CN112910698A - CDN coverage scheme adjusting method, device and equipment - Google Patents
CDN coverage scheme adjusting method, device and equipment Download PDFInfo
- Publication number
- CN112910698A CN112910698A CN202110107636.6A CN202110107636A CN112910698A CN 112910698 A CN112910698 A CN 112910698A CN 202110107636 A CN202110107636 A CN 202110107636A CN 112910698 A CN112910698 A CN 112910698A
- Authority
- CN
- China
- Prior art keywords
- coverage
- initial
- scheme
- population
- pop
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Genetics & Genomics (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method, a device and equipment for adjusting a CDN coverage scheme, wherein the method comprises the following steps: initializing an initial coverage scheme of the CDN; constructing at least two initial populations respectively comprising one initial coverage scheme; respectively establishing a process/thread/coroutine for each initial population; in each process/thread/coroutine, based on the objective of reducing the coverage cost, respectively performing an evolution operation on an initial coverage scheme in an initial population to obtain a child population corresponding to each initial population, wherein the child population comprises a candidate coverage scheme obtained by performing the evolution operation; and selecting a candidate coverage scheme with the lowest coverage cost from all the candidate coverage schemes as a target coverage scheme of the CDN. According to the technical scheme, the coverage cost of the CDN coverage scheme can be saved, and the adjustment difficulty and labor cost of the coverage scheme are reduced.
Description
Technical Field
The invention relates to the technical field of internet, in particular to a method, a device and equipment for adjusting a CDN coverage scheme.
Background
In the technology of a Content Delivery Network (CDN), a user can obtain required Content nearby by setting up a node or a service cluster between a service provider and a consumer, so that Network congestion is relieved and the response speed of the user to access a website is increased.
Further, using a set of service clusters to provide services to customers is also referred to as the customer's overlay resources. One customer needs to be served by a group of service clusters, while one service cluster can also serve multiple customers. Because the charging coefficients of the nodes/internet access points (POPs) to which the service cluster belongs are different, when a plurality of candidates of one client are covered, the nodes/POPs with low price are selected to provide services, and the purpose of saving cost can be achieved.
However, due to the many-to-many relationship between the service clusters and the clients, it becomes difficult for the clients to select a service cluster providing a service among the plurality of service clusters to achieve a state of optimal cost by selecting an appropriate service cluster group to cover different clients. Generally, the following methods are used to optimize the combination to save the cost:
1) a human experience is relied upon to manually attempt to select among different coverage combinations to save costs. This approach is inefficient and, with limited effort and time, only few attempts can be made;
2) the related search algorithm is used for trying, and as the problem belongs to the category of non-deterministic Polynomial (NP) problems, the complexity of common exhaustive search time is high, and the purpose of realizing an optimal solution by designing a good heuristic function is difficult to achieve by a general heuristic method.
Disclosure of Invention
The application aims to provide a method, a device and equipment for adjusting a CDN coverage scheme, so that the coverage cost of the CDN coverage scheme can be saved, and the difficulty in adjusting the coverage scheme and the labor cost can be reduced.
In order to achieve the above object, an aspect of the present application provides a method for adjusting a CDN coverage scheme, where the method includes:
initializing an initial coverage scheme of the CDN;
constructing at least two initial populations respectively comprising one initial coverage scheme;
respectively establishing a process/thread/coroutine for each initial population;
in each process/thread/coroutine, based on the objective of reducing the coverage cost, respectively performing an evolution operation on an initial coverage scheme in an initial population to obtain a child population corresponding to each initial population, wherein the child population comprises a candidate coverage scheme obtained by performing the evolution operation;
and selecting a candidate coverage scheme with the lowest coverage cost from all the candidate coverage schemes as a target coverage scheme of the CDN.
In order to achieve the above object, another aspect of the present application further provides an adjusting device for a CDN coverage scheme, where the adjusting device for the CDN coverage scheme includes a processor and a memory, where the memory is used to store a computer program, and when the computer program is executed by the processor, the adjusting method for the CDN coverage scheme is implemented.
In order to achieve the above object, another aspect of the present application further provides an adjusting apparatus for a CDN coverage scheme, where the apparatus includes:
the coverage scheme initialization module is used for initializing an initial coverage scheme of the CDN;
the initial population building module is used for building at least two initial populations respectively comprising one initial coverage scheme;
the parallel module is used for respectively establishing a process/thread/coroutine for each initial population;
a child population evolution module, configured to perform, in each process/thread/coroutine, an evolution operation on an initial coverage scheme in one of the initial populations based on a target for reducing coverage cost, so as to obtain a child population corresponding to each of the initial populations, where the child population includes a candidate coverage scheme obtained by performing the evolution operation;
a target coverage scheme determining module, configured to select, from all the candidate coverage schemes, a candidate coverage scheme with the lowest coverage cost as a target coverage scheme of the CDN.
In order to achieve the above object, another aspect of the present application further provides an adjusting device for a CDN coverage scheme, where the adjusting device for the CDN coverage scheme includes a processor and a memory, where the memory is used to store a computer program, and when the computer program is executed by the processor, the adjusting method for the CDN coverage scheme is implemented.
It can be seen from above that, the technical scheme that this application provided, concrete can realize following technological effect:
(1) automation: the adjusting method adopted by the invention combines a random algorithm and a heuristic search method to adjust the coverage cost of the coverage scheme, generates an adjusted target coverage scheme by inputting an initial coverage scheme and carrying out finite iterations, and provides the adjusted target coverage scheme for a back-end scheduling system to use, thereby realizing the end-to-end coverage cost adjusting process. In the whole tuning process, the adjustment scheme combines human experience, searching is carried out in a heuristic mode, and a random exploration mode is integrated, so that the limitation of the human experience is avoided, the adjustment cost of the CDN coverage scheme is saved, and the adjustment difficulty and the labor cost of the coverage scheme are reduced.
(2) Fast convergence: because the coverage cost optimization belongs to the NP problem category, a common exhaustive search mode cannot be solved in polynomial time, and some heuristic search algorithms rely on the precise design of heuristic functions, so that the heuristic functions are difficult to take efficiency and exploratory into account. The tuning method adopted by the invention adopts a parallel genetic algorithm, can simultaneously carry out multi-path search in different modes and different angles, and meets the solving efficiency and the exploratory appeal of the problem.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic step diagram of a method for adjusting a CDN coverage scheme according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of an adjusting apparatus for CDN coverage planning in an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an adjusting device for a CDN coverage scheme in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
Referring to fig. 1, in an embodiment of the present application, the method for adjusting a CDN coverage scheme may include the following steps.
And S110, initializing an initial coverage scheme of the CDN.
Wherein, the overlay, also called overlay resource, is used to represent the corresponding relationship between the client and the service cluster providing service for the client. In particular, the overlay scheme may be represented as a mapping between a domain name VIEW (VIEW) and a POP.
Therein, the VIEW (VIEW) may be used to represent operator + regions, such as telecom-fuzhou, mobile-mansion, etc. The domain name VIEW (VIEW) may then be used to represent the operator that the customer domain name is using in a certain area.
The POP is called point of presence, also called a node/internet access point/network access point to which the service cluster belongs.
Further, the mapping between domain name VIEW (VIEW) and POP indicates that the client domain name is in a certain area using a certain network entry point provided by the operator.
In the present embodiment, the method is based on Genetic Algorithm (GA), which combines random algorithm and heuristic search, and adjusts the initial coverage scheme to a target coverage scheme with lower coverage cost based on the goal of reducing coverage cost.
The genetic algorithm is a search algorithm for calculating the optimal in computational mathematics, and is one of evolutionary algorithms. Genetic algorithms are typically implemented by computer simulation. For an optimization problem, a certain number of candidate solutions (called individuals) can be abstractly represented as chromosomes, allowing the population to evolve towards better solutions.
In this embodiment, the initialization operation is to generate an initial coverage scheme, construct an initial population of a genetic algorithm based on the initial coverage scheme, and further, perform an evolution operation on the initial population to make the initial population evolve toward a better solution.
Further, the initial coverage scheme may be initialized in various ways, such as randomly generated by a genetic algorithm; as another example, the coverage scheme used by the current date or history is used as the initial coverage scheme.
And S120, constructing at least two initial populations respectively comprising the initial coverage scheme.
In this embodiment, the initial coverage scheme may be used to construct an initial population, and further, a plurality of coverage schemes for adjustment may be obtained by performing an evolution operation on the initial population.
In this embodiment, the number of the initial population is not limited. Furthermore, when the number of the initial populations is multiple, the parallel evolution operation can be performed on the multiple initial populations, and the multi-path search can be performed in different modes and at different angles at the same time, so that the calculation efficiency is improved.
In this embodiment, each initial population may adopt the same initial coverage scheme; of course, the initialization of the initial coverage schemes in each initial population may be performed in different manners.
And S130, respectively establishing a process/thread/coroutine for each initial population.
In this embodiment, a plurality of processes/threads/coroutines may be newly created and used for the evolutionary operation of each initial population, so as to implement parallel multi-path search of each initial population in different manners and at different angles, and the child populations corresponding to each initial population have different candidate coverage schemes, thereby satisfying the problem solving efficiency and the exploratory appeal.
And S140, in each process/thread/coroutine, based on the objective of reducing the coverage cost, respectively performing an evolution operation on an initial coverage scheme in one initial population to obtain a child population corresponding to each initial population, wherein the child population comprises candidate coverage schemes obtained by performing the evolution operation.
In the present embodiment, the coverage cost may include the following: and calculating the POP cost brought by bandwidth allocation, the internal consumption cost brought by coverage scheme adjustment and the like according to the POP charging mode, the charging coefficient and the like.
Illustratively, the coverage cost FinalCost may be expressed as:
wherein CostPopi is the POP cost of the ith coverage scheme; CostModifyi is the cost internal consumption brought by the adjustment of the ith coverage scheme; n is the number of coverage schemes; α is the cost weight.
In this embodiment, with reference to the genetic algorithm, the initial coverage scheme may be used as the initial population based on the objective of reducing the coverage cost, and the evolutionary operation is iteratively performed to obtain a progeny population composed of the progeny coverage schemes; and when the preset evolution termination condition is met, determining the child coverage scheme in the child population of the last generation as a candidate coverage scheme.
The preset evolution termination condition may include: and when the specified iteration times are reached, the coverage cost is lower than a preset cost threshold value. The preset cost threshold can be determined according to the cost optimization degree of the relative initial coverage scheme and the difference between the ideal POP planning line.
Further, the operation of evolving may include: selection operations and recombination operations.
In one embodiment, when performing an evolution operation on the initial coverage solution in the initial population to obtain a child population composed of candidate coverage solutions, the following steps may be specifically performed:
and S1, taking the initial population as the current generation population.
And S2, selecting the coverage schemes in the current generation population so as to select a preset number of coverage schemes from a plurality of coverage schemes.
In the present embodiment, the coverage cost of each coverage scheme in the current generation population is evaluated; and selecting a preset number of coverage schemes from the current generation population according to a preset selection mode based on the coverage cost.
Further, the preset selection mode may be one of the following:
1. roulette: calculating the probability of each coverage scheme entering the next generation population according to the coverage cost of each coverage scheme, and randomly selecting M coverage schemes based on uniform distribution, wherein the probability of each coverage scheme entering the next generation populationiReference may be made to the following ways:
wherein, CurCostpopiAnd N is the number of the coverage schemes in the current generation population.
2. Best parent preservation strategy: the coverage scheme with the lowest coverage cost in the current generation population does not participate in the subsequent recombination operation and is only used for replacing the coverage scheme with the highest coverage cost in the child generation population, and the remaining M-1 coverage is selected by roulette.
3. TOP K strategy: and sorting the coverage schemes in the current generation population according to the coverage cost, and selecting M parts with the lowest coverage cost to enter the offspring population.
Further, it should be noted that, when the current generation population is the initial population, there is only one coverage solution of the initial coverage solution in the current generation population, and the initial coverage solution can be directly selected to enter the next generation population. When the current generation population is a child population, the child population has a plurality of coverage schemes, and then selection is performed according to one of the three selection modes.
S3, based on the goal of reducing the coverage cost, the selected coverage scheme is recombined to be used as the coverage scheme in the next generation child population.
In the present embodiment, the reorganization operation reallocates resources such as POP to a certain proportion of domain name views without reducing the quality of service.
In this embodiment, based on the objective of reducing the coverage cost, an adjustment operation on the mapping relationship between the domain name view and the POP may be performed on the selected coverage scheme to obtain a reorganized coverage scheme, so as to complete the reorganization operation.
In this example, the recombination operation is similar to gene crossover and gene variation in the genetic algorithm, and aims to obtain more coverage schemes through the recombination operation. In this embodiment, through the reassembly operation, the randomness of the coverage scheme may be increased to increase the search space of the target coverage scheme, and avoid a local optimization trap that is common in the optimization algorithm in the search process.
In a specific embodiment, step S3 can be subdivided into the following sub-steps:
and S31, selecting POP with cost exceeding the preset cost line from the selected overlay schemes as the POP to be adjusted.
In this embodiment, a POP that runs beyond the cost line may be selected to be placed in the POP set to be adjusted.
And S32, selecting the domain name view to be overloaded from at least two domain name views mapped by the POP to be adjusted.
In the embodiment, the probability that the domain name view is selected to participate in overload can be calculated according to the priority of each domain name view under the POP to be adjusted; and determining the view of the domain name to be overloaded by adopting a roulette mode based on the probability of overload participation.
The overload process refers to adjusting the domain name view from the original POP node to map other POP nodes.
Further, the higher the priority of the domain name view, the higher the probability of participating in overload. Wherein, the ith domain name view dviProbability of participation in overload CP (dv)i) Can be expressed as:
wherein, the leveldviFor the ith domain name view dviN is the number of all domain views under POP.
Further, a view of the domain name to be overloaded may be determined by roulette based on the probability of participation in the overload. Based on the probability, the domain name view to be overloaded is selected, the randomness of the adjusting method is increased, and the limitation of human experience is further avoided.
And S33, overloading the domain name view to be overloaded to the POP of the neighborhood.
In this embodiment, the neighborhood of the domain name view to be overloaded is the resource pool range of the domain name view to be overloaded, and the resource pool range specifies the optional POP resources.
In this embodiment, all candidate POPs in the neighborhood of the domain view to be overloaded can be determined; calculating a reception degree of each candidate POP based on the cost line of each candidate POP,
in the present embodiment, the reception degree RP (POP) of the ith POPi) Can be expressed as:
wherein, CurCostpopiAnd the coverage cost corresponding to the ith coverage scheme in the current generation population. N represents the number of all candidate POPs.
Further, a POP can be selected from all the candidate POPs as a target POP in a roulette manner based on the receptivity; and overloading the domain name view to be overloaded to the target POP.
Further, whether chain adjustment is needed can also be judged:
1. if the cost of the target POP does not exceed the cost line after the target POP receives the domain name view to be overloaded, the adjustment of the current coverage scheme is finished;
2. and if the cost of the target POP exceeds the cost line after the target POP receives the domain name view to be overloaded, judging whether the current adjustment times exceed a preset adjustment threshold value of chain adjustment.
And when the current adjustment times is lower than the preset adjustment time threshold, taking the target POP as the POP to be adjusted, and continuing to adjust the POP to be adjusted, namely corresponding to the steps S32 to S33.
And when the current adjustment times exceed a preset adjustment threshold, indicating that the adjustment of the current coverage scheme is finished.
And S4, when the preset evolution termination condition is not met, taking the next generation child population as the next current generation population, and continuing to execute the step S2.
S150, selecting a candidate coverage solution with the lowest coverage cost from all the candidate coverage solutions as a target coverage solution of the CDN.
In this embodiment, the coverage cost of each candidate coverage scheme may be estimated, and similarly, the coverage cost FinalCost may be expressed as:
wherein CostPop is the POP cost, and CostModify is the cost internal consumption brought by the adjustment of the coverage scheme.
Further, a candidate coverage scheme with the lowest coverage cost is selected as a target coverage scheme of the CDN, so that the coverage cost of the CDN coverage scheme is saved.
In this embodiment, the adopted adjustment method is to adjust the coverage cost of the coverage scheme by applying a random algorithm and a heuristic search method, and by inputting an initial coverage scheme and performing finite iterations, a target coverage scheme after adjustment is generated with the aim of reducing the coverage cost as a target, and is provided for a back-end scheduling system to use, so as to implement an end-to-end coverage cost adjustment process.
Firstly, in the whole tuning process, the adjusting scheme combines with human experience, and the heuristic mode of taking coverage cost reduction as a target is used for searching, so that the coverage scheme corresponding to the lowest coverage cost of population solving is facilitated.
In addition, the adjustment method is further integrated into a random exploration mode, such as randomly generating an initial coverage scheme; for another example, the selection operation and the recombination operation adopted in the evolution process are both performed in a probabilistic manner, such as a roulette manner, so that the limitation of human experience can be further avoided, the adjustment cost of the CDN coverage scheme is saved, and the adjustment difficulty and labor cost of the coverage scheme are reduced.
Secondly, the adjustment method has fast convergence: because the coverage cost optimization belongs to the NP problem category, a common exhaustive search mode cannot be solved in polynomial time, and some heuristic search algorithms rely on the precise design of heuristic functions, so that the heuristic functions are difficult to take efficiency and exploratory into account.
In addition, the tuning method adopted in the invention adopts a plurality of newly-built processes/threads/coroutines which are respectively used for the evolution operation of each initial population so as to realize the parallel multi-path search of each initial population in different modes and different angles, and the filial generation populations corresponding to each initial population have different candidate coverage schemes, so that the parallel genetic algorithm meets the solving efficiency and the exploratory appeal of the problem.
Referring to fig. 2, the present application further provides an adjusting device for a CDN coverage scheme, where the adjusting device specifically includes the following structure: a coverage scheme initialization module 210, an initial population construction module 220, a parallelization module 230, a child population evolution module 240, and a target coverage scheme determination module 250.
An overlay scheme initialization module 210, configured to initialize an initial overlay scheme of the CDN;
an initial population constructing module 220, configured to construct at least two initial populations respectively including one of the initial coverage schemes;
a parallel module 230, configured to respectively create a new process/thread/coroutine for each initial population;
a child population evolution module 240, configured to perform, in each process/thread/coroutine, an evolution operation on an initial coverage scheme in one initial population based on a target of reducing coverage cost, so as to obtain a child population corresponding to each initial population, where the child population includes a candidate coverage scheme obtained by performing the evolution operation;
a target coverage solution determining module 250, configured to select a candidate coverage solution with the lowest coverage cost from all the candidate coverage solutions as the target coverage solution of the CDN.
In one embodiment, the progeny population evolution module 240 includes:
and the evolution unit is used for iteratively executing the evolution operation by taking the initial coverage scheme as the initial population based on the target of reducing the coverage cost to obtain the filial generation population consisting of the filial generation coverage schemes.
And the candidate coverage scheme determining unit is used for determining the child coverage scheme in the child population of the last generation as the candidate coverage scheme when a preset evolution termination condition is met.
In one embodiment, the preset evolution termination condition comprises: and when the specified iteration times are reached, the coverage cost is lower than a preset cost threshold value.
In one embodiment, the evolutionary operation includes: selecting operation and recombining operation;
the evolution unit is specifically configured to perform the following steps:
s1, taking the initial population as a current generation population;
s2, selecting the coverage schemes in the current generation population to select a preset number of coverage schemes from the coverage schemes;
s3, based on the goal of reducing the coverage cost, executing recombination operation on the selected coverage scheme to use the recombined coverage scheme as the coverage scheme in the next generation filial population;
and S4, when the preset evolution termination condition is not met, taking the next generation child population as the next current generation population, and continuing to execute the step S2.
Further, step S2 is specifically configured to:
evaluating the coverage cost of each coverage scheme in the current generation population;
and selecting a preset number of coverage schemes from the current generation population according to a preset selection mode based on the coverage cost.
Further, step S3 is specifically configured to:
and based on the target of reducing the coverage cost, performing adjustment operation on the mapping relation between the domain name view and the POP on the selected coverage scheme to obtain a recombined coverage scheme so as to finish the recombination operation.
Further, the adjusting operation of the mapping relationship between the domain name and the logical node is performed on the selected overlay scheme, and specifically, the adjusting operation is used to perform the following steps:
selecting POP with cost exceeding a preset cost line from the selected coverage scheme as POP to be adjusted;
selecting a domain name view to be overloaded from at least two domain name views mapped by the POP to be adjusted;
and overloading the domain name view to be overloaded to POP of a neighborhood, wherein the neighborhood is the resource pool range of the domain name view to be overloaded.
In an embodiment, the selecting a domain name view to be overloaded from at least two domain name views mapped by the POP to be adjusted includes:
calculating the probability of the selected domain name view participating in overload according to the priority of each domain name view under the POP to be adjusted;
and determining the view of the domain name to be overloaded by adopting a roulette mode based on the probability of overload participation.
In one embodiment, the overloading the domain name view to be overloaded to a POP in a neighborhood includes:
determining all candidate POPs in the neighborhood of the domain name view to be overloaded;
calculating the receiving degree of each candidate POP based on the cost line of each candidate POP;
selecting a POP from all candidate POPs as a target POP in a roulette mode based on the receiving degree;
and overloading the domain name view to be overloaded to the target POP.
In one embodiment, after the overloading the domain name view to be overloaded to the target POP, the method further includes:
and after the target POP receives the domain name view to be overloaded, when the cost of the target POP exceeds a cost line and the current adjustment frequency is lower than a preset adjustment threshold value, taking the target POP as the POP to be adjusted, and continuously executing the adjustment of the POP to be adjusted.
Referring to fig. 3, the present application further provides an adjusting device for a CDN coverage scheme, where the adjusting device for a CDN coverage scheme includes a processor and a memory, where the memory is used to store a computer program, and when the computer program is executed by the processor, the adjusting method for the CDN coverage scheme may be implemented, for example, the following steps are executed:
initializing an initial coverage scheme of the CDN;
constructing at least two initial populations respectively comprising one initial coverage scheme;
respectively establishing a process/thread/coroutine for each initial population;
in each process/thread/coroutine, based on the objective of reducing the coverage cost, respectively performing an evolution operation on an initial coverage scheme in an initial population to obtain a child population corresponding to each initial population, wherein the child population comprises a candidate coverage scheme obtained by performing the evolution operation;
and selecting a candidate coverage scheme with the lowest coverage cost from all the candidate coverage schemes as a target coverage scheme of the CDN.
It can be seen from above that, the technical scheme that this application provided, concrete can realize following technological effect:
(1) automation: the adjusting method adopted by the invention combines a random algorithm and a heuristic search method to adjust the coverage cost of the coverage scheme, generates an adjusted target coverage scheme by inputting an initial coverage scheme and carrying out finite iterations, and provides the adjusted target coverage scheme for a back-end scheduling system to use, thereby realizing the end-to-end coverage cost adjusting process. In the whole tuning process, the tuning scheme combines human experience, searches in a heuristic mode, and is integrated with a random exploration mode, so that the limitation of the human experience is avoided, the coverage cost of the CDN coverage scheme is saved, and the tuning difficulty and the labor cost of the coverage scheme are reduced.
(2) Fast convergence: because the coverage cost optimization belongs to the NP problem category, a common exhaustive search mode cannot be solved in polynomial time, and some heuristic search algorithms rely on the precise design of heuristic functions, so that the heuristic functions are difficult to take efficiency and exploratory into account. The tuning method adopted by the invention adopts a parallel genetic algorithm, can simultaneously carry out multi-path search in different modes and different angles, and meets the solving efficiency and the exploratory appeal of the problem.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for embodiments of the system and the apparatus, reference may be made to the introduction of embodiments of the method described above in contrast to the explanation.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (12)
1. A method for adjusting a CDN coverage scheme is characterized by comprising the following steps:
initializing an initial coverage scheme of the CDN;
constructing at least two initial populations respectively comprising one initial coverage scheme;
respectively establishing a process/thread/coroutine for each initial population;
in each process/thread/coroutine, respectively performing evolution operation on an initial coverage scheme in one initial population to obtain a child population corresponding to each initial population, wherein the child population comprises a candidate coverage scheme obtained by performing the evolution operation;
and selecting a candidate coverage scheme with the lowest coverage cost from all the candidate coverage schemes as a target coverage scheme of the CDN.
2. The method of claim 1, wherein performing an evolutionary operation on the initial coverage schemes in the initial populations to obtain the child populations corresponding to each of the initial populations respectively comprises:
taking the initial coverage scheme as an initial population, and iteratively executing an evolution operation to obtain a filial generation population consisting of filial generation coverage schemes;
and when the preset evolution termination condition is met, determining the child coverage scheme in the child population of the last generation as a candidate coverage scheme.
3. The method of claim 2, wherein the preset evolution termination condition comprises: and when the specified iteration times are reached, the coverage cost is lower than a preset cost threshold value.
4. The method of claim 3, wherein the evolutionary operation comprises: selecting operation and recombining operation;
the iteratively executing an evolutionary operation by taking the initial coverage scheme as an initial population based on the objective of reducing the coverage cost to obtain a child population composed of child coverage schemes includes:
s1, taking the initial population as a current generation population;
s2, selecting the coverage schemes in the current generation population to select a preset number of coverage schemes from the coverage schemes;
s3, based on the goal of reducing the coverage cost, executing recombination operation on the selected coverage scheme to use the recombined coverage scheme as the coverage scheme in the next generation filial population;
and S4, when the preset evolution termination condition is not met, taking the next generation child population as the next current generation population, and continuing to execute the step S2.
5. The method of claim 4, step S2, comprising:
evaluating the coverage cost of each coverage scheme in the current generation population;
and selecting a preset number of coverage schemes from the current generation population according to a preset selection mode based on the coverage cost.
6. The method of claim 4, step S3, comprising:
and based on the target of reducing the coverage cost, performing adjustment operation on the mapping relation between the domain name view and the POP on the selected coverage scheme to obtain a recombined coverage scheme so as to finish the recombination operation.
7. The method according to claim 6, wherein the performing of the adjustment operation on the mapping relationship between the domain name and the logical node on the selected overlay scheme comprises:
selecting POP with cost exceeding a preset cost line from the selected coverage scheme as POP to be adjusted;
selecting a domain name view to be overloaded from at least two domain name views mapped by the POP to be adjusted;
and overloading the domain name view to be overloaded to POP of a neighborhood, wherein the neighborhood is the resource pool range of the domain name view to be overloaded.
8. The method of claim 7, wherein selecting the domain name view to be overloaded from the at least two domain name views to which the POP to be adjusted is mapped comprises:
calculating the probability of the selected domain name view participating in overload according to the priority of each domain name view under the POP to be adjusted;
and determining the view of the domain name to be overloaded by adopting a roulette mode based on the probability of overload participation.
9. The method of claim 7, wherein overloading the domain name view to be overloaded to a POP of a neighborhood comprises:
determining all candidate POPs in the neighborhood of the domain name view to be overloaded;
calculating the receiving degree of each candidate POP based on the cost line of each candidate POP;
selecting a POP from all candidate POPs as a target POP in a roulette mode based on the receiving degree;
and overloading the domain name view to be overloaded to the target POP.
10. The method of claim 9, further comprising, after the overloading the domain name view to be overloaded to the target POP:
and after the target POP receives the domain name view to be overloaded, when the cost of the target POP exceeds a cost line and the current adjustment frequency is lower than a preset adjustment threshold value, taking the target POP as the POP to be adjusted, and continuously executing the adjustment of the POP to be adjusted.
11. An apparatus for adjusting a CDN coverage scheme, the apparatus comprising:
the coverage scheme initialization module is used for initializing an initial coverage scheme of the CDN;
the initial population building module is used for building at least two initial populations respectively comprising one initial coverage scheme;
the parallel module is used for respectively establishing a process/thread/coroutine for each initial population;
a child population evolution module, configured to perform, in each process/thread/coroutine, an evolution operation on an initial coverage scheme in one of the initial populations based on a target for reducing coverage cost, so as to obtain a child population corresponding to each of the initial populations, where the child population includes a candidate coverage scheme obtained by performing the evolution operation;
a target coverage scheme determining module, configured to select, from all the candidate coverage schemes, a candidate coverage scheme with the lowest coverage cost as a target coverage scheme of the CDN.
12. An adjustment device for a CDN coverage plan, characterized in that the adjustment device for a CDN coverage plan comprises a processor and a memory, the memory being configured to store a computer program, which, when executed by the processor, implements the method according to any one of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110107636.6A CN112910698B (en) | 2021-01-27 | 2021-01-27 | CDN coverage scheme adjusting method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110107636.6A CN112910698B (en) | 2021-01-27 | 2021-01-27 | CDN coverage scheme adjusting method, device and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112910698A true CN112910698A (en) | 2021-06-04 |
| CN112910698B CN112910698B (en) | 2023-08-22 |
Family
ID=76120460
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110107636.6A Active CN112910698B (en) | 2021-01-27 | 2021-01-27 | CDN coverage scheme adjusting method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112910698B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140279765A1 (en) * | 2013-03-12 | 2014-09-18 | International Business Machines Corporation | Early generation of individuals to accelerate genetic algorithms |
| CN108712302A (en) * | 2018-04-19 | 2018-10-26 | 上海帝联网络科技有限公司 | The computational methods and device of zone bandwidth, computer-readable medium |
| CN108762927A (en) * | 2018-05-29 | 2018-11-06 | 武汉轻工大学 | The multiple target method for scheduling task of mobile cloud computing |
| CN109246032A (en) * | 2018-09-28 | 2019-01-18 | 杭州领智云画科技有限公司 | A kind of CDN dispatching method and equipment based on genetic algorithm |
| CN109992355A (en) * | 2019-01-30 | 2019-07-09 | 北京理工大学 | A kind of multiple target cloud workflow schedule method based on the non-dominant genetic algorithm of improvement |
| US20190244110A1 (en) * | 2018-02-06 | 2019-08-08 | Cognizant Technology Solutions U.S. Corporation | Enhancing Evolutionary Optimization in Uncertain Environments By Allocating Evaluations Via Multi-Armed Bandit Algorithms |
| CN111162954A (en) * | 2020-01-22 | 2020-05-15 | 厦门网宿有限公司 | Path planning method, system and central server |
| CN111475274A (en) * | 2020-04-20 | 2020-07-31 | 北京邮电大学 | Cloud collaborative multi-task scheduling method and device |
| CN111538570A (en) * | 2020-05-12 | 2020-08-14 | 广东电网有限责任公司电力调度控制中心 | VNF deployment method and device for energy conservation and QoS guarantee |
-
2021
- 2021-01-27 CN CN202110107636.6A patent/CN112910698B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140279765A1 (en) * | 2013-03-12 | 2014-09-18 | International Business Machines Corporation | Early generation of individuals to accelerate genetic algorithms |
| US20190244110A1 (en) * | 2018-02-06 | 2019-08-08 | Cognizant Technology Solutions U.S. Corporation | Enhancing Evolutionary Optimization in Uncertain Environments By Allocating Evaluations Via Multi-Armed Bandit Algorithms |
| CN108712302A (en) * | 2018-04-19 | 2018-10-26 | 上海帝联网络科技有限公司 | The computational methods and device of zone bandwidth, computer-readable medium |
| CN108762927A (en) * | 2018-05-29 | 2018-11-06 | 武汉轻工大学 | The multiple target method for scheduling task of mobile cloud computing |
| CN109246032A (en) * | 2018-09-28 | 2019-01-18 | 杭州领智云画科技有限公司 | A kind of CDN dispatching method and equipment based on genetic algorithm |
| CN109992355A (en) * | 2019-01-30 | 2019-07-09 | 北京理工大学 | A kind of multiple target cloud workflow schedule method based on the non-dominant genetic algorithm of improvement |
| CN111162954A (en) * | 2020-01-22 | 2020-05-15 | 厦门网宿有限公司 | Path planning method, system and central server |
| CN111475274A (en) * | 2020-04-20 | 2020-07-31 | 北京邮电大学 | Cloud collaborative multi-task scheduling method and device |
| CN111538570A (en) * | 2020-05-12 | 2020-08-14 | 广东电网有限责任公司电力调度控制中心 | VNF deployment method and device for energy conservation and QoS guarantee |
Non-Patent Citations (1)
| Title |
|---|
| 胡国龙等: "基于改进遗传算法的无线传感网覆盖优化", 《计算机测量与控制》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112910698B (en) | 2023-08-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Apostolopoulos et al. | Risk-aware data offloading in multi-server multi-access edge computing environment | |
| CN112512056B (en) | Multi-objective optimization calculation unloading method in mobile edge calculation network | |
| CN111182570B (en) | User association and edge computing unloading method for improving utility of operator | |
| CN111641891B (en) | Task peer-to-peer unloading method and device in multi-access edge computing system | |
| US20120140636A1 (en) | Methods and apparatus to determine network link weights | |
| Cohen et al. | Following the perturbed leader for online structured learning | |
| CN112073452B (en) | Mobile edge computing task allocation method with effective energy and limited resources | |
| CN113918240A (en) | Task unloading method and device | |
| WO2024113974A1 (en) | Computing power network routing allocation method and apparatus, electronic device, and storage medium | |
| CN114567895A (en) | Method for realizing intelligent cooperation strategy of MEC server cluster | |
| CN117519995B (en) | Star network mobile edge computing resource allocation method | |
| CN112559171B (en) | Multi-user task unloading method based on delayed acceptance in mobile edge computing environment | |
| CN117155942A (en) | Micro-service dynamic self-adaptive client load balancing method and system | |
| Lafeta et al. | MEANDS: A Many-objective Evolutionary Algorithm based on Non-dominated Decomposed Sets applied to multicast routing | |
| CN109298930A (en) | A kind of cloud workflow schedule method and device based on multiple-objection optimization | |
| CN113992677A (en) | MEC calculation unloading method for delay and energy consumption joint optimization | |
| CN116089091A (en) | Resource allocation and task unloading method based on edge calculation of Internet of things | |
| CN113672372B (en) | Multi-edge collaborative load balancing task scheduling method based on reinforcement learning | |
| Crichigno et al. | Multiobjective multicast routing algorithm | |
| CN114785692A (en) | Virtual power plant aggregation regulation and control communication network flow balancing method and device | |
| CN112910698B (en) | CDN coverage scheme adjusting method, device and equipment | |
| Wang et al. | Multi-objective joint optimization of communication-computation-caching resources in mobile edge computing | |
| CN114615705B (en) | Single-user resource allocation strategy method based on 5G network | |
| CN113543225A (en) | Method and system for safe dynamic resource allocation of power wireless private network | |
| CN114385359B (en) | Cloud edge task time sequence cooperation method for Internet of things |
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 |