CN113938930B - Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene - Google Patents
Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene Download PDFInfo
- Publication number
- CN113938930B CN113938930B CN202111540182.8A CN202111540182A CN113938930B CN 113938930 B CN113938930 B CN 113938930B CN 202111540182 A CN202111540182 A CN 202111540182A CN 113938930 B CN113938930 B CN 113938930B
- Authority
- CN
- China
- Prior art keywords
- vnf
- service
- time
- node
- constructing
- 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
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- 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
- H04L41/0806—Configuration setting for initial configuration or provisioning, e.g. plug-and-play
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45595—Network integration; Enabling network access in virtual machine instances
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a method for constructing a virtual network function forwarding graph adaptive to a 5G network multi-service scene, which relates to the technical field of virtual network functions and is used for constructing the virtual network function forwarding graph to guarantee service requests of various service scenes. The invention divides the scene requirements into four types, and respectively provides four VNF-FG design methods based on the four types: 1) the scenario applicable to the free VNF-FG emphasizes the resource optimization efficiency, such as least bandwidth consumption, highest utilization rate of computing resources and the like; 2) the scenes suitable for the directed VNF-FG emphasize the strict sequence among the functional processes; 3) the determined VNF-FG is applicable to a scene which not only requires the precedence relationship of function processing, but also requires that the service can be completed timely, on time and cooperatively; 4) the hybrid VNF-FG includes many or all of the above characteristics.
Description
Technical Field
The invention relates to the technical field of virtual network functions, in particular to a method for constructing a virtual network function forwarding graph suitable for a 5G network multi-service scene.
Background
The fifth generation communication system era is met by mobile communication networks, the functions of traditional network equipment can be reconstructed by using virtualization (NFV) and Software Defined Networking (SDN) technologies, guaranteed network capacity is provided for the coming Network Services (NS), the mobile communication services are expanded from mobile phones, mobile broadband and large-scale machine type communication to the vertical field with special requirements on the communication services, and technical support is provided for the coming intelligent society.
The network service is composed of a series of Virtual Network Functions (VNFs) connected in series, the VNFs can be dynamically created and deleted, flexibility is high, the VNFs can be expanded, deployment is more convenient, resource utilization is more efficient, and operation cost and capital expenditure are reduced. The construction and deployment of the network service mainly consider the number of virtual network functions, the sequence of the network functions in the service function chain and the resource allocation situation of the whole service function chain in the virtualization infrastructure (NFVI), and the logical topology including a plurality of service chain structures can be represented by a virtual network function forwarding graph (VNF-FG), so that the guarantee of the service request can be realized through the virtual function forwarding graph.
At present, in network service deployment, a VNF-FG is generally used as a known input quantity to research mapping and scheduling problems between a forwarding graph and underlying infrastructure resources, and there is a proposed VNF-FG design process, which is directed to a single telecommunication network or a data center network mostly, and service scenarios in a 5G network are diversified, and features and requirements thereof are also different, for example, an industrial network is an important field of 5G vertical industry digital transformation, and is small in industrial application scale and has a deterministic feature, and a data center scenario emphasizes resource utilization optimization, and a single VNF-FG design method is difficult to support service requirements in a 5G network under multiple scenarios.
Disclosure of Invention
In view of this, the present invention provides a method for constructing a virtual network function forwarding graph adapted to a 5G network multi-service scenario, where the constructed virtual network function forwarding graph includes a free VNF-FG, a directed VNF-FG, a deterministic VNF-FG, and a hybrid VNF-FG, and is respectively adapted to four different 5G network service scenarios.
In order to solve the technical problem, the method for constructing the virtual network function forwarding graph suitable for the 5G network multi-service scene comprises the following steps:
s1, the system configuration module obtains the resource capability provided by the device template, parses the service requirement description file given by the service itself, extracts the service parameters from the service requirement description file, maps the extracted service parameters to the referenced VNFD, the internal and external connection points CP and the virtual link descriptor VLD of the communication network service request, and stores the mapped service parameters in the network service description file, i.e., the NSD file, where the VNFD is the referenced virtual network function VNF descriptor.
And S2, virtualizing the device template into a VNF template in the cloud 5G network, pre-configuring a VNF node and instantiating a returned device instance to the system configuration module by the device configuration module according to the device template capability and parameter requirements issued by the system configuration module, and mapping the parameter requirements of the device template into the VNFD by the system configuration module in the cloud 5G network to instantiate the VNF instance.
S3, analyzing the NSD file, designing a virtual network function forwarding graph VNF-FG according to the network service request, mapping VNF instances and virtual links in the virtual network function forwarding graph VNF-FG to a virtualization infrastructure NFVI, namely mapping a plurality of VNF-FGs to a physical server or a host, distributing resources on the physical server or the host on the bottom layer, and starting a virtual machine to complete the deployment of the VNF.
And constructing different virtual network function forwarding graphs (VNF-FG) according to different service scenes.
And when the service scene is a telecommunication service or cloud data center scene, constructing a free VNF-FG.
And when the business scene is an industrial internet, constructing a directed VNF-FG.
When the service scene is a scene needing to ensure the completion time of the service full flow service or the sequence of completion among services, a determined VNF-FG is constructed, and the application scene of the determined VNF-FG comprises a timely service scene, an on-time service scene and a collaborative service scene.
When the service scene is combined by multiple scenes, a mixed VNF-FG and a mixed VNF-FG are constructed, and a plurality of topological areas are simultaneously provided and are respectively used for constructing a free VNF-FG, a directed VNF-FG and a determined VNF-FG.
Further, for the free-form VNF-FG, the following steps are employed:
step 101) calculating the flow companding ratio of each involved VNF node; the traffic companding ratio refers to the ratio of the flow rates before and after flowing through the VNF node.
Step 102) arranging the VNF nodes from small to large by adopting quick sorting according to the flow companding ratio of each type of VNF nodes.
Step 103) initializes an empty priority queue M.
Step 104) enqueuing the arranged nodes in sequence to M, and selecting the VNF node of the same type with the lowest load rate during enqueuing.
Step 105) dequeuing and connecting VNF nodes in a first-in first-out order aiming at the queue M until the queue is empty, and constructing a free type VNF-FG.
Further, when mapping the multiple VNF-FGs to the physical server or the host, in order to stagger peaks and troughs between VNF loads in the VNF-FGs, resource utilization is balanced in a time dimension, specifically, the following steps are performed:
step 201) the load of a single VNF-FG is the linear superposition of all VNF nodes owned by the single VNF-FG, and in the ith VNF-FG, the load in one period is sampled and a load matrix is establishedM i ：
In the formula, the first row to the third row respectively represent load sampling vectors of a CPU, a memory and a bandwidth, and beta is the sampling frequency in one period; wherein in the ith VNF-FG,c i 0 ~ c i β-1 the load sampling value is sampled at 0-beta-1 time by the CPU in the period;m i 0 ~m i β-1 load sampling values of 0-beta-1 sampling exist in the period;b i 0 ~b i β-1 the sampling value is a load sampling value of which the bandwidth is sampled from 0 to beta-1 times in a period.
Step 202) if the operation is in the initial stage, setting each physical server or host machine to be empty.
Step 203) measures the correlation between load sample vectors of CPU, memory and bandwidth by using Pearson correlation coefficient.
Step 204) calculating the correlation between VNF-FGs, wherein the correlation between the ith VNF-FG and the jth VNF-FG isμ i,j ：
μ i,j =Pearson ij (M i 0 ,M j 0 )+Pearson ij (M i 1 ,M j 1 )+Pearson ij (M i 2 ,M j 2 )
Wherein i represents the ith VNF-FG, j represents the jth VNF-FG,M i 0 ,M j 0 load sampling vectors of the CPU corresponding to the ith VNF-FG and the jth VNF-FG respectively,Pearson ij (M i 0 ,M j 0 ) Representing the Pearson correlation coefficient of the load sampling vector of the CPU corresponding to the ith and jth VNF-FG,M i 1 ,M j 1 load sampling vectors of memories corresponding to the ith and the j VNF-FG respectively,Pearson ij (M i 1 , M j 1 ) Representing the Pearson correlation coefficient of the load sampling vector of the memory corresponding to the ith and the jth VNF-FG,M i 2 ,M j 2 load sample vectors of bandwidths corresponding to the ith and j VNF-FG respectively,Pearson ij (M i 2 ,M j 2 ) And representing Pearson correlation coefficients of load sample vectors of bandwidths corresponding to the ith and jth VNF-FGs.
Step 205), traversing all VNF-FGs to be mapped, combining all VNF-FG loads and the VNF-FGs to be mapped for one physical server or host, calculating correlation coefficients between every two VNF-FGs in the combination, adding the correlation coefficients to obtain the correlation coefficient sum of the current physical server or host, selecting the physical server or host with the minimum correlation coefficient sum in each mapping, and mapping the VNF-FGs to be mapped.
Further, the directed VNF-FG is constructed by the following steps:
step 301) connecting VNF nodes according to the order of the service processing functions, with VNFs in the network as vertices and connected virtual links as edges, thereby constructing a directed acyclic graph in graph theory.
Step 302) initializes an array a and stores the VNF node in the array.
Step 303) initializes an empty linked list M.
Step 304) puts VNFs with an incoming number of 0 into the linked list M, where an incoming number of 0 indicates that no other VNF depends on the node, and subtracts 1 from the incoming numbers of other nodes connected to the node.
Step 305) there are multiple instances of the same type of VNF, scheduling the underlying infrastructure to provide the corresponding resources when the VNF is put into the linked list M.
Step 306) if the current array A is not empty, jumping to Step304, otherwise, entering Step 307.
Step 307) connecting the node elements in the linked list M from the beginning in sequence, thereby obtaining the topological sequence of the forwarding graph and constructing a directed VNF-FG.
Further, the deterministic VNF-FG was constructed using the following steps:
step 401) initializing the topological graph by adopting a directed VNF-FG design method.
Step 402) obtaining service time and time sequence constraint, wherein the constraint conditions of different service types are different, and the time constraint conditions of the service are limited based on the three types of services.
Step 403) establishing a cyclic forwarding queue to plan and control the waiting time of the service in each node, wherein the cyclic forwarding queue is two queues with the same size, the service is alternately received and sent in an odd-even time slot, the period size Q of a single queue is initialized according to the steady state.
ServiceThe completion time of the node is determined by the transmission time of the virtual link and the total waiting time of each node, the waiting time of a single node is determined by the waiting times of the service in a circular forwarding queue in the node, the transmission time of the link is obtained by a distance measurement mode, and the waiting time of each VNF node is calculated
Calculating formula and completion time of serviceTThe calculation formula of (a) is as follows:
D i =Q×k i ；
in the formulak i The waiting times in the VNF node i, t is the measurable virtual link transmission time and the total time delay jitter when the link is transmitted, and n is the number of nodes contained in the VNF-FG.
Step 404) successively increasing the total waiting times in the queue, and storing the service completion time from small to large in an array B.
Step 405) if the service is a timely service, searching for the worst service completion time by adopting binary search in the group B, selecting a total waiting time scheme corresponding to the service completion time before the worst service completion time, and uniformly distributing the waiting times in each node, thereby realizing time constraint on each VNF node on the topological graph.
If the service is the punctual service, searching for a corresponding time interval in the group B by adopting binary search, selecting a total waiting time scheme corresponding to the service completion time in the corresponding time interval, and uniformly distributing the waiting times in each node, thereby realizing the time constraint on each VNF node on the topological graph.
If the service is a collaborative service, service completion time is sequentially selected in the array B according to the arrival sequence of the collaborative service, and the total waiting times corresponding to the selected service completion time are averagely distributed to each VNF node, so that the time constraint of each VNF node is realized on the topological graph.
Has the advantages that:
1. the invention provides a virtual network function forwarding graph constructed to guarantee service requests of various service scenes, and provides a design method of the virtual network function forwarding graph suitable for 5G network multi-service scenes. The invention divides the scene requirements into four types, and respectively provides four VNF-FG design methods based on the four types: 1) the scenario applicable to the free VNF-FG emphasizes the resource optimization efficiency, such as least bandwidth consumption, highest utilization rate of computing resources and the like; 2) the scenes suitable for the directed VNF-FG emphasize the strict sequence among the functional processes; 3) the determined VNF-FG is applicable to a scene which not only requires the precedence relationship of function processing, but also requires that the service can be completed timely, on time and cooperatively; 4) the hybrid VNF-FG includes many or all of the above characteristics.
2. In the invention, parameter information such as time, connection points, VNFD and the like is read from the NSD file, and the matched VNF-FG design method is judged. 1) The free type VNF-FG has no requirements on parameters such as time, connection points and the like, so the VNFs are connected according to the traffic companding ratio of the nodes, the free type VNF-FG is formed, and the traffic companding ratio influences the consumption of computing resources of the nodes and bandwidth resources of virtual links.
3. The directed VNF-FG constructed by the invention has clear requirements on parameters of connection points, the directed VNF-FG is a directed graph in a graph theory on a logic topology, a VNFD descriptor in a previous VNF node specifies adjacent connection points and attributes, therefore, the number of entries of the node is put into a priority queue according to the concept of the graph theory, the number of entries of the node corresponds to the number of accesses of the node, when a certain node is put into the priority queue, the number of entries of the next node adjacent to the node is reduced by one, the value is zero and is put into the queue, the steps are repeated until all the nodes are stored into the queue, VNF nodes are popped out in a first-in first-out sequence and are connected in series, and the directed VNF-FG is constructed.
4. The established and finalized VNF-FG has clear requirements on parameters such as connection points, time and the like, not only is a directed graph on a logic topology, but also the service completion time needs to be guaranteed when the VNF-FG is designed. Because the time of the virtual link can be measured, counted or predicted, the time requirement is refined and decomposed to the calculation processing process of each VNF node, when the traffic flow reaches each VNF node, the waiting processing time of the traffic flow is directly controlled by using the circular forwarding queue, and the time constraint is given to each VNF node in the directed graph, so that the deterministic VNF-FG is constructed.
Drawings
FIG. 1 is a diagram of a network service deployment architecture;
FIG. 2 is a flow diagram of a free-form VNF-FG design;
FIG. 3 is a flow diagram of a directed VNF-FG design;
FIG. 4 is a schematic diagram of deterministic traffic scene classification;
FIG. 5 is a flowchart of a deterministic VNF-FG design;
fig. 6 is a schematic diagram of AMI service deployment.
Detailed Description
The invention is described in detail below by way of example with reference to the accompanying drawings.
The invention provides a design method of a virtual network function forwarding graph suitable for a 5G network multi-service scene, which comprises the following specific steps:
s1, the system configuration module obtains the resource capability provided by the device template, analyzes the XML description file given by the service itself, extracts the time, quality parameter, logic node type and processing function, maps the parameters into VNFD, internal and external connection point CP and virtual link descriptor for the communication network service request, stores the related parameters into the NSD file of the network service description file, and stores the related parameters into the VNFD, internal and external connection point CP and virtual link descriptor VLD of the referenced VNF.
And S2, virtualizing the device template into a VNF template in the cloud network, pre-configuring the VNF and instantiating a returned device instance by the device configuration module according to the device template capability and parameter requirements issued by the system configuration module, and mapping the parameter requirements of the device into the VNFD and instantiating the VNF instance in the cloud 5G network after the system configuration module receives the device instance. The parameters of the device include the computation, storage and network resource parameters required by the device.
S3, parsing the NSD file, designing a VNF-FG according to the network service request, mapping a VNF instance and a virtual link in the VNF-FG to the NFVI, allocating resources on a bottom physical server or a host, and starting a virtual machine to complete deployment of the VNF, as shown in fig. 1.
It should be emphasized that, due to different business scenarios, the VNF-FG designs differently, and can be divided into a free VNF-FG, a directed VNF-FG, a deterministic VNF-FG, and a hybrid VNF-FG.
Aiming at the free VNF-FG, the use scenes of the free VNF-FG are mostly telecommunication services or cloud data center scenes, such as IMS services and GI-LAN services, the problems of when a virtual network function node VNF is accessed and issued strategies, the reliability of the nodes and the like are only needed to be considered, the sequential dependency relationship among the nodes is not needed to be concerned, therefore, the free VNF-FG network topology is not unique, and the optimization of resources is designed and emphasized.
The traffic companding ratio refers to the ratio of the flow rate before and after the VNF node, which affects the VNF processing power and link bandwidth consumption. The VNFs of different network functions have different traffic companding ratios due to different working properties, for example, the traffic companding ratio of a firewall is smaller than 1 due to a discarding mechanism; the video transcoder may cause the outgoing rate of the stream to be greater than the incoming rate due to the different formats; the VPN agent adds IPsec header overhead to the data packet, thereby increasing the bandwidth of the traffic. Based on this, a design method of the free-form VNF-FG is proposed, as shown in fig. 2. The method comprises the following specific steps:
step 101) calculating the flow companding ratio of each involved VNF node;
step 102) arranging the VNF nodes from small to large by adopting quick sorting according to the flow companding ratio of each type of VNF nodes;
step 103) initializing an empty priority queue M;
step 104) enqueuing the arranged nodes in sequence, wherein VNF nodes of the same type are selected to have a lower load rate during enqueuing;
step 105) dequeuing and connecting VNF nodes in a first-in first-out sequence aiming at the queue M until the queue is empty, and constructing a VNF-FG; (existing methods do not consider flow companding ratio)
Further, when mapping a plurality of VNF-FGs to a physical server or a host, in order to stagger peaks and troughs between VNF loads in each VNF-FG, resource utilization is balanced in a time dimension, the specific steps include:
step 201) the load of a single VNF-FG is the linear superposition of all VNF nodes owned by the single VNF-FG, and in the ith VNF-FG, the load in one period is sampled and a load matrix is establishedM i ：
In the formula, the first row to the third row respectively represent load sampling vectors of a CPU, a memory and a bandwidth, and beta is the sampling frequency in one period; wherein in the ith VNF-FG,c i 0 ~ c i β-1 the load sampling value is sampled at 0-beta-1 time by the CPU in the period;m i 0 ~m i β-1 load sampling values of 0-beta-1 sampling exist in the period;b i 0 ~b i β-1 the sampling value is a load sampling value of which the bandwidth is sampled from 0 to beta-1 times in a period.
Step 202), if the operation is in the initial stage, setting each physical server or host machine to be empty;
step 203) measuring the correlation among the load vectors by using Pearson correlation coefficients, wherein X and Y are two vectors respectively, and the larger the correlation among the vectors is, the correlation among the load vectors isPearson xy The larger the value:
step 204) calculating the correlation between VNF-FGs, wherein the correlation between the ith VNF-FG and the jth VNF-FG isμ i,j ：
μ i,j =Pearson ij (M i 0 ,M j 0 )+Pearson ij (M i 1 ,M j 1 )+Pearson ij (M i 2 ,M j 2 )
Wherein i represents the ith VNF-FG, j represents the jth VNF-FG,M i 0 ,M j 0 load sampling vectors of the CPU corresponding to the ith VNF-FG and the jth VNF-FG respectively,Pearson ij (M i 0 ,M j 0 ) Representing the Pearson correlation coefficient of the load sampling vector of the CPU corresponding to the ith and jth VNF-FG,M i 1 ,M j 1 load sampling vectors of memories corresponding to the ith and the j VNF-FG respectively,Pearson ij (M i 1 , M j 1 ) Representing the Pearson correlation coefficient of the load sampling vector of the memory corresponding to the ith and the jth VNF-FG,M i 2 ,M j 2 load sample vectors of bandwidths corresponding to the ith and j VNF-FG respectively,Pearson ij (M i 2 ,M j 2 ) And representing Pearson correlation coefficients of load sample vectors of bandwidths corresponding to the ith and jth VNF-FGs.
Step 205), traversing all VNF-FGs to be mapped, combining all VNF-FG loads and the VNF-FGs to be mapped for one physical server or host, calculating correlation coefficients between every two VNF-FGs in the combination, adding the correlation coefficients to obtain the correlation coefficient sum of the current physical server or host, selecting the physical server or host with the minimum correlation coefficient sum in each mapping, and mapping the VNF-FGs to be mapped.
Fig. 2 shows a free-form VNF-FG design flow diagram.
Aiming at the directed VNF-FG, the use scenes of the directed VNF-FG are mostly industrial internet, and the directed VNF-FG has strict sequential logic relationship when traversing the virtual network function nodes, for example, in the power acquisition service, firstly, according to the service application of the metering master station, the Web server initiates a service request to an acquisition server, the acquisition server translates the service requirement into a network requirement index by using an acquisition protocol coding function and sends the network requirement index to a front-end processor server, the front-end processor server codes messages to ensure that the messages are suitable for bearing formats of different technical systems, and accessing different transmission channels into a heterogeneous public network through a firewall function, wherein the functions comprise monitoring of network states, cooperative switching and the like, finally, the service message is issued to the acquisition terminal, the acquisition terminal analyzes the acquisition requirement and then executes the acquisition action of the power consumption information, and the unidirectional power consumption information acquisition service requirement issuing process is ended. After the functions are virtualized into VNF nodes, the functions need to be executed according to the sequence of function processing. Based on this, a design method of the directed VNF-FG is proposed, as shown in fig. 3. The method comprises the following specific steps:
step 301) regarding a VNF in a network as a vertex in a graph, regarding connected virtual links as edges in the graph, and connecting VNF nodes according to a sequence of service processing functions, thereby constructing a directed acyclic graph in graph theory;
step 302) initializing an array A and storing the VNF node into the array;
step 303) initializing an empty linked list M;
step 304) placing the VNF with the degree of entry of 0 into the linked list M, wherein the degree of entry of 0 indicates that no other VNF depends on the node, and subtracting the degree of entry of other nodes connected with the node by one;
step 305) multiple instances may exist for the same type of VNF, scheduling the underlying infrastructure to provide the corresponding resources when the VNF is put in, assuming for virtualFirst of network function resource requirement listkIndividual network function unit VNF-C k Which is required tod C K A unit of node computing resources. For two adjacent network function units on the function chain SFCC k AndC K+1which requires allocationd C K,K+1 A unit of link communication resources;
step 306), if the current array A is not empty, jumping to Step304, otherwise, entering Step 307;
step 307) connecting the node elements in the linked list M from the beginning in sequence, thereby obtaining the topological sequence of the forwarding graph and constructing a directed VNF-FG.
Figure 3 shows a directed VNF-FG design flow diagram.
Aiming at a deterministic VNF-FG, traversing virtual network function nodes is required to have strict logical relations, and meanwhile, the completion time of the service full flow or the completion sequence among services, such as closed-loop control, real-time cloud driving control, remote operation, on-off control services and the like of industrial application, are still required to be ensured.
1) The timely type service needs to ensure that the time for completing the service is short, namely the service needs to arrive as soon as possible;
2) the punctual service needs to ensure that the time jitter for completing the service is small, namely, the service is required to come fast and be punctual;
3) the service coordination of the same type needs to ensure the arrival time among different concurrent services, that is, the time for completing the service among the services needs to have a sequential relationship.
Fig. 4 shows a deterministic traffic scenario classification diagram.
Based on the above, a design method of a deterministic VNF-FG is provided, a topological graph is initialized by adopting a design method of a directed VNF-FG, and then each node is given time constraint, as shown in FIG. 5. The method comprises the following specific steps:
step 401) initializing a topological graph by adopting a directed VNF-FG design method;
step 402) obtaining service time and time sequence constraint, wherein the constraint conditions of different service types are different, and the time constraint conditions of the services are limited based on the three types of services:
wherein J = {1,2,3} is a set of different traffic types, x, y are adjacent VNF nodes, and binary variablesX x,y j Time =1 indicates that traffic stream J flows through the virtual link<x，y>To nodes x, y and vice versaX x,y j =0。J int Timely type service, requiring guarantee of lower bound of delay
； j ont On-time service, the upper and lower time delay bounds need to be guaranteed
And
；j co for the cooperative service, the time sequence logic between service arrivals needs to be guaranteed,d x,y which is the time of transmission between the adjacent nodes,
is the network access time of the service.
Step 403), establishing a cyclic forwarding queue to plan and control the waiting time of the service in each node, wherein the cyclic forwarding queue is two queues with the same size, alternately receiving and sending the service in an odd-even time slot, the period size Q of a single queue is initialized according to a steady-state criterion;
d ave =Q/8(1-2q)+x 2/2q(1-x)-0.65(Q/q 2)1/3 x 9/2
in the formulad ave In order to average the time delay of the service flow,qfor the VNF node the actual throughput,xis the throughput saturation, i.e., the ratio of the actual throughput to the maximum throughput. The completion time of the service is determined by the transmission time of the virtual link and the total waiting time of each node, further, the waiting time of a single node is determined by the waiting times of the service in the cyclic forwarding queue in the node, the transmission time of the link is obtained by means of ranging and the like, and the waiting time of each VNF nodeD i Calculating formula and completion time of serviceTThe calculation formula of (a) is as follows:
D i =Q×k i ；
in the formulak i The waiting times in the node i are, t is the measurable transmission time and the measurable transmission time of the virtual link, t is the total time delay jitter during the link transmission, and n is the number of nodes contained in the VNF-FG.
Step 404) gradually increasing the total waiting times in the queue, and storing the service completion time into an array B from small to large; the length of the array B is set as the waiting timesk i +2。
Step 405) if the service is a timely service, searching for the worst service completion time by adopting binary search in the groups, selecting a total waiting time scheme corresponding to the service completion time before the worst service completion time, and uniformly distributing the waiting times in each node, thereby realizing the time constraint on each VNF node on the topological graph.
If the service is the punctual service, searching for a corresponding time interval in the group B by adopting binary search, selecting a total waiting time scheme corresponding to service completion time in the time interval, and uniformly distributing the waiting times in each node, thereby realizing time constraint on each VNF node on the topological graph.
If the service is a collaborative service, service completion time is sequentially selected in the array B according to the arrival sequence of the collaborative service, and the total waiting times corresponding to the selected service completion time are averagely distributed to each VNF node, so that the time constraint of each VNF node is realized on the topological graph.
For a hybrid VNF-FG, which may have many or all of the characteristics of a free type, a directed type, and a hybrid type at the same time, the above-mentioned manner is adopted for designing in the topology areas of the respective types.
Figure 5 shows a deterministic VNF-FG design flow diagram.
Example 1
First, for the design method of the free VNF-FG, it is assumed that there are three VNFs in the network, which are respectively a firewall, an IDS, and a video transcoder, and for convenience of description, the function type VNF-A, VNF-B, VNF-C is used in this embodiment. The three VNFs have no strict sequential processing relationship in function execution, the traffic companding ratio of function type a is 0.5, the processing capacity is 10, the traffic companding ratio of function type B is 1, the processing capacity is 8, the traffic companding ratio of function type C is 2, the processing capacity is 15, and the initial traffic rate of the user is set to 10.
The steps in the free VNF-FG design method are adopted, ascending arrangement is carried out according to the flow compression-expansion ratio, the VNF-FG is formed, the logic connection of the VNF-FG is a function type A, a function type B and a function type C, the effectiveness and the superiority of the method are verified, and the influences on the VNF processing capacity and bandwidth consumption in different design schemes are calculated respectively.
Calculating the consumption condition of the resources of the VNF according to the initial flow, the flow companding ratio and the processing capacity of each functional node:
1) the logic connection solved by the free VNF-FG design method is a function type A, a function type B and a function type C, the initial flow rate is 10, the processing capacity of the function type A is 10, no additional VM is needed to be started for bearing, the rate is changed to 5 after the flow reaches the function type B, the processing capacity of the function type B is 8, no additional VM is needed to be started for bearing, the rate is changed to 5 after the flow reaches the function type C in the same way, and no additional VM is needed to be started;
2) when the logical link is a function type a, a function type C, or a function type B, first, the rate when the traffic reaches the function type a is 10, the VM does not need to be additionally started, the rate when the traffic reaches the function type C is 20, 2 VMs need to be started to carry, the rate when the traffic reaches the function type B is 20, and 3 VMs need to be started to carry;
3) when the logical links are a function type B, a function type A and a function type C, firstly, the rate when the flow reaches the function type B is 10, 2 VMs are required to be started for carrying, the rate when the flow reaches the function type A is 10, no additional starting is required, the rate when the flow reaches the function type C is 5, and no additional VM carrying is required to be started;
4) when the logical link is a function type B, a function type C, or a function type a, first, the rate when the traffic reaches the function type B is 10, 2 VMs need to be started to carry, the rate when the traffic reaches the function type C is 10, no additional start is needed, and the rate when the traffic reaches the function type a is 2 VMs need to be started to carry;
5) when the logical link is a function type C, a function type a, or a function type B, first, the rate when the traffic reaches the function type C is 10, no additional VM is needed, the rate when the traffic reaches the function type a is 20, 2 VMs need to be started to carry, the rate when the traffic reaches the function type B is 10, and 2 VMs need to be started to carry;
6) when the logical link is a function type C, a function type B, or a function type a, first, the rate when the traffic reaches the function type C is 10, no additional VM is needed, the rate when the traffic reaches the function type B is 20, 3 VMs need to be started to carry, the rate when the traffic reaches the function type a is 20, and 2 VMs need to be started to carry;
calculating the consumption of link bandwidth resource according to the initial flow and the flow companding ratio, wherein the logical connection mode is as above, adoptingB1~B6 respectively replacing:
B1=10+5+5=20
B2=10+5+10=25
B3=10+10+5=25
B4=10+10+20=40
B5=10+20+10=40
B6=10+20+20=50
therefore, compared with the method for designing the free type VNF-FG, the bandwidth consumption is relatively minimum, the number of the started VMs is relatively minimum, and the resource utilization efficiency is highest.
For the traversal VNF-FG design method, in this embodiment, an Advanced Metering Infrastructure (AMI) is taken as an example for description, and an AMI service deployment process is shown in fig. 6, where execution of functions has strict sequential logic. The design method of the traversal VNF-FG is as follows:
1) respectively virtualizing functions of metering, collecting, gateway, front-end processor, database and application display into virtual network nodes of vCPE, MEC, PSA, UDM and DN, planning a directed acyclic graph, initializing an array A, and storing the VNF node into the array;
2) initializing an empty linked list M;
3) putting vCPE with the degree of incidence of 0 into a linked list M, wherein the degree of incidence of connected MEC nodes is 0, and the other nodes are unchanged;
4) if the current array A is not empty, jumping to the step 3), putting the MEC node into a linked list, wherein the input number of the PSA node is 0, and the remaining nodes are unchanged;
5) repeating the operation until the array A is empty, which indicates that all the nodes are placed in the linked list;
6) and because the linked list has a traversal type, sequentially connecting the node elements in the linked list M from the beginning to construct a traversal type VNF-FG.
The initial topology is based on a traversal-type VNF-FG. In this embodiment, assuming that there is a switch service A, B, C, D, E, and it is required that traversal virtual network function nodes have a strict logical relationship, it is still necessary to ensure that switches a to E between services reach a destination end in ascending order according to the alphabet, and the method for designing the traversal VNF-FG is specifically as follows:
1) initializing a topological graph by adopting a traversal VNF-FG design method;
2) acquiring service timing constraints from NSD, wherein the correct arrival sequence among services is switching service A, switching service B, switching service C, switching service D and switching service E;
3) establishing a circular forwarding queue to plan and control the waiting time of the service in each node, and setting the period size of the initial circular queue to be 1msT is the sum of measurable virtual link transmission times, set to 10msΔ t is the total delay jitter when the link is transmitting, set to + -0.3ms；
4) Gradually increasing the total waiting times in the queue, storing the service completion time into the array from small to large, and intercepting the first 20 bits of the array for display: 11.2878, 11.9633, 12.7667, 13.8548, 14.9452, 16.0569, 16.8573, 18.0617, 19.1267, 20.6661, 21.5409, 22.7560, 23.7825, 24.9090, 26.0094, 26.5026, 27.7150, 29.3612, 30.1526, 32.7719
5) And sequentially selecting 5 times, such as 11.2878, 12.7667, 14.9452, 16.8573 and 18.0617, from the array, respectively corresponding to 1, 3, 5, 7 and 8 total waiting times, distributing the times into each node according to the node load condition, and finally, effectively ensuring that the services A-E among the switches correctly reach a destination end in an ascending order according to the alphabet.
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. The method for constructing the virtual network function forwarding graph adapting to the 5G network multi-service scene is characterized by comprising the following steps of:
s1, acquiring resource capability provided by the device template in the system configuration module, analyzing a service requirement description file given by the service, extracting service parameters from the service requirement description file, mapping the extracted service parameters into a VNFD, an internal and external connection point CP and a virtual link descriptor VLD which are quoted by the communication network service request, and storing the VNFD as a quoted virtual network function VNF descriptor in a network service description file, namely an NSD file;
s2, virtualizing the device template into a VNF template in the cloud 5G network, pre-configuring a VNF node and instantiating a returned device instance to the system configuration module by the device configuration module according to the device template capability and parameter requirements issued by the system configuration module, mapping the parameter requirements of the device template into a VNFD by the system configuration module in the cloud 5G network, and instantiating the VNF instance;
s3, analyzing the NSD file, designing a virtual network function forwarding graph VNF-FG according to the network service request, mapping VNF instances and virtual links in the virtual network function forwarding graph VNF-FG to a virtualization infrastructure NFVI, namely mapping a plurality of VNF-FGs to a physical server or a host, distributing resources on the physical server or the host on the bottom layer, and starting a virtual machine to complete the deployment of the VNF;
constructing different virtual network function forwarding graphs (VNF-FG) according to different service scenes;
when the service scene is a telecommunication service or cloud data center scene, constructing a free VNF-FG;
when the service scene is an industrial internet, constructing a directed VNF-FG;
when the service scene is a scene needing to ensure the completion time of the service full flow service or the sequence of completion among services, constructing a determined VNF-FG, wherein the application scene of the determined VNF-FG comprises a timely service scene, an on-time service scene and a collaborative service scene;
and when the service scene is a combination of various scenes, constructing a mixed VNF-FG, wherein the mixed VNF-FG is simultaneously provided with a plurality of topological areas and is respectively used for constructing a free VNF-FG, a directed VNF-FG and a determined VNF-FG.
2. The method for constructing the forwarding graph of virtual network functions according to claim 1, wherein the forwarding graph of virtual network functions is constructed by using the following steps for a free-form VNF-FG:
step 101) calculating the flow companding ratio of each involved VNF node; the traffic companding ratio refers to the ratio of flow rates before and after flowing through a VNF node;
step 102) arranging the VNF nodes from small to large by adopting quick sorting according to the flow companding ratio of each type of VNF nodes;
step 103) initializing an empty priority queue M;
step 104) enqueuing the arranged nodes in sequence by M, and selecting the VNF node of the same type with the lowest load rate during enqueuing;
step 105) dequeuing and connecting VNF nodes in a first-in first-out order aiming at the queue M until the queue is empty, and constructing a free type VNF-FG.
3. The method for constructing a forwarding graph for virtual network functions according to claim 2, wherein, when mapping a plurality of VNF-FGs to physical servers or hosts, in order to stagger peaks and troughs of VNF loads in each VNF-FG, resource utilization is balanced in a time dimension, specifically comprising the following steps:
step 201) the load of a single VNF-FG is the linear superposition of all VNF nodes owned by the single VNF-FG, and in the ith VNF-FG, the load in one period is sampled and a load matrix is establishedM i ：
In the formula, the first row to the third row respectively represent load sampling vectors of a CPU, a memory and a bandwidth, and beta is the sampling frequency in one period; wherein in the ith VNF-FG,c i 0 ~ c i β-1 the load sampling value is sampled at 0-beta-1 time by the CPU in the period;m i 0 ~ m i β-1 load sampling values of 0-beta-1 sampling exist in the period;b i 0 ~b i β-1 the load sampling value is sampled within the 0-beta-1 th time of the bandwidth in the period;
step 202), if the operation is in the initial stage, setting each physical server or host machine to be empty;
step 203) measuring the correlation among load sampling vectors of a CPU, a memory and a bandwidth by adopting a Pearson correlation coefficient;
step 204) calculating the correlation between VNF-FGs, wherein the correlation between the ith VNF-FG and the jth VNF-FG isμ i,j ：μ i,j =Pearson ij (M i 0 ,M j 0 )+Pearson ij (M i 1 ,M j 1 )+Pearson ij (M i 2 ,M j 2 )
Wherein i represents the ith VNF-FG, j represents the jth VNF-FG,M i 0 ,M j 0 load sampling vectors of the CPU corresponding to the ith VNF-FG and the jth VNF-FG respectively,Pearson ij (M i 0 ,M j 0 ) Representing the Pearson correlation coefficient of the load sampling vector of the CPU corresponding to the ith and jth VNF-FG,M i 1 ,M j 1 load sampling vectors of memories corresponding to the ith and the j VNF-FG respectively,Pearson ij (M i 1 ,M j 1 ) Representing the Pearson correlation coefficient of the load sampling vector of the memory corresponding to the ith and the jth VNF-FG,M i 2 ,M j 2 load sample vectors of bandwidths corresponding to the ith and j VNF-FG respectively,Pearson ij (M i 2 ,M j 2 ) Is shown asi. Pearson correlation coefficients of load sampling vectors of j VNF-FGs corresponding to the bandwidth;
step 205), traversing all VNF-FGs to be mapped, combining all VNF-FG loads and the VNF-FGs to be mapped for one physical server or host, calculating correlation coefficients between every two VNF-FGs in the combination, adding the correlation coefficients to obtain the correlation coefficient sum of the current physical server or host, selecting the physical server or host with the minimum correlation coefficient sum in each mapping, and mapping the VNF-FGs to be mapped.
4. The method of claim 3, wherein the directed VNF-FG is constructed by:
step 301) connecting VNF nodes according to the order of service processing functions, with VNFs in the network as vertices and connected virtual links as edges, thereby constructing a directed acyclic graph in graph theory;
step 302) initializing an array A and storing the VNF node into the array;
step 303) initializing an empty linked list M;
step 304) placing the VNF with the degree of entry of 0 into the linked list M, wherein the degree of entry of 0 indicates that no other VNF depends on the node, and subtracting 1 from the degree of entry of other nodes connected with the node;
step 305) multiple instances of the same type of VNF exist, and the bottom-layer infrastructure is scheduled to provide corresponding resources when the VNF is put into a linked list M;
step 306), if the current array A is not empty, jumping to Step304, otherwise, entering Step 307;
step 307) connecting the node elements in the linked list M from the beginning in sequence, thereby obtaining the topological sequence of the forwarding graph and constructing a directed VNF-FG.
5. The method of claim 4, wherein the deterministic VNF-FG is constructed by:
step 401) initializing a topological graph by adopting a directed VNF-FG construction method;
step 402) obtaining service time and time sequence constraint, wherein the constraint conditions of different service types are different, and the time constraint conditions of the services are limited based on the three types of services:
step 403), establishing a cyclic forwarding queue to plan and control the waiting time of the service in each node, wherein the cyclic forwarding queue is two queues with the same size, alternately receiving and sending the service in an odd-even time slot, and initializing the Q according to the stable state;
the service completion time is determined by the transmission time of the virtual link and the total waiting time of each node, the waiting time of a single node is determined by the waiting times of the service in the cyclic forwarding queue in the node, the transmission time of the link is obtained by a distance measurement mode, and the waiting time of each VNF node is calculatedD i Calculating formula and completion time of serviceTThe calculation formula of (a) is as follows:
D i =Q×k i ；
in the formulak i The waiting times in the VNF node i are, t is the measurable transmission time and the measurable transmission time of the virtual link, t is the total time delay jitter during the link transmission, and n is the number of nodes contained in the VNF-FG;
step 404) gradually increasing the total waiting times in the queue, and storing the service completion time into an array B from small to large;
step 405) if the service is a timely service, searching for the worst service completion time by adopting binary search in the group B, selecting a total waiting time scheme corresponding to the service completion time before the worst service completion time, and uniformly distributing the waiting times in each node, thereby realizing time constraint on each VNF node on the topological graph;
if the service is the punctual service, searching a corresponding time interval in the group B by adopting binary search, selecting a total waiting time scheme corresponding to the service completion time in the corresponding time interval, and uniformly distributing the waiting times in each node, thereby realizing the time constraint on each VNF node on the topological graph;
if the service is a collaborative service, service completion time is sequentially selected in the array B according to the arrival sequence of the collaborative service, and the total waiting times corresponding to the selected service completion time are averagely distributed to each VNF node, so that the time constraint of each VNF node is realized on the topological graph.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111540182.8A CN113938930B (en) | 2021-12-16 | 2021-12-16 | Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111540182.8A CN113938930B (en) | 2021-12-16 | 2021-12-16 | Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113938930A CN113938930A (en) | 2022-01-14 |
| CN113938930B true CN113938930B (en) | 2022-03-25 |
Family
ID=79289151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111540182.8A Active CN113938930B (en) | 2021-12-16 | 2021-12-16 | Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113938930B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115514657B (en) * | 2022-11-15 | 2023-03-24 | 阿里云计算有限公司 | Network modeling method, network problem analysis method and related equipment |
| CN117155803B (en) * | 2023-09-01 | 2024-03-12 | 广东九博科技股份有限公司 | Multi-level scene-oriented router and adaptive optimization method thereof |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108696373A (en) * | 2017-04-06 | 2018-10-23 | 华为技术有限公司 | Virtual resource allocation method, NFVO and system |
| WO2019082085A1 (en) * | 2017-10-24 | 2019-05-02 | Telefonaktiebolaget Lm Ericsson (Publ) | Methods for defining a network service descriptor (nsd) for a network service (ns), and network functions virtualization (nfv) orchestrator (nfvo) using said nsd |
| CN110190987A (en) * | 2019-05-08 | 2019-08-30 | 南京邮电大学 | Based on backup income and the virtual network function reliability dispositions method remapped |
| WO2020092174A1 (en) * | 2018-11-02 | 2020-05-07 | Intel Corporation | Systems, methods, and devices for enabling live media production services |
| CN111385114A (en) * | 2018-12-28 | 2020-07-07 | 华为技术有限公司 | VNF service instantiation method and device |
| CN111917567A (en) * | 2019-05-07 | 2020-11-10 | 慧与发展有限责任合伙企业 | Network function virtualization |
| CN112333035A (en) * | 2020-12-30 | 2021-02-05 | 中国人民解放军国防科技大学 | Real-time hybrid service function chain embedding cost optimization method and device |
| CN112968794A (en) * | 2021-01-28 | 2021-06-15 | 广州杰赛科技股份有限公司 | Network function chain deployment method, device, terminal device and storage medium |
| CN113347030A (en) * | 2021-05-26 | 2021-09-03 | 重庆大学 | Unified template management system of cloud native network scene arrangement platform |
-
2021
- 2021-12-16 CN CN202111540182.8A patent/CN113938930B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108696373A (en) * | 2017-04-06 | 2018-10-23 | 华为技术有限公司 | Virtual resource allocation method, NFVO and system |
| WO2019082085A1 (en) * | 2017-10-24 | 2019-05-02 | Telefonaktiebolaget Lm Ericsson (Publ) | Methods for defining a network service descriptor (nsd) for a network service (ns), and network functions virtualization (nfv) orchestrator (nfvo) using said nsd |
| AU2021245203A1 (en) * | 2017-10-24 | 2021-11-04 | Telefonaktiebolaget Lm Ericsson (Publ) | Methods for defining a Network Service Descriptor (NSD) for a Network Service (NS), and Network Functions Virtualization (NFV) Orchestrator (NFVO) using said NSD |
| WO2020092174A1 (en) * | 2018-11-02 | 2020-05-07 | Intel Corporation | Systems, methods, and devices for enabling live media production services |
| CN111385114A (en) * | 2018-12-28 | 2020-07-07 | 华为技术有限公司 | VNF service instantiation method and device |
| CN111917567A (en) * | 2019-05-07 | 2020-11-10 | 慧与发展有限责任合伙企业 | Network function virtualization |
| CN110190987A (en) * | 2019-05-08 | 2019-08-30 | 南京邮电大学 | Based on backup income and the virtual network function reliability dispositions method remapped |
| CN112333035A (en) * | 2020-12-30 | 2021-02-05 | 中国人民解放军国防科技大学 | Real-time hybrid service function chain embedding cost optimization method and device |
| CN112968794A (en) * | 2021-01-28 | 2021-06-15 | 广州杰赛科技股份有限公司 | Network function chain deployment method, device, terminal device and storage medium |
| CN113347030A (en) * | 2021-05-26 | 2021-09-03 | 重庆大学 | Unified template management system of cloud native network scene arrangement platform |
Non-Patent Citations (4)
| Title |
|---|
| 5G Military Application Scenarios and Private Network Architectures;Jingjing Liao et al.;《 2020 IEEE International Conference on Advances in Electrical Engineering and Computer Applications( AEECA)》;20200827;全文 * |
| Design and Implementation of Heterogeneous Orchestration for 5G’s Scenarios;Jingjing Liao et al.;《2021 IEEE 3rd International Conference on Frontiers Technology of Information and Computer (ICFTIC)》;20211114;全文 * |
| 网络功能虚拟化资源配置及优化研究综述;邵维专等;《计算机应用研究》;20170818(第02期);全文 * |
| 面向多业务需求的NFV和SDN融合的资源优化算法;朱晓荣等;《通信学报》;20181125(第11期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113938930A (en) | 2022-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Goyal et al. | Start-time fair queueing: A scheduling algorithm for integrated services packet switching networks | |
| CN113938930B (en) | Construction method of virtual network function forwarding graph adapting to 5G network multi-service scene | |
| Sariowan et al. | SCED: A generalized scheduling policy for guaranteeing quality-of-service | |
| US6393026B1 (en) | Data packet processing system and method for a router | |
| Shi et al. | Real-time communication analysis with a priority share policy in on-chip networks | |
| Zhao et al. | A unified modeling framework for distributed resource allocation of general fork and join processing networks | |
| CN115022182B (en) | Real-time flow scheduling optimization method for train communication network based on QSILP algorithm | |
| CN111030942B (en) | TTE network offline scheduling method based on response constraint | |
| Al-Turjman et al. | SAHCI: scheduling approach for heterogeneous content-centric IoT applications | |
| Serdaroglu et al. | An efficient multipriority data packet traffic scheduling approach for fog of things | |
| CN102355422A (en) | Multicore, parallel and lock-free quality of service (QOS) flow control method | |
| Chai et al. | Real-time communication model based on OPC UA wireless network for intelligent production line | |
| Gärtner et al. | On the incremental reconfiguration of time-sensitive networks at runtime | |
| CN116743658A (en) | Industrial Internet deterministic communication method and device based on OPC UA-TSN | |
| CN102594670B (en) | Multiport multi-flow scheduling method, device and equipment | |
| CN113553279B (en) | RDMA communication acceleration set communication method and system | |
| Zhang et al. | Scheduling coflows with incomplete information | |
| CN114039935A (en) | Message scheduling system and method based on distributed real-time bus configuration | |
| Mi et al. | A scheduling algorithm of maximize the number of porosity for the time-triggered DIMA system | |
| Loshkarev et al. | Performance modeling and optimization of flexible QoS-guaranteed multifunctional MAC for rural profitable ubiquitous 5G IoT/M2M systems | |
| Shi et al. | Load balance based network bandwidth allocation for delay sensitive services | |
| Chai et al. | Research on Deterministic Communication Model for Industrial Internet Based on OPC UA-TSN | |
| Wang et al. | Traffic Scheduling Method for Power Low-Latency Business in TSN Based on Sending Time Scheduled | |
| Guck | Centralized Online Routing for Deterministic Quality of Service in Packet Switched Networks | |
| CN116800620A (en) | Block chain simulation system and method |
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 |