CN116208567A - Method and system for flow scheduling of SDN network resources of cross-domain data center - Google Patents
Method and system for flow scheduling of SDN network resources of cross-domain data center Download PDFInfo
- Publication number
- CN116208567A CN116208567A CN202310126810.0A CN202310126810A CN116208567A CN 116208567 A CN116208567 A CN 116208567A CN 202310126810 A CN202310126810 A CN 202310126810A CN 116208567 A CN116208567 A CN 116208567A
- Authority
- CN
- China
- Prior art keywords
- data center
- cross
- domain data
- energy consumption
- sdn network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/78—Architectures of resource allocation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to the technical field of network communication, in particular to a method and a system for flow scheduling of SDN network resources of a cross-domain data center, wherein the method comprises the following steps: adding the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model; establishing a corresponding objective function and constraint conditions for each SDN network feature of SDN network resources respectively, and adding the objective function and constraint conditions into a flow scheduling optimization objective model; solving the flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on SDN network resources, realizing a network resource timely adjustment technology suitable for network flow dynamics, and saving the flow bandwidth cost in a data center.
Description
Technical Field
The invention relates to the technical field of network communication, in particular to a method and a system for flow scheduling of SDN network resources of a cross-domain data center.
Background
And providing a unified management and agile scheduling theory of the cross-domain SDN network resources for the integration of the computing network. Firstly, based on the difference between different power-computing services in service characteristics and service quality requirements, a differentiated service system and an evaluation method are established according to a service principle of moderate demand and global optimum, and a theoretical basis is provided for the design and optimization of a computing network integration resource efficient scheduling system.
In particular, under the condition of meeting strict time delay requirements, the multi-type resources such as computing resources, storage resources, transmission link communication resources and the like in the whole network are cooperatively scheduled, so that strict service quality guarantee is provided for the computing network convergence service, and the optimal multi-dimensional resource configuration of the whole network is realized. And then, taking time delay and energy consumption as performance indexes, establishing a mathematical model by taking edge node computing power, cloud computing node computing power, computing task allocation proportion, user allocation scheduling strategies, differential delay and the like as constraint targets through establishing different objective functions, and reasonably allocating network resources and computing resources to ensure that computing task time delay and system energy consumption are optimized under constraint conditions of limited resources, task priority and the like.
The scheduling problem of network resources is always a fundamental key problem of internet research. The advent of SDN has brought new opportunities for research in this regard. For traffic scheduling in an SDN environment, cui et al, cornell university, usa, propose an extensible multicast solution, i.e., dual structure multicast, based on analysis of multicast traffic heterogeneity of an SDN data center. An SDN resource optimization scheme applied to a data center is provided. Experimental results show that compared with the traditional scheme, the scheme can improve the link efficiency by 12% and reduce the packet loss rate by 51%. The SDN controller is used for controlling group members and related management information, and a control strategy based on a greedy algorithm is provided for dynamically distributing optical signal network resources. An equivalent multi-route scheme (ECMP) under a single controller is provided, and the overall topology is optimized from the global network resource consideration under the same network domain, so that the network resource utilization rate is improved, and the packet loss rate is reduced. The method indicates that the current network resource multipath optimization algorithm can be implemented only for a fixed topology in a certain data center, and the generalization requirement is difficult to achieve, so that the document provides an improved ant colony optimization algorithm, thereby sensing the topology state under the condition of unknown topology and meeting the universality requirement of the multipath optimization algorithm. Aiming at multipath routing arrangement, some researchers are also performing trial study and propose hierarchical multipath arrangement optimization so as to improve network performance; another approach optimizes route boost resource utilization through congestion awareness and automatic identification techniques. In 2019, multipath adaptive routing algorithms for cross-domain distributed data centers have also begun to be of interest to researchers. Therefore, in terms of unified network resource scheduling, the unified scheduling theory of the cross-domain data center of the software defined network combined with the future cloud edge fusion is still in the initial stage of research, and the following problems mainly exist at present:
1) How to combine SDN network characteristic indexes to establish a software-defined network resource optimization objective function and adjust and plan corresponding network flow and resource strategies is one of the technical problems to be solved by the patent.
2) How to combine the computing power and the storage power of the data center and the edge computing node in the cloud edge fusion-oriented cross-domain data center environment to realize low-delay and low-load global traffic scheduling based on the global network condition is also another technical problem solved by the patent.
Disclosure of Invention
The invention aims to solve the technical problem of providing a method and a system for flow scheduling of SDN network resources of a cross-domain data center aiming at the defects of the prior art.
The technical scheme of the flow scheduling method of the SDN network resources of the cross-domain data center is as follows:
adding the computing capacity of a cross-domain data center and the network performance energy consumption of an edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
establishing a corresponding objective function and constraint conditions for each SDN network feature of SDN network resources respectively, and adding the objective function and constraint conditions into the flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
and solving the flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on the SDN network resources.
The technical scheme of the system for flow scheduling of SDN network resources of the cross-domain data center is as follows:
the system comprises a first adding and establishing module, a second adding and establishing module and a solving module;
the first adding and establishing module is used for: adding the computing capacity of a cross-domain data center and the network performance energy consumption of an edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
the second adding and establishing module is used for: establishing a corresponding objective function and constraint conditions for each SDN network feature of SDN network resources respectively, and adding the objective function and constraint conditions into the flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
the solving module is used for: and solving the flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on the SDN network resources.
The beneficial effects of the invention are as follows:
in a cross-domain data center environment oriented to future cloud edge fusion, low-delay and low-load global flow scheduling based on global network conditions can be effectively realized, internal correlation among resource scheduling such as calculation, storage, network and the like is revealed, the utilization rate of software defined network resources and user experience are optimized, a network resource timely adjustment technology suitable for network flow dynamics is realized, the flow bandwidth cost in the data center is saved, the cost can be saved for related data centers and services, and indirect economic benefits are obtained.
Drawings
Fig. 1 is a flow chart of a method for flow scheduling of SDN network resources in a cross-domain data center according to an embodiment of the present invention;
FIG. 2 is a schematic flow diagram of a solution to a flow scheduling optimization mathematical model using a non-dominant genetic algorithm with elite strategy;
fig. 3 is a schematic structural diagram of a system for traffic scheduling of SDN network resources in a cross-domain data center according to an embodiment of the present invention.
Detailed Description
Software defined network features, SDN, are some of the features that can be detected directly, but some are known through computation (i.e., metrics). Furthermore, feature metrics of SDN network features typically require a detection process. The characteristic detection adopts the following technical route:
1) Analyzing network state related characteristics of the OpenFlow protocol, such as counting switch port information (number of received bytes, number of lost packets, duration time and the like), and guiding collection of each SDN network characteristic data;
2) Investigation of sampling algorithms commonly used at present, such as algorithms of uniform sampling, self-adaptive random sampling, grouping sampling, threshold sampling and the like, analysis of advantages and disadvantages of each sampling algorithm, and further combination of characteristics of SDN networks and different network characteristic measurement indexes to be studied, the sampling algorithm suitable for different SDN network characteristics is provided;
3) Qualitative and quantitative relations between sampling frequency and load and accuracy in the analysis sampling algorithm are researched, and load expenditure brought by data sampling to a controller is reduced. The flow entries are preinstalled in the network to form a detection route covering the whole network. And optimizing each mathematical model and providing a new model according to the research and pre-research results of the project group on the network characteristic mathematical model. The following describes the mathematics of each SDN network feature preliminarily defined in the present invention, specifically:
(1) link bandwidth:
symbol interpretation: where UB represents the link bandwidth, bt 1 And Bt 2 Respectively representing the initial number of bytes and the end number of bytes in the sampling interval, and p represents the sampling interval.
(2) Packet loss rate:
symbol interpretation: the sampling sample interval is [ i-1, i ], the sampling time interval is [ time (i-1), time (i) ], and droped represents the number of discarded bytes contained in the switch response information.
(3) Delay:
Delay s1&s2 ＝Totaltime-Delay c&s1 -Delay c&s2
symbol interpretation: delay of s1&s2 Representing the Delay between switches s1 and s2, total time represents the total time from the source of packet initiation to the termination of the packet, delay c&s1 ，Delay c&s2 Representing the delay between the controller and the switches s1, s2, respectively.
(4) Dithering:
symbol interpretation: the sample interval is [ i-1, i ], the Delay of this interval is Delay (i-1, i), and θ is the noise contribution.
When the calculation power resources of the data center are subjected to optimization constraint in the present stage, only index factors such as time delay, bandwidth, packet loss rate and the like are generally considered, the calculation capacity of the data center and the storage capacity of the edge server are added into a flow dispatching optimization target model, the time delay, the energy consumption, the bandwidth and the like are used as performance indexes, different target functions are established, cloud edge node calculation power, differential delay, storage capacity and the like are used as constraint targets, a mathematical model is established, reasonable distribution of network resources and calculation resources is ensured, and the optimization of calculation task time delay, system energy consumption and the like under constraint conditions such as limited resources, task priority and the like is ensured.
The multi-objective optimization problem consists of decision variable parameters, objective functions and constraint conditions, and the optimization objective functions of the whole network are divided into minimizing service delay, maximizing quality of experience of a data center, minimizing energy consumption of an edge server, minimizing a link packet loss rate, maximizing network path bandwidth and minimizing link jitter rate:
minf 1 ＝T average (1)
maxＱoE(2)
min∑ l g l (x l )＝C (3)
D T x uv ≤d (13)
J T x uv ≤j (14)
∏LossRate(i)≤L (15)
the explanation of the above formulas (1) to (15) is as follows:
1) Equation (1) represents minimizing the average latency of user services, including the latency of uploading to the cloud computing data center, expressed as
De m Representing the latency of the application m upload to the cloud data center,
representing a total number of requests of application m uploaded to the cloud computing data center on the ith edge server, further including Delay s1&s2 Representing the delay between switches s1 and s 2. In the formula (1), f 1 Represents a time delay function, minf 1 Representing a minimized time delay objective function, T average Representing the total average delay of the network flow schedule.
2) Equation (2) represents and maximizes the quality of experience Qoe of the data center, which is related to computational power. In the formula (2), qoE represents a quality of service index function of the data center network service, maxQoE represents maximized quality of service QoE, and maxQoE is an artificially set evaluation standard of network quality of experience;
3) Equation (3) represents minimizing edge server network performance energy consumption, which is related to edge server storage capacity, in equation (3), Σ l g l (x l ) Representing the energy consumption compliance function on the link, min l g l (x l ) Representing minimizing link energy consumption, C represents the total power consumption in the entire edge network.
4) Equation (4) represents minimizing the link packet loss rate and list represents the link set. In the formula (4), L list A packet loss rate function representing a link list,
representing minimizing the packet loss rate for the sampling intervals i through n in the link,
5) Equation (5) maximizes the bandwidth utilization of the network path and UB represents the link bandwidth. In the formula (5), U list The bandwidth calculation objective function representing the link List,
representing maximized network path l e UB represents a bandwidth calculation formula, l e Representing a certain link;
6) Equation (6) represents the link packet loss rate. In the formula (6), J list The jitter rate calculation representing the link List,
jitter rate representing minimized link List, jitter (i) is the Jitter rate calculation formula, l e Representing a certain link.
7) Equation (7) represents the maximum computational power limit of the data center versus the energy consumption of the time t node. In the formula (7) of the present invention,
representing each T time node in time TAn average energy consumption calculation function of N calculation nodes, wherein Q is a data center calculation capability limit, < ->
The energy consumption of the time T node is T, the total time is T, and N is the calculation node.
8) Equation (8) represents a constraint of limitation of the maximum storage capacity of the data center; in the formula (8), the expression "a",
representing the total storage capacity calculation within the storage K node, and (2)>
Data center computing task representing time node t, C n Representing storage capacity limitations, c k Representing a capacity representation of k storage nodes.
9) Equation (9) represents the maximum limiting constraint of the network bandwidth of the data center. In the formula (9) of the present invention,
representing the bandwidth requirement between the data center node a and node b tasks at time node t,
representing the maximum limit of the network bandwidth of n nodes.
10 Equation (10) (11) (12) represents a minimized edge server network performance energy consumption penalty associated with conditions such as edge server storage capacity and network node performance. In the formula (10) of the present invention,
is sigma in the energy consumption compliant function (3) l g l (x l ) Independent variable of x l For storage capacity, +.>
Representing k service node resource utilization of i and j nodes under l task setThe utilization rate, i and j, represents an edge server, l represents a certain task, E is a task set, and k is a service node.
In the formula (11), the color of the sample is,
represents x l Limiting the maximum storage capacity to c l 。
In the formula (12) of the present invention,
performance assessment representation representing network nodes of edge servers i and j under k, d ij Is a performance evaluation value.
11 Equation (13) represents D representing the boundary of the network application to the delay constraint, D representing the column vector of the delays of the sides, x uv Representing a link. In formula (13), D T x uv D is less than or equal to D, D represents the boundary of the network application on the time delay constraint, D represents the column vector of the time delay of each side, and x uv Representing a link.
12 Equation (14) link jitter is below the constraint limit, J represents the boundary of the network application to jitter constraint, and J represents the column vector of jitter on each side. In formula (14), J T x uv And J is less than or equal to the constraint limit, J represents the boundary of the jitter constraint applied by the network, and J represents the column vector of the jitter of each side.
13 Equation (15) L represents the maximum requirement for packet loss rate. In the formula (15), pi LossRate (i) is less than or equal to L, lossRate (i) represents the sum of packet loss rates, and L represents the maximum requirement on the packet loss rate.
As shown in fig. 1, a method for traffic scheduling of SDN network resources in a cross-domain data center according to an embodiment of the present invention includes the following steps:
s1, adding computing capacity of a cross-domain data center and network performance energy consumption of an edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
the calculation process of the cross-domain data center is as follows:
s010, establishing a first constraint condition according to performance parameters of the cross-domain data center;
the performance parameters of the cross-domain data center comprise buffer space, network link bandwidth and the like. The first constraint established includes equation (7), equation (8) and equation (9), namely:
where Q is a computational power limit, C n Is the maximum storage capacity of the device,
maximum limit of network bandwidth, +.>
Energy consumption of time t node.
S011, calculating the computing capacity of the cross-domain data center by utilizing a multi-objective optimization model added with a first constraint condition and combining an objective function established for the computing capacity of the cross-domain data center;
the multi-objective optimization model is explained as follows:
when the computing power resources of the data center are constrained at the present stage, the data center is crossed by the scheduling of a pure network, and the factors such as time delay, bandwidth and the like are generally considered. Comprehensively considering the computing power resources, network performance and the like of the data center and the edge server. SDN architecture is arranged between the score data center and the total schedule, and an openflow protocol needs to be expanded; the individual data centers are connected with the total scheduling through the traditional Internet, the special situation needs to be considered in data acquisition, and the computing capacity of the cross-domain data center and the network performance energy consumption storage capacity of the edge server are added into a multi-objective optimization model NSGA-II algorithm on the basis of traditional computing power resource regulation.
Wherein, the objective function set for the computing power of the cross-domain data center is formula (2);
in the algorithm implementation process of the multi-objective optimization model added with the first constraint condition, the computing capacity of the cross-domain data center can be calculated through the formula (2) and the first constraint condition.
The calculation process of the optimal network performance energy consumption of the edge server is as follows:
s020, establishing a second constraint condition according to the performance parameters of the edge server;
the performance parameters of the edge server include the capacity and network performance, and the second constraint conditions established include formula (10), formula (11) and formula (12) are:
s021, calculating the optimal network performance energy consumption of the edge server by combining the energy consumption compliance function under the second constraint condition with an objective function corresponding to the network performance energy consumption of the edge server.
Wherein, the objective function established by the network performance energy consumption of the edge server is formula (3);
the energy consumption obeying function is g l (x l )＝μ l x l Alpha, wherein mu l Parameter related to computing power network node equipment of alpha is edge serverNumber, c l D for storage capacity ij In the implementation process of the algorithm, the optimal energy consumption solution of the edge server can be obtained by calculating according to a formula (3) for performance evaluation of the network node.
The specific explanation of the multi-objective optimization model NSGA-II algorithm is as follows:
the multi-objective optimization model NSGA-II algorithm is a rapid non-dominant multi-objective optimization algorithm with elite retention strategy, is a multi-objective optimization algorithm based on Pareto optimal solution, reduces the complexity of a non-inferior sorting genetic algorithm, has high operation speed, has good convergence of solution sets, introduces a crowding degree and crowding degree comparison operator, overcomes the defect that the NSGA algorithm needs to manually specify shared parameters, and takes the crowding degree as a comparison criterion among individuals in a population, so that the individuals in the quasi-Pareto domain can be uniformly expanded to the whole Pareto domain, and the diversity of the population is ensured.
In the multi-objective optimization solution, some of the optimization objective functions are maximum values and some are minimum values, and for the convenience of calculation, all objective functions are converted into maximum values to be minimum values. The mathematical description of the multi-objective optimization problem herein is:
MinF(X)＝[f 1 (X),f 2 (X),…,f m (X)]
stg i (X)≤0,i＝1,2.,…k
h j (X)＝0,j＝1,2…l
wherein F (X) = [ F 1 (X),f 2 (X),…,f m (X)] T The space where the object is located is called the object space, which is the number of sub-objects, x= [ X ] 1 ,x 2 ,…,x n ] T For a given R n N-dimensional vectors on the space, namely the space where X is located is called as a decision space of the problem, h j (X) =0, j=1, 2, … l is an equality constraint, g i (X) is less than or equal to 0, i=1, 2 … k is an inequality constraint.
S2, establishing a corresponding objective function and constraint conditions for each SDN network characteristic of SDN network resources respectively, and adding the objective function and constraint conditions into a flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
wherein, SDN network characteristics of SDN network resources include: link bandwidth, packet loss rate, delay, and jitter. Specifically:
the objective function established for the link bandwidth is formula (5), the constraint condition established for the link bandwidth is formula (9), the objective function established for the packet loss rate is formula (4), the constraint condition established for the packet loss rate is formula (15), the objective function established for the delay is formula (1), the constraint condition established for the delay is formula (13), the objective function established for the jitter is formula (6), and the constraint condition established for the link bandwidth is formula (14).
S3, solving a flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on SDN network resources, as shown in FIG. 2, specifically:
the goal of the optimization is that the global network path deployment after the optimization strategy using multipath allocation is minimal relative to the network packet loss rate, transmission delay and edge server energy consumption before the optimization, and traffic will be more prone to flow onto lighter-loaded links, the network has higher admission capacity for future arriving connection requests without requiring rerouting of existing connections, the specific process of solving the flow scheduling optimization mathematical model using the non-dominant genetic algorithm with elite strategy is as follows:
s31, an initial definition parameter is SDN network topology diagram G (V, E, C), wherein V represents a switch node set, E represents a detected link set, a bandwidth set C, the number of user hosts is P, the number of edge servers is N, and the edge server capacity is x l Data center computing power limit Q, and uploading data center time delay
Number of data centers D n ，/>
Representing data center link bandwidth limitsAnd (3) the number K of controllers.
S32, initializing a population, randomly creating a parent population P with a population size of N, wherein each individual contains L chromosomes, L represents the number of candidate data streams, and each individual can be represented by the following genes:
| 1 | 2 | 3 | 4 | ... | N |
wherein the numbers in each chromosome represent the paths selected by the candidate path, the numbers and candidate paths have the following simple correspondence as shown in table 1 below.
Table 1:
| | Path | |
| 1 | |
|
| 2 | Candidate Path 2 | |
| ... | ... | |
| M | Candidate path M |
S33, evaluating each individual according to the proposed multi-constraint condition by the model, calculating the values of six objective functions of minimizing average time delay, maximizing quality of experience of the data center, minimizing network performance energy consumption on an edge server, minimizing link packet loss rate, jitter rate and maximizing bandwidth utilization of a network path through defined algorithm parameters.
S34, carrying out rapid non-dominant sorting on the population by combining the objective function, and endowing each individual with a corresponding non-dominant fitness value. According to the fitness function:
f＝Af 1 +BQoE+C∑ l g l (x l )+DL list +EU list +JH list the six optimization targets are constructed, wherein A, B, C, D, E and F are weight coefficients, different weight coefficients represent the importance of optimizing the six targets, and the weight coefficients are all set to be 1 for representing the overall balance. The smaller the objective function value, the more advantageous the individual is.
The rapid non-dominant ranking performs non-dominant ranking on each chromosome in the initial solution space according to different objective functions, and solution sets of different grades are generated according to ranking results. For example, the different grades obtained after non-dominant ranking of a part of the initial population given in step S32 may be classified as:
Rank0＝{Label2，Label3}，Rank1＝{Label1，Label3}，Rank2＝{Label3}，Rank3＝{Label5，Label7}......
s35, generating a first generation sub-population of the population by utilizing a binary tournament algorithm, wherein the first generation sub-population comprises the operations of selecting, crossing and mutating the population. The tournament method selection strategy first randomly selects k individuals from a population, where k is N/2. Then 2 individuals are taken out of the population of k individuals at a time, the best individual is selected to enter the pairing pool, and the step is repeated until the number of individuals in the pairing pool reaches a set value, here N. The adopted crossover operator and mutation operator are respectively two-point crossover and single-point mutation, and the 2 operators are utilized to operate individuals in the pairing pool to generate a child population. The crossover rate was 0.8 and the mutation rate was 0.1.
S36, combining the populations of the parent and the offspring to generate a new population, wherein the number of the combined populations is the sum of the numbers of the parent and the offspring populations.
S37, calculating non-dominant grades through a rapid non-dominant sorting algorithm again, and simultaneously calculating the crowding degree of individuals in the population; and selecting excellent individuals in the population according to the non-dominant ranking and the calculated crowding degree to form a new parent population.
S38, judging the iteration times, wherein the iteration times can be set to be T= 50,100,150, when a preset value of the code is reached, clustering the generated population, selecting one body from each class for strengthening, and then preferentially leaving the parent population as a new parent population compared with the parent population; when the iteration times of the preset value are not reached, the step S39 is executed;
s39, selecting, crossing and mutating the newly generated population by utilizing a binary tournament algorithm to generate a new child population, judging whether the maximum iteration number is reached, ending the algorithm if the maximum iteration number is reached, continuing iteration if the maximum iteration number is not reached, and jumping to S36 for execution.
S40, analyzing the importance degree of each target according to actual conditions, obtaining a final Pareto optimal solution set, and selecting a proper solution from the Pareto optimal solution set as a final optimal scheduling scheme.
Optionally, in the above technical solution, the process for acquiring the computing power of the cross-domain data center includes:
because the network traffic scheduling problem is a multi-objective optimization problem, the obtained final solution is not the only solution, but an optimal solution set for multiple choices, the importance degree of each objective needs to be analyzed according to actual conditions, and an optimal scheduling scheme is selected. The model comprehensively considers various conditions according to different user demands, when the real-time requirement on flow scheduling is high, the importance degree of a time delay function can be increased, and a scheduling scheme with the minimum time delay function value is selected so as to ensure the real-time performance of data transmission. When the services deployed in the edge network are increased, more user requests are processed at the edge nodes connected with the services, and in this case, transmission delay and delay of uploading to the cloud computing data center are not generated, so that the user can obtain lower network delay and better user experience. However, the corresponding power consumption increases, because to obtain a lower network delay, the deployed services increase, so that the number of edge servers started up increases, and the load also increases, so that the power consumption increases. Therefore, the algorithm can improve the utilization efficiency of network resources and save the flow bandwidth cost in the data center, thereby saving the cost for the related data center and service and obtaining indirect economic benefit.
In practical applications, a specific scheme is selected for deployment as needed from among these non-dominant solutions available from NSGA-II. For example, solutions with some compromises can be selected, and under the condition that the time delay of users and applications to a network is met, solutions with smaller power consumption are selected as much as possible to perform actual traffic scheduling deployment, so that the maximum benefit is obtained.
The beneficial effects of the present inventors are as follows:
the application range of the software-defined network not only comprises the fields of cloud computing and enterprise networks, but also the backbone network operators will partially apply SDN technology. Therefore, the SDN can have deep influence on the business mode of network application, can further promote industrial innovation of Internet application, and has wide and far-reaching application value and industrialization prospect. The cloud computing and edge computing fusion service is provided by focusing on a cross-domain data center with cloud edge fusion in the future, and the cloud edge fusion service mode synthesizes complementary advantages of the cloud computing and the edge computing, and has definite market development potential along with continuous deep Internet application.
In the aspect of economic benefit, the technical achievement of the project can directly provide a software-defined network deployment solution and technical support for a cross-domain data center of cloud edge fusion in the future, so that related direct economic benefit is obtained. In addition, the unified management of the software-defined network resources and the flow scheduling theory studied by the patent can improve the utilization efficiency of the network resources and save the flow bandwidth cost in the data center, so that the cost can be saved for the related data center and service, and the indirect economic benefit is obtained.
In the above embodiments, although steps S1, S2, etc. are numbered, only specific embodiments are given herein, and those skilled in the art may adjust the execution sequence of S1, S2, etc. according to the actual situation, which is also within the scope of the present invention, and it is understood that some embodiments may include some or all of the above embodiments.
As shown in fig. 3, a system 200 for traffic scheduling of SDN network resources in a cross-domain data center according to an embodiment of the present invention includes a first adding and establishing module 210, a second adding and establishing module 220, and a solving module 230;
the first addition setup module 210 is configured to: adding the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
the second addition setup module 220 is configured to: establishing a corresponding objective function and constraint conditions for each SDN network characteristic of SDN network resources respectively, and adding the objective function and constraint conditions into a flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
the solving module 230 is configured to: solving a flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on SDN network resources.
Optionally, in the above technical solution, the device further includes a first acquisition module, where the first acquisition module is configured to:
establishing a first constraint condition according to the performance parameters of the cross-domain data center;
and calculating the computing capacity of the cross-domain data center by utilizing a multi-objective optimization model added with the first constraint condition and combining an objective function established for the computing capacity of the cross-domain data center.
Optionally, in the above technical solution, the device further includes a second acquisition module, where the second acquisition module is configured to:
establishing a second constraint condition according to the performance parameters of the edge server;
and calculating the optimal network performance energy consumption of the edge server by using the energy consumption obeying function under the second constraint condition and combining an objective function corresponding to the network performance energy consumption of the edge server.
Optionally, in the above technical solution, the objective function of the storage capability of the edge server is: optimal network performance energy consumption of edge servers.
Optionally, in the above technical solution, SDN network characteristics of SDN network resources include: link bandwidth, packet loss rate, delay, and jitter.
The steps for implementing corresponding functions by using the parameters and the unit modules in the system for flow scheduling of the SDN network resource of the cross-domain data center according to the present invention may refer to the parameters and the steps in the embodiment of the method for flow scheduling of the SDN network resource of the cross-domain data center, which are not described herein.
The electronic device of the embodiment of the invention comprises a memory, a processor and a program stored on the memory and running on the processor, wherein the processor realizes the steps of the method for dispatching the flow of the SDN network resource of the cross-domain data center implemented by any one of the above steps when executing the program.
The electronic device may be a computer, a mobile phone, or the like, and the program is corresponding to computer software or mobile phone APP, and each parameter and step in the above electronic device according to the present invention may refer to each parameter and step in the above embodiment of a method for traffic scheduling of an SDN network resource of a cross-domain data center, which is not described herein.
Those skilled in the art will appreciate that the present invention may be implemented as a system, method, or computer program product.
Accordingly, the present disclosure may be embodied in the following forms, namely: either entirely hardware, entirely software (including firmware, resident software, micro-code, etc.), or entirely software, or a combination of hardware and software, referred to herein generally as a "circuit," module "or" system. Furthermore, in some embodiments, the invention may also be embodied in the form of a computer program product in one or more computer-readable media, which contain computer-readable program code.
Any combination of one or more computer readable media may be employed. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium can be, for example, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of the computer-readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
While embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the invention.
Claims (10)
1. A method for traffic scheduling of SDN network resources in a cross-domain data center, comprising:
adding the computing capacity of a cross-domain data center and the network performance energy consumption of an edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
establishing a corresponding objective function and constraint conditions for each SDN network feature of SDN network resources respectively, and adding the objective function and constraint conditions into the flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
and solving the flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on the SDN network resources.
2. The method for traffic scheduling of SDN network resources of a cross-domain data center of claim 1, wherein the process of obtaining computing power of the cross-domain data center comprises:
establishing a first constraint condition according to the performance parameters of the cross-domain data center;
and calculating the computing capacity of the cross-domain data center by utilizing a multi-objective optimization model added with the first constraint condition and combining an objective function established for the computing capacity of the cross-domain data center.
3. The method for traffic scheduling of SDN network resources of a cross-domain data center of claim 2, wherein the process of obtaining the optimal network performance energy consumption of the edge server includes:
establishing a second constraint condition according to the performance parameters of the edge server;
and calculating the optimal network performance energy consumption of the edge server by utilizing the energy consumption obeying function under the second constraint condition and combining an objective function corresponding to the network performance energy consumption of the edge server.
4. A method for traffic scheduling of SDN network resources of a cross-domain data center according to claim 3, wherein an objective function of network performance energy consumption of the edge server is: and the optimal network performance energy consumption of the edge server.
5. A method of traffic scheduling of SDN network resources of a cross-domain data center according to any of claims 1 to 4, characterized in that SDN network characteristics of the SDN network resources comprise: link bandwidth, packet loss rate, delay, and jitter.
6. The system for flow scheduling of SDN network resources of the cross-domain data center is characterized by comprising a first adding and establishing module, a second adding and establishing module and a solving module;
the first adding and establishing module is used for: adding the computing capacity of a cross-domain data center and the network performance energy consumption of an edge server into a multi-objective optimization model NSGA-II algorithm, and respectively establishing corresponding objective functions for the computing capacity of the cross-domain data center and the network performance energy consumption of the edge server to obtain a flow dispatching optimization objective model;
the second adding and establishing module is used for: establishing a corresponding objective function and constraint conditions for each SDN network feature of SDN network resources respectively, and adding the objective function and constraint conditions into the flow scheduling optimization objective model to obtain a flow scheduling optimization mathematical model;
the solving module is used for: and solving the flow scheduling optimization mathematical model by adopting a non-dominant genetic algorithm with elite strategy to obtain a flow scheduling scheme for performing flow scheduling on the SDN network resources.
7. The system for traffic scheduling of SDN network resources of a cross-domain data center of claim 6, further comprising a first acquisition module configured to:
establishing a first constraint condition according to the performance parameters of the cross-domain data center;
and calculating the computing capacity of the cross-domain data center by utilizing a multi-objective optimization model added with the first constraint condition and combining an objective function established for the computing capacity of the cross-domain data center.
8. The system for traffic scheduling of SDN network resources of a cross-domain data center of claim 7, further comprising a second acquisition module configured to:
establishing a second constraint condition according to the performance parameters of the edge server;
and calculating the optimal network performance energy consumption of the edge server by utilizing the energy consumption obeying function under the second constraint condition and combining an objective function corresponding to the network performance energy consumption of the edge server.
9. The system for traffic scheduling of SDN network resources of a cross-domain data center of claim 8, wherein an objective function of network performance energy consumption of the edge server is: and the optimal network performance energy consumption of the edge server.
10. A system for traffic scheduling of SDN network resources of a cross-domain data center according to any of claims 6 to 9, characterized in that SDN network characteristics of the SDN network resources comprise: link bandwidth, packet loss rate, delay, and jitter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126810.0A CN116208567A (en) | 2023-02-15 | 2023-02-15 | Method and system for flow scheduling of SDN network resources of cross-domain data center |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310126810.0A CN116208567A (en) | 2023-02-15 | 2023-02-15 | Method and system for flow scheduling of SDN network resources of cross-domain data center |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116208567A true CN116208567A (en) | 2023-06-02 |
Family
ID=86515558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310126810.0A Pending CN116208567A (en) | 2023-02-15 | 2023-02-15 | Method and system for flow scheduling of SDN network resources of cross-domain data center |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116208567A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116708446A (en) * | 2023-08-03 | 2023-09-05 | 山东省计算中心（国家超级计算济南中心） | Network performance comprehensive weight decision-based computing network scheduling service method and system |
| CN117201536A (en) * | 2023-09-08 | 2023-12-08 | 中国矿业大学 | SDN cross-domain internet of things private network sensing equipment management and control computing power resource allocation system |
-
2023
- 2023-02-15 CN CN202310126810.0A patent/CN116208567A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116708446A (en) * | 2023-08-03 | 2023-09-05 | 山东省计算中心（国家超级计算济南中心） | Network performance comprehensive weight decision-based computing network scheduling service method and system |
| CN116708446B (en) * | 2023-08-03 | 2023-11-21 | 山东省计算中心（国家超级计算济南中心） | Network performance comprehensive weight decision-based computing network scheduling service method and system |
| CN117201536A (en) * | 2023-09-08 | 2023-12-08 | 中国矿业大学 | SDN cross-domain internet of things private network sensing equipment management and control computing power resource allocation system |
| CN117201536B (en) * | 2023-09-08 | 2024-05-17 | 中国矿业大学 | SDN cross-domain internet of things private network sensing equipment management and control computing power resource allocation system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109818865B (en) | SDN enhanced path boxing device and method | |
| CN116208567A (en) | Method and system for flow scheduling of SDN network resources of cross-domain data center | |
| Sun et al. | Autonomous resource slicing for virtualized vehicular networks with D2D communications based on deep reinforcement learning | |
| CN111182570B (en) | User association and edge computing unloading method for improving utility of operator | |
| CN114338504B (en) | Micro-service deployment and routing method based on network edge system | |
| CN109151077B (en) | Calculation unloading method based on target guidance | |
| CN108023840B (en) | OVS network flow acceleration optimization method and OVS network flow acceleration optimization system | |
| Younus et al. | Optimizing the lifetime of software defined wireless sensor network via reinforcement learning | |
| Mollahasani et al. | Dynamic CU-DU selection for resource allocation in O-RAN using actor-critic learning | |
| EP4024212B1 (en) | Method for scheduling inference workloads on edge network resources | |
| Wei et al. | Joint participant selection and learning scheduling for multi-model federated edge learning | |
| Wang et al. | Multi-objective joint optimization of communication-computation-caching resources in mobile edge computing | |
| Tennakoon et al. | Q-learning approach for load-balancing in software defined networks | |
| CN114615705B (en) | Single-user resource allocation strategy method based on 5G network | |
| CN114785692A (en) | Virtual power plant aggregation regulation and control communication network flow balancing method and device | |
| Chu et al. | Reinforcement learning based multi-access control with energy harvesting | |
| Ren et al. | Feat: Towards fast environment-adaptive task offloading and power allocation in mec | |
| Lu et al. | Resource-efficient distributed deep neural networks empowered by intelligent software-defined networking | |
| Abkenar et al. | Transaction throughput maximization under delay and energy constraints in fog-IoT networks | |
| Zhang et al. | Supervised learning based resource allocation with network slicing | |
| López et al. | Priority flow admission and routing in sdn: Exact and heuristic approaches | |
| An et al. | EVOLUTIONARY ALGORITHM FOR TASK OFFLOADING IN VEHICULAR FOG COMPUTING | |
| Li et al. | Road Side Unit-Assisted Learning-Based Partial Task Offloading for Vehicular Edge Computing System | |
| Boonyopakorn et al. | A hybrid GA and tabu search approach to find optimal node placement in IP networks | |
| CN111542078A (en) | Core network control plane elastic resource allocation method in NFV environment |
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 |