a limited number of switches.
For selection of snowflake structure level, referring to FIG. 5, the following examples are now provided to specifically illustrate this selection step:
step 1-1) counting the number of virtual nodes in a virtual network, and selecting an mth-level snowflake structure with the number of servers being greater than and closest to the number of the virtual nodes; assuming the existence of a virtual network
Can be expressed as a weighted undirected graph +.>
Wherein->
Virtual representing virtual network requestsNode set, ->
A set of virtual links representing virtual network requests, +.>
Attribute constraint set representing virtual node, +.>
The set of property constraints representing the virtual link, need to be +.>
Finding out the m-th stage snowflake structure with the same node number in the whole network, wherein the total number of the virtual nodes is marked as +.>
The m value satisfies that the number of servers in the m-1 level snowflake structure is less than +.>
And the number of servers in the m-th level snowflake structure is greater than or equal to +.>
；
Step 1-2) counting the number of virtual links in the virtual network as
Comparing the number of real connections in the mth level snowflake structure with the number of virtual links, if the number of real connections is not less than the number of virtual links +.>
Taking the mth level as the final selected level, otherwise, indicating that the number of real connections in the mth level structure is insufficient, if the direct allocation can influence the virtual network operation efficiency, executing the step 1-3);
step 1-3) reselecting the m+1st stage snowflake structure, and continuing to execute step 1-2), namely comparing the number of real connections in the m+1st stage snowflake structure with the number of real connections in the m+1st stage snowflake structure
If the number of real connections is still smaller than +.>
And further selecting the m+2-level snowflake structure upwards until the level corresponding to the number of the virtual links is selected, wherein the number of the real connections is not smaller than the number of the virtual links.
It should be noted that in the above embodiment, the real connection is used to participate in the number comparison, and the real connection does not include the connection link between the server and the switch in the 0-level snowflake structure, if the part of the link is also participated in counting, especially as k increases, the ratio of the number of the part of the link is also exponentially increased, for the selected m-level snowflake structure, the ratio of the number of links between the internal modules and between the server is naturally reduced, and then the ratio of the links between the server and the switch is increased, which necessarily affects the data processing and transfer efficiency between the modules and between the intra-domain servers, so that the number of links between the server and the switch in the 0-level snowflake structure is not counted as an effective link participation, and in the finally selected m-level snowflake structure, the number of links actually existing is significantly greater than the total number of virtual links requested by the virtual network, thereby providing the user with an extensible network bandwidth requirement.
After the class confirmation of the snowflake structures with the same scale is completed, according to the resource attribute constraint condition of the virtual network, one snowflake structure with the highest similarity is selected from the snowflake structure set with the same class to be used as the virtual network mapping object. Wherein, the resource attribute related to the resource constraint condition of the virtual node may include: the node number, the node computing capacity, the node storage capacity, the node transmission capacity and the node distance; the resource attributes involved in the resource constraint of the virtual link may include: the number of links, the link bandwidth and the link delay.
In one embodiment of the present invention, the similarity calculation process of the peer snowflake structure and the virtual network about the resource attribute includes the following steps:
step 2-1), calculating the average value of all nodes and links in the current snowflake structure on each resource attribute, and collecting the average value of all the resource attributes to generate a node cluster center and a link cluster center;
step 2-2) calculating the similarity between each virtual node and the node clustering center and the similarity between each virtual link and the link clustering center respectively by using a cosine similarity algorithm;
step 2-3) setting up weight coefficients for the nodes and the links, carrying out weighted summation on the similarity of all the virtual nodes and the virtual links, and further taking the homogenized result as a similarity value of the virtual network and the current snowflake structure.
The similarity calculation process is further described below in conjunction with specific resource attributes:
assume that an m-level snowflake structure possesses a setWherein->
M-level snowflake structure->
A weighted undirected graph can be used +.>
Indicating (I)>
，/>
Representing node set, ++>
Representing a link set,/->
Attribute constraint set representing node +.>
Representing a set of attribute constraints for the link; attribute constraint set->
Comprising the attribute sequence of each node +.>
，/>
Representing node computing power,/->
Representing node transmission capabilities->
Representing node storage capacity; attribute constraint set->
Comprising a sequence of attributes for each link>
，/>
Representing link bandwidth, +.>
Representing link delay; calculating snowflake Structure>
Cluster centers of all nodes in (a):
wherein the method comprises the steps of
Corner mark->
Or->
Or->
，/>
Representing node set +.>
Is>
Indicate->
Personal node about->
The attribute values can be calculated to obtain the clustering centers of the nodes about 3 attributes, and the average value is calculated
And->
And->
Combining to obtain a node clustering center;
calculating snowflake structure
Cluster centers of all links in (a):
wherein the method comprises the steps of
Corner mark->
Or->
，/>
Representing Link set->
Is>
Indicate->
Individual links about->
Attribute values, the calculation can obtain the clustering center of 2 attributes of the link, and the average value +.>
And
combining to obtain a link clustering center;
extracting virtual network datasets
Likewise, the attribute constraint set +.>
Comprising a custom attribute sequence for each virtual node->
Attribute constraint set->
Comprising a custom property sequence for each virtual link>
，/>
、/>
、/>
Representing computing, transmitting and storing capabilities, respectively, of the virtual node request, +.>
、/>
Respectively representing the bandwidth and delay of the virtual link request;
computing each virtual node
Similarity to node cluster center:
computing each virtual link
Similarity to link cluster center:
because the requirements of different user terminals on the requested virtual network performance are different, for example, for network layer applications with high calculation and storage concentration requirements, the requirements on the performance of the nodes are high, and for communication intensive applications, the requirements on the bandwidth of the links are high, so that the similarity is weighted and summed by setting up weight coefficients for the nodes and the links so as to adapt to different virtual network application environments;
further, calculating virtual network by weighting algorithm and mean method
And snow flake structure->
Is a comprehensive similarity of:
wherein the method comprises the steps of
Representing virtual node->
Weight coefficient of>
Representing virtual Link->
Weight coefficient of (2), and satisfy
The ownership coefficient is set by the user terminal during the virtual network request phase, and is +.>
Representing the total number of virtual nodes>
Representing the total number of virtual links; finally, a similarity set can be obtained>
By selecting the maximum similarity value from the collection after descending order +.>
And the corresponding snowflake network structure and the virtual network establish a mapping relation.
Some network applications require strict geographical location requirements, such as high-density operation process requiring instant messaging and time requirement sensitivity, physical distance exceeding a certain range can cause returned processing data to fail, therefore, the network applications need to take the inter-node distance change as another constraint condition to participate in selection of a modularized network structureThe process, in particular, by defining a boolean variable b, for each virtual node
Setting a distance threshold +.>
When the distance between the virtual node and the node cluster center is not more than the distance threshold value +.>
When the Boolean variable is equal to 1, otherwise, the Boolean variable is equal to 0; taking the Boolean variable as constraint parameter to participate in similarity product operation of the corresponding virtual node, and transforming the similarity calculation formula to obtain the following steps:
wherein the method comprises the steps of
Representing virtual node->
The variable assignment depends on the physical distance between two nodes, by implementing constraint on distance attribute on the nodes, if all virtual nodes are located at positions (i.e. client terminal positions) generally far away from the node clustering center of a snowflake structure, the calculated comprehensive similarity value is calculated>
Will be smaller, thereby excluding the network resources at the far end from the selectable range, on the contrary, the snowflake structure with closer distance has a similarity value +.>
The larger the more the selection criteria for the mapping object are met.
In order to implement the virtual network mapping method, the present invention also provides a virtual network mapping system based on cloud computing service, as shown in fig. 6, the system includes: the network construction unit and the virtual network mapping unit are arranged on the cloud computing data center, and the virtual network request unit is arranged on the user terminal;
the network construction unit is used for establishing a snowflake type network topological structure for a cloud computing bottom physical network, wherein the network topological structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and the nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
a virtual network request unit for generating a virtual network request according to the application requirement of the user terminal, wherein the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
and the virtual network mapping unit is used for selecting snowflake structure levels consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each snowflake structure with the same level and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity onto the virtual network.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Claims (9)
1. The virtual network mapping method based on the cloud computing service is characterized by comprising the following steps of:
the method comprises a network construction stage, wherein a snowflake type network topological structure is established for a bottom physical network in a cloud computing environment, the network topological structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and an nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
in the virtual network request stage, the user terminal generates a virtual network request, and the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
and in the virtual network mapping stage, selecting a snowflake structure level consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each snowflake structure with the same level and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity to the virtual network.
2. The cloud computing service-based virtual network mapping method according to claim 1, wherein the nth stage snowflake structure generating process comprises: and (3) disconnecting all virtual connections and real connections on the n-1-level snowflake structure, adding a 0-level snowflake structure at each breakpoint, and reestablishing two real connections at two ends of the 0-level snowflake structure and the original connection relationship to form the n-level snowflake structure, wherein the virtual connections are derived from the 0-level snowflake structure added in the n-1-level snowflake structure forming stage.
3. The virtual network mapping method based on cloud computing service according to claim 2, wherein the 0 th level snowflake structure comprises a switch and a plurality of cloud servers in a central position, all cloud servers are scattered on the periphery of the switch and establish internal connection with the switch, and virtual connection is established between any two adjacent cloud servers.
4. The cloud computing service-based virtual network mapping method of claim 2, wherein the level 0 snowflake structure comprises a central switch disposed at a central location, and peripheral switches distributed around the central switch and establishing an internal connection with the central switch, and a virtual connection is established between any two adjacent peripheral switches.
5. The cloud computing service-based virtual network mapping method according to claim 1, wherein the resource attributes involved in the resource constraint condition of the virtual nodes include the number of nodes, the computing capacity of the nodes, the storage capacity of the nodes, the transmission capacity of the nodes and the distance of the nodes; the resource attribute related to the resource constraint condition of the virtual link comprises: the number of links, the link bandwidth and the link delay.
6. The cloud computing service-based virtual network mapping method according to claim 1, wherein the snowflake structure level selection rule is:
step 1-1) counting the number of virtual nodes in a virtual network, and selecting an mth-level snowflake structure with the number of servers being greater than and closest to the number of the virtual nodes;
step 1-2) counting the number of virtual links in the virtual network, comparing the number of real connections in the mth stage snowflake structure with the number of virtual links, if the number of real connections is not less than the number of virtual links, taking the mth stage as the final selected stage, otherwise, executing the step 1-3);
and 1-3) reselecting the m+1st stage snowflake structure, and continuing to execute the step 1-2) until the number of the real connections is not smaller than the level corresponding to the number of the virtual links.
7. A virtual network mapping method based on cloud computing services as recited in claim 1, wherein the similarity calculation process of the peer snowflake structure and the virtual network with respect to resource attributes includes:
step 2-1), calculating the average value of all nodes and links in the current snowflake structure on each resource attribute, and collecting the average value of all the resource attributes to generate a node cluster center and a link cluster center;
step 2-2) calculating the similarity between each virtual node and the node clustering center and the similarity between each virtual link and the link clustering center respectively by using a cosine similarity algorithm;
step 2-3) setting up weight coefficients for the nodes and the links, carrying out weighted summation on the similarity of all the virtual nodes and the virtual links, and further taking the homogenized result as a similarity value of the virtual network and the current snowflake structure.
8. The cloud computing service-based virtual network mapping method of claim 7, further comprising a constraint process for node distance:
defining a Boolean variable, setting a distance threshold for each virtual node, and when the distance between the virtual node and the node clustering center is not greater than the distance threshold, enabling the Boolean variable to be equal to 1, otherwise enabling the Boolean variable to be equal to 0; and taking the Boolean variable as a constraint parameter to participate in similarity product operation of the corresponding virtual node.
9. A virtual network mapping system based on cloud computing services, the system comprising: the network construction unit and the virtual network mapping unit are arranged on the cloud computing data center, and the virtual network request unit is arranged on the user terminal;
a network construction unit: the method comprises the steps that a snowflake type network topological structure is established for a cloud computing bottom physical network, wherein the network topological structure comprises a multi-stage snowflake structure which is expanded layer by layer from low to high, and an nth stage snowflake structure comprises a plurality of nth-1 to 0 stage snowflake structures;
virtual network request unit: generating a virtual network request according to the application requirement of the user terminal, wherein the virtual network comprises virtual nodes and virtual links of resource constraint conditions;
virtual network mapping unit: and selecting a snowflake structure level consistent with the requested virtual network scale from the network topology structure, calculating the similarity of each peer snowflake structure and the virtual network about the resource attribute, and then mapping the snowflake structure with the highest similarity onto the virtual network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310694407.8A CN116436730B (en) | 2023-06-13 | 2023-06-13 | Virtual network mapping method and mapping system based on cloud computing service |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310694407.8A CN116436730B (en) | 2023-06-13 | 2023-06-13 | Virtual network mapping method and mapping system based on cloud computing service |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116436730A true CN116436730A (en) | 2023-07-14 |
| CN116436730B CN116436730B (en) | 2023-09-29 |
Family
ID=87081891
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310694407.8A Active CN116436730B (en) | 2023-06-13 | 2023-06-13 | Virtual network mapping method and mapping system based on cloud computing service |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116436730B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104363253A (en) * | 2014-12-12 | 2015-02-18 | 北京奇虎科技有限公司 | Website security detecting method and device |
| CN107005482A (en) * | 2014-09-22 | 2017-08-01 | 沃尔廷控股公司 | For software defined network, storage and the compiler and method for calculating execution operation |
| CN107196806A (en) * | 2017-06-09 | 2017-09-22 | 西安电子科技大学 | The topology radiated based on subgraph closes on matching virtual network mapping method |
| US20210160097A1 (en) * | 2019-11-21 | 2021-05-27 | Oracle International Corporation | System and method for providing a multi-dimensional ring-lattice network topology |
| CN112887156A (en) * | 2021-02-23 | 2021-06-01 | 重庆邮电大学 | Dynamic virtual network function arrangement method based on deep reinforcement learning |
| US20210409277A1 (en) * | 2020-06-29 | 2021-12-30 | Cisco Technology, Inc. | Generation and deployment of inherited network topology models |
-
2023
- 2023-06-13 CN CN202310694407.8A patent/CN116436730B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107005482A (en) * | 2014-09-22 | 2017-08-01 | 沃尔廷控股公司 | For software defined network, storage and the compiler and method for calculating execution operation |
| CN104363253A (en) * | 2014-12-12 | 2015-02-18 | 北京奇虎科技有限公司 | Website security detecting method and device |
| CN107196806A (en) * | 2017-06-09 | 2017-09-22 | 西安电子科技大学 | The topology radiated based on subgraph closes on matching virtual network mapping method |
| US20210160097A1 (en) * | 2019-11-21 | 2021-05-27 | Oracle International Corporation | System and method for providing a multi-dimensional ring-lattice network topology |
| US20210409277A1 (en) * | 2020-06-29 | 2021-12-30 | Cisco Technology, Inc. | Generation and deployment of inherited network topology models |
| CN112887156A (en) * | 2021-02-23 | 2021-06-01 | 重庆邮电大学 | Dynamic virtual network function arrangement method based on deep reinforcement learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116436730B (en) | 2023-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | Fog-computing-enabled cognitive network function virtualization for an information-centric future Internet | |
| Liu et al. | DRL-R: Deep reinforcement learning approach for intelligent routing in software-defined data-center networks | |
| JP5390413B2 (en) | Hierarchically clustered P2P streaming system | |
| Huang et al. | A services routing based caching scheme for cloud assisted CRNs | |
| CN111130858B (en) | Dynamic multi-target virtual network mapping method in SD-WAN scene | |
| CN108923979B (en) | Software defined network virtual network mapping method | |
| CN116389491B (en) | Cloud edge computing power resource self-adaptive computing system | |
| Kumar et al. | Using clustering approaches for response time aware job scheduling model for internet of things (IoT) | |
| Kang et al. | Application of adaptive load balancing algorithm based on minimum traffic in cloud computing architecture | |
| CN115665227B (en) | Universal heterogeneous integrated computing network resource intelligent adaptation network architecture and method | |
| Ning et al. | Social-oriented resource management in cloud-based mobile networks | |
| Khodaparas et al. | A software-defined caching scheme for the Internet of Things | |
| Simoens et al. | Challenges for orchestration and instance selection of composite services in distributed edge clouds | |
| Xu et al. | Near-optimal and collaborative service caching in mobile edge clouds | |
| Xie et al. | Mesh-of-Torus: a new topology for server-centric data center networks | |
| Shreyas et al. | Selection of optimal path for the communication of multimedia data in internet of things | |
| CN116436730B (en) | Virtual network mapping method and mapping system based on cloud computing service | |
| Wang et al. | An efficient and reliable service customized routing mechanism based on deep learning in IPv6 network | |
| Wang et al. | Research on load balancing technology for microservice architecture | |
| CN110932920A (en) | Network topology structure | |
| CN116455817A (en) | Software-defined cloud network fusion architecture and route implementation method | |
| CN114401199B (en) | Hierarchical network capability model management method | |
| Xiao et al. | A sub-action aided deep reinforcement learning framework for latency-sensitive network slicing | |
| CN107948070B (en) | A kind of mobile P 2 P network virtual link choosing method based on QoS | |
| Lei et al. | Cost and qos balanced service function graph scaling algorithm in cloud-edge collaborative multimedia iot |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230903 Address after: Building F1, China Soft International Xi'an Science and Technology Park, No. 3639 Yunshui 1st Road, High tech Zone, Xi'an City, Shaanxi Province, 710061 Applicant after: China soft International Technology Service Co.,Ltd. Address before: Room O-412, 4th Floor, Lido Hotel, No. 6 Jiangtai Road, Chaoyang District, Beijing, 100016 Applicant before: Thales Digital Technology (Beijing) Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |