CN106357739A - Two-stage composition and scheduling method specific to lot-sizing cloud service request - Google Patents
Two-stage composition and scheduling method specific to lot-sizing cloud service request Download PDFInfo
- Publication number
- CN106357739A CN106357739A CN201610747165.4A CN201610747165A CN106357739A CN 106357739 A CN106357739 A CN 106357739A CN 201610747165 A CN201610747165 A CN 201610747165A CN 106357739 A CN106357739 A CN 106357739A
- Authority
- CN
- China
- Prior art keywords
- service
- request
- cloud
- qos
- cloud service
- 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
- 238000000034 method Methods 0.000 title claims abstract description 63
- 239000000203 mixture Substances 0.000 title claims abstract description 14
- 238000004513 sizing Methods 0.000 title abstract 4
- 238000005457 optimization Methods 0.000 claims abstract description 28
- 230000008569 process Effects 0.000 claims description 33
- 230000014509 gene expression Effects 0.000 claims description 16
- 238000011156 evaluation Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 14
- 230000004044 response Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 7
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000002360 preparation method Methods 0.000 claims description 2
- 239000000463 material Substances 0.000 abstract description 2
- 238000004519 manufacturing process Methods 0.000 description 15
- 238000012545 processing Methods 0.000 description 6
- 238000011160 research Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 238000013178 mathematical model Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 241000287196 Asthenes Species 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
- H04L67/61—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources taking into account QoS or priority requirements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention involves a two-stage composition and scheduling method specific to lot-sizing cloud service. firstly, to establish a plan and a model according to the selection of QoS optimization cloud service composition patch, to formulate the description types of cloud service and cloud service composition as well as the assessment standards of quality of service (QoS); secondly, to generate strategies and make a model for the scheduling of lot-sizing service requests, and to formulate a corresponding mathematic model; finally, to maximize according to the two-stage cloud composition and scheduling which is proposed above. The invention proposes the plan and model for the selection of quality of service optimization cloud service composition path and formulates the strategies and model specific to lot-sizing service request scheduling problem, and on this basis, it is able to offer the optimal realization of cloud service composition and scheduling on the basis of two stages, able to realize an overall optimization of cloud service materials scheduling without detriment to the cloud service scheduling quality, uniting both effectively with high practicability and execution efficiency.
Description
Technical Field
The invention belongs to the technical field of digital information transmission, such as telegraph communication, and particularly relates to a two-stage combination and scheduling method for a batch cloud service request, which adopts an efficient scheme to schedule cloud services to meet the multifunctional requirement of the service request on the premise of ensuring the quality of the service combination.
Background
With the rapid development of the internet service industry and the blowout type increase of internet users, the requests and the usage amount of cloud services are greatly increased, and the situations that the cloud service management platform processes and simultaneously processes batch service requests are more and more.
The response to the bulk service request includes two aspects: (1) selecting an optimal service combination for each cloud service request submitted by an internet user; (2) in the face of batch service requests, a cloud service resource scheduling scheme with the minimum total time is provided on the premise of ensuring the quality. On the whole, in the face of the proliferation of internet users and large batch of service requests, the cloud service management platform needs to reduce the total time overhead and cost overhead while ensuring the quality of the service requests, and not only the quality but also the efficiency problem needs to be considered for the execution of the service requests; meanwhile, for the cloud service platform, how to schedule service resources is needed to be considered to improve efficiency on the premise of ensuring service quality.
Under the background, the establishment of an optimization model of the cloud service combination and a corresponding service resource scheduling model and algorithm has important theoretical significance and research value.
In the prior art, most of the optimization models for establishing the cloud service combination and the corresponding service resource scheduling models and algorithms are considered independently, and the premise of ensuring the execution quality of the service request is not considered in the process of scheduling the cloud service resources. Most of the existing cloud services take web services and manufacturing cloud services as their expressions, and when the cloud services take the two expressions, the cloud service management platform often handles batch service requests, so the cloud service platform often faces the following problems: on the premise of ensuring the quality of service combination, the cloud service is scheduled by adopting an efficient scheme to meet the multifunctional requirement of the service request. In order to solve this problem, it is necessary to consider two stages in combination, but the prior art has been less studied on this aspect, and the manufacturing cloud service, which is one expression of the cloud service, is not developed for a long time, so that the studies on these two stages are not much.
Disclosure of Invention
The technical problem solved by the present invention is that, in the prior art, most of the optimization models for establishing cloud service combinations and corresponding service resource scheduling models and algorithms are considered independently, and the premise of ensuring the execution quality of service requests is not considered in the process of scheduling cloud service resources, but most of the existing cloud services take web services and manufacturing cloud services as their expression forms, and when the cloud services are in the two expression forms, the cloud service management platform handles a lot of batch service requests, so the cloud service platform is often faced with the problems that: on the premise of ensuring the quality of service combination, the cloud service is scheduled by adopting an efficient scheme to meet the multifunctional requirement of the service request. In order to solve the problem, two stages need to be considered comprehensively, but the research on the aspect is less in the prior art, and the manufacturing cloud service which is one expression of the cloud service is not long in development time, so that the research on the two stages is not much, and an optimized two-stage combination and scheduling method for the batch cloud service request is further provided.
The technical scheme adopted by the invention is that the two-stage combination and scheduling method for the batch cloud service requests comprises the following steps:
step 1.1: the method comprises the steps that a user describes detail requirements and an overall target of the user on a service combination through an ontology description language own-s to form a service request, and the service request is submitted to a cloud service management platform to form a cloud service request sequence;
step 1.2: the cloud service management platform logically divides the service request according to the detail requirement and the whole target of the service combination of the user to form a sub-request stream, and performs service rough selection on the sub-request;
step 1.3: selecting a service composition path for the service request based on the QoS;
step 1.4: the cloud service management platform makes a scheduling plan through a service scheduling model according to a service combination path corresponding to each request in the service request sequence, and enters a scheduling preparation stage;
step 1.5: and the cloud service management platform carries out service resource scheduling.
Preferably, in step 1.2, when a user submits a large-scale complex request with diversified functions or a request with personalized customization, the cloud service management platform combines a plurality of single cloud services for the user, including the following steps:
step 1.2.1: logically dividing coarse-grained service requests into fine-grained sub-request sets T ═ T1,T2,…TmEach sub-request can use a cloud service with a single function; the sub-requests are sequentially associated and do not form a ring;
step 1.2.2: forming a sub-request T according to the corresponding information of j cloud service resource suppliers registered by the cloud service management platformiCorresponding candidate service resource set
Step 1.2.3: to be provided withRepresents a sub-request TiIs accomplished by cloud service L (i), where L (i) is sub-request TiIn a candidate cloudSelected services of the service set, i.e.The sub-request set T ═ T1,T2,…TmThe path of the complete set of cloud service combinations used can be expressed as
Preferably, in the step 1.2, the service resource interacts with the data reading layer through XML data to read and write data; the data reading layer comprises a plurality of distributed databases.
Preferably, in the step 1.3, selecting the service combination path for the service request based on the QoS includes the following steps:
step 1.3.1: establishing a cloud service combination evaluation mechanism based on QoS;
step 1.3.2: and establishing a cloud service combination model based on Qos global optimization.
Preferably, in the step 1.3.1, the cloud service combination evaluation mechanism includes a functional QoS attribute and a non-functional QoS attribute, and the functional QoS attribute includes a response timeAnd costThe non-functional QoS attributes include a reliability indexAnd availability index
Preferably, in step 1.3.1, the response time isAnd costThe reliability index is a reverse Qos attributeAnd availability indexAnd assigning the reverse Qos attribute and the forward Qos attribute for the forward Qos attribute, wherein the total Qos must meet the requirement
Where max and min are the maximum and minimum values in the same QoS attribute.
Preferably, in step 1.3.2, the set Q is assumed to represent several local QoS attributes included in the global QoS attribute, i.e. Q ═ QoS1,QoS2,…QoSi…,QoSk}；αwIs the weight ratio of each of several QoS attributes in the global QoS, andsub-request TiThe sum of the respective QoS attribute metric values using the corresponding service resources L (i) isWherein,represents a sub-request TjAttribute QoS in Using service resources L (j)wThe measurement value of (a); sub-request set T ═ T1,T2,…TmThe path of a complete set of cloud service combinations currently selected can be represented asWhen not subject to local QoS constraints, oneThe global QoS expression of the cloud service set combination can be described asWhen the current service combination is limited by local QoS,
preferably, the cloud service management platform makes a scheduling plan through a service scheduling model according to a service combination corresponding to each request in the service request sequence, where the scheduling model includes service switching time overhead and total time overhead.
Preferably, the service switching time is a matrix logn×nWherein, logijThe service switching time cost from the virtual service point i to the virtual service point j, the service switching time matrix logn×nThe switching time overhead between services provided by virtual service nodes is recorded, logijRepresenting the service switching time overhead from virtual service point i to virtual service point j.
Preferably, the cloud service S corresponding to the jth service point in the service combinationjThe time overhead spent for completing the sub-request corresponding to the service request No. i is TimeTablem×n(ii) a The cloud service management platform sets JQ as { JQ ═ JQ according to the service request sequence1,JQ2,JQ3......JQnScheduling cloud services to complete execution of respective service requests toAndcorresponding service S representing the kth virtual service PointkCompleting execution service request JQiStart time and end time of, for a service request JQi，
JQ executed if cloud service provided by kth virtual service pointiTime 0, i.e. it does not need to goThe process is performed, then
JQ executed if cloud service provided by kth virtual service pointiTime is not 0, i.e. it requires the virtual service point to provide cloud service, service request JQiThe starting time of the kth virtual service point is the sum of the time of the execution end of a virtual service point with the execution time not being 0 before the kth virtual service point and the switching time of the virtual service point to the kth virtual service point;
to be provided withThe virtual service point which represents the k-th virtual service point process and has the previous execution time not equal to 0 has the following constraint relationship:
the technical scheme of the invention is divided into three steps:
firstly, a scheme and a model for solving a first stage, namely a selection problem of a QoS-based optimal cloud service combination path, are provided, and a description mode of cloud service and cloud service combination and an evaluation standard of quality of service (QoS) are provided;
a strategy and a model for the scheduling problem facing the batch service request in the second stage are provided and solved, and a corresponding mathematical model is provided;
and thirdly, optimizing based on the two-stage cloud service combination and scheduling.
Regarding the step one, the concrete contents include:
1. quality of service (QoS) based optimal cloud service portfolio selection
The method mainly aims at the research of an optimal service combination selection method based on the service quality, analyzes the characteristics of QoS in the cloud computing environment and the specific situation of the current batch cloud service request, researches the specific steps of service combination selection, and provides a specific mathematical model of the service combination selection.
1.1 cloud service Combined architecture and implementation Process
The cloud service architecture consists of three important roles: the cloud service system comprises a cloud service management platform, a cloud service requester and a cloud service resource provider. The cloud service requester initiates a request to the cloud service management platform, the cloud platform selects the most appropriate single cloud service or cloud service combination for the service request through analysis and processing of service request information, sends request information to the corresponding cloud service resource provider, and schedules the corresponding cloud service to the corresponding service node. For the cloud platform, it is very important to have rich cloud service resources, and meanwhile, a cloud service provider also needs to register service information on the cloud service management platform in time, so that a user searches for corresponding cloud services on the cloud platform according to corresponding requirements.
The cloud service management platform needs to have abundant cloud service resources, and basic information, performance, cost overhead, response time and the like of the corresponding cloud service are provided for the cloud service management platform by a cloud resource provider. The method comprises the following steps of (1) obtaining the service which is expected to be obtained by a user of the cloud platform to be high-quality, and obtaining the service in the fastest time and paying the cost as little as possible, wherein in order to achieve the aim, the specific steps are as follows:
(1) cloud service providers need to publish cloud service information which can be provided by the cloud service providers on the internet, and the cloud service management platform registers QoS attributes such as stability, characteristics, cost and the like of cloud services in a management system according to the corresponding published information;
(2) when a user submits a cloud service request to the cloud platform, the cloud platform starts a search process, selects an optimal cloud service or cloud service combination according to the user requirement and the personalized customization requirement, and after the user submits the service request with a small scale or single request content, the cloud service management platform only needs to select a corresponding cloud service with a single function for the cloud service management platform, and certainly, the most matched cloud service can also be selected for the service request by adopting methods such as semantic matching and the like;
(3) if the submitted request is large-scale complex and requires diversified functions or has higher personalized customization degree, single services need to be combined, and meanwhile, from the perspective of the whole cloud platform, the efficient and high-quality cloud service combination model and algorithm can also save resources for service providers.
From the functional goal, the cloud platform completes the coarse-grained service requests which are more complex and have higher frequency and are proposed by the users through the service combination. Selecting a corresponding cloud service combination for a more complex cloud service request generally goes through three processes:
(1) firstly, dividing the service request logic with coarse granularity into sub-requests with fine granularity, and ensuring that each sub-request only needs to use service resources with single function;
(2) and forming a corresponding candidate service resource set according to the corresponding information of the candidate resource suppliers registered in the cloud service management platform. The efficiency of service selection can be improved by classifying the services, and the complexity of service combination is reduced;
(3) and determining an optimal path of the cloud service combination in each candidate service set.
1.2 QoS-based cloud service combination evaluation mechanism
The QoS is used as an important measurement standard in a service-oriented system, and reflects the functional and non-functional service quality level of the combination of the cloud service with single fine-grained function and the coarse-grained service. Service requests with multi-functionality requirements on the cloud platform select service combinations are also based on quality of service (QoS). The QoS attributes are divided into functional QoS attributes and non-functional QoS attributes. Common functional QoS attributes mainly include time and cost, etc. The functional QoS attributes mainly describe the time and cost overhead of providing services by a provider, which is of greater concern to users, and are reference factors which are very common and important to users in a cloud platform service-oriented architecture. Non-functional QoS attributes include reliability, availability, etc. The non-functional indicators represent whether a service matches a request, and the security of data and programs using the service, among other things.
1.2.1 service evaluation description
The cloud service evaluation standard is an index describing cloud service functionality and non-functionality, and is necessarily an aggregate comprising a plurality of indexes, and the specific use environment of each service index is different. Part of service indexes belong to basic common indexes, and are used as basic indexes in most cloud service evaluation systems, functional QoS attributes belong to the class, the other service indexes belong to runnability metrics (runtime metrics), and non-functional QoS attributes belong to the class, and the service indexes have very important reference significance for reliable, stable and high-integrity cloud service selection by users. Besides the two types of QoS service attributes, the QoS service attributes are obtained by means of historical evaluation, such as the credibility of the cloud service and the like.
1.2.2 QoS protocol
The units of different QoS attributes are different, and the way of describing cloud services is also different. The diversity of the QoS attributes is represented by the diversity of the QoS attribute acquisition modes and the diversity of the QoS metric value expression. From the aspect of the manner of obtaining the QoS attribute metric, there are two main ways:
(1) type of provisioning in advance
The metric value of the QoS attribute of this type is provided to the cloud service management platform by the cloud resource provider in advance, and belongs to a constant value preset in advance. Generally as functional QoS attributes: response time, cost overhead, etc. QoS attributes.
(2) Statistical type of historical data
The measurement value of the QoS attribute of this type is obtained by performing data statistics on the history of the cloud resource provider, and this type of statistical data can better reflect the service stability of the cloud service for a long time, so generally, the QoS attribute such as non-functional attribute reliability, security, etc. needs history data to be obtained.
The diversity of the QoS attribute metric value expression is mainly reflected in the inconsistency of the evaluation expression modes of different QoS indexes, and is mainly divided into the following three types:
(1) specific value types, such as integer and floating point type, for example, the metric values of QoS attributes such as response time and cost overhead, and the specific value type is also the most common and basic metric value expression type;
(2) in a certain interval, for example, a service has an attribute of a service time interval, which means that it provides the corresponding cloud service only in the certain time interval, that is, the attribute of the service measures the interval in the service time interval.
(3) Fuzzy type, some QoS attribute metric values use descriptors of some evaluation classes such as "general", "better", "excellent", "not qualified", etc. In addition, they are described in a class language, such as "first class", "high class", "intermediate class", etc., and these rating and class descriptions express the comparison between different metric values.
1.3 Qos global optimization-based cloud service combination model
The problem of selecting the optimal service combination for the service request with multifunctional requirements in the cloud environment is a multi-objective optimization problem in nature, and particularly under the condition of local QoS constraints, the metric value of the local QoS attribute (particularly, the non-functional QoS attribute) may be limited to be higher than a certain metric value or higher than the average level of candidate suppliers. The configurable type based on the QoS cloud service combination is also embodied, and different users have different degrees of importance on different service indexes. For example, a user of the cloud manufacturing industry may be more concerned about the production cycle and production cost of manufacturing cloud services, while a user using distributed computing has higher requirements for stability, security, and integrity. For different users, the QoS attributes which are regarded as important are different, so that the objective function can be adjusted by adjusting each weight value, and the requirements of different users are met.
Regarding the second step, the concrete contents include:
considering the scheduling policy of the cloud service management platform when batch-oriented service requests need to be processed simultaneously, in the scheduling process, the execution quality of the service requests is guaranteed by guaranteeing that the service combinations used by the service requests are the corresponding service combinations with the highest global quality of service (QoS) metric selected in the first stage, and the failure times are reduced so as to reduce the waiting time for executing the batch service requests and the overall execution time.
2.1 scheduling architecture
The cloud platform service scheduling policy needs to fully consider the busy and idle states of a cloud service provider, the overall waiting time of a whole service request queue and the service quality of the service provider, and select the optimal service combination corresponding to the service request with the multifunctional requirement, namely, to provide the cloud service combination with the highest QoS metric value for the service request.
The method comprises the following specific steps:
(1) through a first-stage optimal cloud service combination selection model and algorithm based on QoS, finding an optimal path connecting each virtual service point for each submitted service request;
(2) when the cloud service management platform processes a batch of cloud service requests with similar types, scheduling corresponding service resources according to a certain strategy to complete corresponding sub-requests;
(3) when the complex service requirement proposed by the user needs a series of complex service combinations composed of single services to complete, service resources are reasonably dispatched, and the waiting time is reduced, so that the service request of a batch can be completed in the shortest total time.
2.2 scheduling model
The main optimization goal of the service scheduling strategy for the batch cloud service requests is to reduce the total execution time of the batch service requests, and the scheduling precondition is that each service request must use the corresponding optimal service combination.
In the process of establishing the cloud service scheduling model, the time overhead problem of cloud service switching must be considered.
The cloud service scheduling problem for the batch service request can be described as follows: under the condition that cloud service resources are limited, the sequence of scheduling cloud service execution service requests is determined, so that the overall execution time of the service requests is the minimum, for the method for determining the scheduling scheme, the invention adopts the method that the total execution time of various scheduling schemes is pre-calculated by using an evolution algorithm before the batch service requests are executed, the optimal scheduling scheme is obtained, and the process of selecting the optimal scheduling scheme belongs to strategy enumeration, so that the problem that the evolution algorithm in artificial intelligence can be well solved. According to the invention, the scheduling model is designed under the optimal condition of considering the service switching time overhead and the total production time minimization, so that the optimal scheduling and optimization of the batch cloud service are achieved.
Regarding step three, the concrete contents include:
3.1 cloud service composition and scheduling architecture and Process
The cloud service management platform processes the batch service requests of the internet users and is completed by a plurality of service processes in a coordinated mode. The whole service combination and scheduling process is driven by user service request submission, and the cloud service management platform is used as a control center to select service combinations for service requests and schedule services of request tasks in the execution process, which are two main tasks of the cloud platform.
In order to ensure that each service request can obtain the optimal service combination, the specific flow of the cloud service platform scheduling system is as follows:
(1) performing data operation on the service request information in the service request sequence one by one according to service provider information of the cloud platform to obtain an optimal service path of each service request;
(2) the optimal service combination path of each service request can be transmitted to a dispatching center of the cloud service platform, and the dispatching center formulates an optimal service resource dispatching scheme according to the specific information of the service requests and the optimal service combination;
(3) and completing the scheduling of the service resources, so that the execution time of the total service request sequence can be minimized on the premise of ensuring that the optimal service combination is used for each service request.
3.2 Multi-objective optimization
Conventional research processing of cloud service requests generally only selects an optimal path for the service request, and does not consider a plurality of specific problems that should be considered when a batch of service requests wait to be processed in an actual problem. In this process, there are also a number of optimizable objectives, for which the following two phases are specifically considered:
(1) considering the cloud service composition phase, it is a main target condition to select the optimal service composition for each service request. Since the evaluation criteria for a cloud service or a combination of cloud services herein are based on QoS metric values, the main goal of the cloud service combination phase can also be considered as selecting a service combination for which the QoS metric values are optimal for the service request, and each user may have some local optimization goal for the service;
(2) considering the cloud service resource scheduling phase, for the cloud service management platform, only one service request may not be processed at any time, and it is often necessary to face that a plurality of service requests need to be processed simultaneously, so the overall goal of the scheduling phase is to minimize the total time in the cloud service resource scheduling facing the batch service requests. The scheduling of course presupposes that each service request uses the service combination for which the selection is optimal.
The invention provides an optimized combination and scheduling method aiming at batch cloud service requests, which is based on the optimization realization of two-stage cloud service combination and scheduling by respectively providing a scheme and a model of a selection problem based on a QoS optimal cloud service combination path and providing a strategy and a model of a scheduling problem facing the batch service requests, and can achieve global optimization of cloud service resource scheduling on the premise of ensuring the execution quality of the cloud service requests and effectively solve the problem of unification of the two.
The invention has the following advantages:
(1) in the background, the cloud service with single function is difficult to meet a multi-demand composite service request provided by an internet user, and in the method, a protocol is established on the basis according to common QoS attributes, and then a QoS-based cloud service combination selection algorithm is provided;
(2) the cloud service scheduling optimization method has the advantages that the practicability is strong, the execution efficiency is high, due to the extremely rapid increase of internet users, the number of cloud service requests provided by the users is greatly increased, the batch cloud service requests are more and more processed for a cloud service platform, and under the limited cloud service resources, the cloud service scheduling optimization model facing the batch service requests is established, and the scheduling model reduces the total time of the overall execution of the cloud service requests and ensures the execution quality of the service requests by using the optimal service combination selected for each service request in the first stage;
(3) the two-stage optimization model provided by the invention is essentially a multi-objective optimization problem, and an artificial intelligence evolution algorithm is often used for solving the multi-objective optimization problem, so that a new idea for solving the two-stage problem is envisaged.
Drawings
Fig. 1 is a structural diagram of a cloud platform service combination system in which the number of sub-requests is 5.
Detailed Description
The present invention is described in further detail with reference to the following examples, but the scope of the present invention is not limited thereto.
According to the cloud service scheduling and optimizing method, analysis and modeling are carried out from two stages of cloud service scheduling and optimizing, and the cloud service in batches can be better processed. The manufacturing cloud service is a cloud service with stronger representativeness, and with the full set of traditional manufacturing industry and cloud computing, the manufacturing cloud service is used more and more frequently, so the three processes are described by using one manufacturing cloud service combination:
an internet user needs to customize a batch of products with higher personalized degree through a cloud platform, and firstly submits detailed information and related requirements of the required products to a server end through a mobile end or a PC end; after receiving the service request, the server divides the service request into a plurality of sub-requests: selecting materials, welding, molding, polishing, testing and the like, and then completing selection of an optimal service combination path by the cloud platform according to the existing service resource set and the requirements of a user on time cost and the like; in the selection process, the cloud platform not only considers functional attributes such as time cost, but also considers non-functional attribute indexes such as reliability and availability, and then schedules corresponding services to complete each sub-request of the user.
The specific process is as follows:
1. a scheme and a model for solving the first stage, namely the selection problem of the optimal cloud service combination path based on the QoS are provided, and a description mode of cloud service and cloud service combination and an evaluation standard of the quality of service (QoS) are provided.
1.1 establishing a cloud service composite structure system and implementing process thereof
The structure diagram of the whole cloud platform service combination system from top to bottom is shown in fig. 1.
The sub-request set is composed of T ═ T1,T2,…TmThe sequence relationship is usually represented by a workflow diagram or a task flow diagram, and the dependency relationship and the logical relationship between the sub-requests are represented by a Directed Acyclic Graph (DAG) in the invention. In a directed acyclic graph, the set of nodes, V, is { V }1,V2,……,VmThe node set represents all sub-requests obtained after a complete service request is divided, so the node set corresponds to a sub-request set T, as shown in the first layer a of fig. 1, where the sub-request set includes sub-requests T1、T2、T3、T4And T5。
In the invention, the division principle of the sub-requests is that each sub-request uses a service with a single function.
The second layer of the architecture is the resource provisioning layer. The cloud platform provider provides cloud services to users through hardware virtualization technology and middleware taking software as a carrier. Suppose that j providers can request T for a sub-serviceiProviding a corresponding cloud service, then TiCan be represented as For representing sub-requests TiIs completed by the cloud service L (i), which is the path of the cloud service combination, wherein L (i) is the sub-request TiIn its candidate cloud service setSelected services in the pool, i.e.So the sub-request set T ═ T1,T2,…TmThe path of the complete set of cloud service combinations used can be expressed asT is shown as the second layer B of FIG. 11The corresponding candidate cloud service set includes rs11、rs12And rs13Finally select rs12To complete, T2The corresponding candidate cloud service set includes rs21、rs22And rs23Finally select rs21To complete, T3The corresponding candidate cloud service set includes rs31、rs32And rs33Finally select rs32To complete, T4The corresponding candidate cloud service set includes rs41、rs42And rs43Finally select rs42To complete, T5The corresponding candidate cloud service set includes rs51、rs52And rs53Finally select rs53To complete.
The third layer is a data reading layer, as shown in the third layer C of fig. 1, in the cloud platform, the data reading layer mainly includes a distributed database, and the cloud service resources of the second layer store data in the database of the bottom layer and also read data from the database. The data reading layer not only bears the data reading task, but also can realize data sharing and data transmission. And a uniform XML data format is adopted in the data exchange process. In addition, the data layer also defines the logic of data control flow, and the automation degree of cloud service execution is improved through the control of multiple data flows.
1.2 establishing a QoS-based cloud service combination evaluation mechanism
Table 1 is a metric expression for the four QoS attributes considered by the present invention:
table 1: metric expressions for QoS attributes
The cloud service combination user has an optimization target of local QoS under the condition of requiring the optimal global QoS value, the local limits are also limiting conditions of a multi-objective optimization problem, the use of cloud service resources with high reliability is a common local QoS requirement of a customer, and for a more complex service request, the user may require that the used cloud service combination must meet a certain reliability requirement (for example, the average level of a candidate resource provider is reached) before the service combination is used.
When performing metric calculation on QoS attributes, data generally needs to be preprocessed first, so that different QoS attributes have the same metric unit. For example, the time spent in the QoS attributes is calculated as "hours", the cost spent is calculated as "elements" or "dollars", and the reliability is described as "high", "normal", "low", etc., so the different QoS attributes describe different measurement units, and for more convenience in the algorithm processing stage, we need to specify the different QoS attributes, i.e. convert the different units of QoS attributes into the value of the measurement [1, 50 ].
The QoS attributes are classified into "forward QoS" and "reverse QoS". For example, the reliability is "forward QoS", the higher the reliability, the higher the probability that the customer will use the service, and the time and cost is "reverse QoS", the more time and cost it takes, the lower the probability that the customer will select the service. The reverse QoS attribute considered in the invention is time and cost, the forward QoS attribute is reliability and availability, the forward QoS is different from the reverse QoS, the specification of the forward QoS and the reverse QoS is achieved through the specification of the QoS attribute, and the forward QoS and the reverse QoS reach the same measurement unit. The following two formulas are the reduction formulas of the forward QoS attribute and the reverse QoS attribute respectively, max and min are the maximum value and the minimum value in the same QoS attribute,
among them, QosforwardIs an unreturned forward Qos attribute value, is an input value, QosreverseIs an unreduced reverse Qos attribute value, is an input value; qosstIs the reduced Qos attribute value and is the output value.
For the assignment problem, the following explanation is made:
suppose that in a certain service, the reliability index range is between [1, 100], the availability index range is [10, 30], the time cost range is between [1, 8], and the cost overhead range is between [20, 50 ].
In the existing four services, the reliability index Qo is respectively obtainedsforward140, availability index Qosforward2To 16, time spent Qosreverse14 hours, cost overhead Qosreverse2Respectively obtaining Qos through Qos reduction formula for 32 elementssst1、Qosst2、Qosst3And Qosst4Respectively 19.70, 15.00, 28.57 and 30.00, wherein Qosst1、Qosst2Qo is calculated according to the forward Qos reduction formulasst3And Qosst4Calculating according to a reverse Qos reduction formula, specifically as follows:
Qosst1＝(40-1)/(100-1)*50＝19.7；
Qosst2＝(16-10)/(30-10)*50＝15；
Qosst3＝(1-(4-1)/(8-1))*50＝28.57；
Qosst4＝(1-(32-20)/(50-20))*50＝30；
according to Qosst4>Qosst3>Qosst1>Qosst2It can be seen that the fourth service is the best.
1.3 Qos global optimization-based cloud service combination model
The invention adjusts the objective function by adjusting each weight value, thereby meeting the requirements of different users.
Let the set Q represent a plurality of local QoS attributes included in the global QoS attribute, i.e., Q ═ QoS1,QoS2,…QoSi…,QoSk}. and αwIs the weight ratio of each QoS attribute in the global QoS, andthen formulaA sub-service request T is describediThe sum of the respective QoS attribute metric values for the corresponding service resources l (i) is used, wherein,represents a subtask TjAttribute QoS in Using service resources L (j)wIs measured, subtask set T ═ T1,T2,…TmThe path of a complete set of cloud service combinations currently selected can be represented asThen the process of the first step is carried out,
without local QoS constraints, a set of global QoS expressions for a cloud service composition may be described as
If the current service composition problem is limited by local QoS, then
As can be seen from the above expression, the difference in the weight value settings of different QoS attributes by the user may largely affect the selection of the optimal service path, and therefore, the customizability of the cloud service combination model limited by the local QoS attributes is also embodied in that the model may adjust the service combination selection model according to the specific request information of the user and the requirements of the local attributes, so as to select the cloud service combination path that best meets the user requirements among the cloud resource providers.
2. And a second stage, namely a strategy and a model for scheduling problems facing the batch service request are provided and solved, and a corresponding mathematical model is provided.
2.1 establishment of scheduling architecture
The scheduling module of the cloud service management platform is not only connected with service requests of users (including a mobile terminal and a PC terminal), but also needs to send scheduling instructions to a cloud service resource provider and monitor data exchange between the cloud service provider and a data layer, after the users submit the service requests to the cloud platform, the users also need to submit corresponding information and corresponding detail requirements, and the cloud service management platform schedules cloud services to coordinate and meet each service request according to specific information and requirements of the service requests.
2.2 establishment of scheduling model
The present invention considers and models scheduling in two aspects, which are the minimization of overhead and total production time for service switching time, respectively.
(1) Service switching time overhead
The cloud service management system centrally coordinates and controls the cloud services of each single service to form a cloud service execution chain, and the time overhead of service switching among different cloud service providers depends on a cloud service management platform and a network environment, and a service switching time matrix logn×nThe switching time overhead between services provided by virtual service nodes is recorded, logijRepresents the service switching time overhead from the virtual service point i to the virtual service point j, since the service completion sub-requestThe order of the calculation is unidirectional, so the service switching time matrix is an upper right corner matrix, and the lower left part and the symmetry axis are 0. The cloud service management platform classifies and arranges the services provided by each resource provider to form various cloud services with different functions, and each service is called a virtual service point.
The service switching time between two virtual service nodes affects the total time for completing the service request to a certain extent. If a certain virtual service point in the cloud service execution chain is not needed for a certain service request, the corresponding service switching time is the switching time from the virtual service point with the previous cloud service execution time not being 0 to the service point with the next cloud service execution time not being 0.
(2) Minimizing total production time
The cloud platform is assumed to have n service points to form a service sequence, each service sequence can provide a class of cloud service for scheduling, and the optimal service combination path is selected in the service sequence for each complex cloud service request based on a QoS (quality of service) cloud service combination model and algorithm.
Assuming that m service requests are to be processed by the cloud service management platform, the multidimensional array TimeTable can be determined after the optimal cloud service combination is selected in the first stagem×nThe time cost of each sub-request of each service request on each service point is recorded, and after the optimal service path of the service request is determined, the cloud service with the single function corresponding to the sub-request is obtained on each service point. TimetableijRepresenting the cloud service S corresponding to the jth service pointjAnd time overhead for completing the sub-request corresponding to the service request No. i.
The cloud service management platform sets JQ as { JQ ═ JQ according to the service request sequence1,JQ2,JQ3......JQnScheduling the cloud services to complete the execution of the respective service request,andcorresponding services S respectively representing kth virtual service pointkCompleting execution service request JQiThe start time and the end time. Assuming that each service point satisfies only one service request at a time, the following constraint relationship exists:
the kth virtual service Point SkProviding service execution service request JQiThen, it needs to be served by the service point Sk+1Continue to provide corresponding cloud services for JQiAnd (6) executing. Virtual service Point S at this timek+1Whether to execute the request JQiDependent on JQiAt Sk+1Whether the required cloud service is currently executing sub-requests of other cloud service requests, and the cloud service management platform schedules the cloud service to complete the service requests according to the sequence array JQ, so if the virtual service point S isk+1While still executing other sub-service requests, then it is necessary that the JQ bei-1. So at the execution virtual service point Sk+1On-provisioning cloud service execution service request JQiIs at the starting time of SkUpper service execution request JQiAt the end of time Sk+1Execute-on JQi-1The larger value of the termination time.
Considering the service switching time at the same time, the following two cases need to be considered:
1) JQ for service requestiIf the JQ executed by the cloud service provided by the kth virtual service pointiThe time is 0, i.e., it is not necessary to perform this process. ThenThis is because the service request JQiService is not required to be provided by the kth virtual service point, and a JQ is requestediThe start time of the kth virtual service point is the end time of the last virtual service point, and the execution time of the k-th virtual service point is 0, so that the service request JQiThe start time of the kth virtual service Point, i.e. the end thereofA beam time;
2) JQ for service requestiIf the execution time of the current virtual service point is not 0, the current virtual service point needs to provide cloud service. Then service request JQiThe starting time of the kth virtual service point is the execution ending time of a virtual service point with a previous execution time of the kth virtual service point being not 0 plus the switching time of the virtual service point to the kth virtual service point.
Suppose thatIndicating a virtual service point whose execution time is not 0 before the kth virtual service point process. Then, combining the above two cases, there are the following constraint relations:
the invention adopts an evolutionary algorithm to pre-calculate the total execution time of various scheduling schemes before the batch service request is executed, and obtains the optimal scheduling scheme from the total execution time.
3. And optimizing based on the proposed two-stage cloud service combination and scheduling to achieve the aim of optimizing.
In the invention, the optimal service combination selection process of the first stage is to improve the execution quality of the service request, the service scheduling scheme of the second stage is to improve the execution efficiency, and the cloud service combination used in the scheduling process of the second stage is selected by the first stage in order to ensure the execution quality of the service request, so that the two-stage model connection is that the metric value of the response time attribute of the optimal service combination selected for each service request by the first stage is stored in the first stageTimeTableijIn the middle, the scheduling model of the second stage is just adopting TimeTableijAs an input source. The two-stage optimization model has the advantages that the two-stage optimization model can be distinguished, the coupling degree is low, the model can be updated and recombined according to actual problems, and the applicability of the two-stage optimization model to the actual problems is greatly improved.
The whole execution process of the cloud service combination and scheduling system driven by the service request of the user can be understood as an automatic processing process of a cloud service management platform consisting of an optimal service combination selection module based on QoS and a service resource scheduling module. The processing mode of the cloud service platform for the service combination can not only have sharp feedback processing on the local requirements of the cloud service request of the user, but also adjust the proportion parameter of the model at any time according to the requirements of the user to select the service combination which is most suitable for the user, and simultaneously solve the problem of service resource conflict caused by service requests of users in batches. The whole service combination and scheduling system is a ring system taking a cloud service platform as a center, namely the cloud platform collects service provider resource information and service request information of users, performs optimization and most effective 'request-service' matching on the service request information, and meets the requirement of high efficiency of shortest total time.
Claims (10)
1. A two-stage combination and scheduling method for a batch cloud service request is characterized in that: the method comprises the following steps:
step 1.1: the method comprises the steps that a user describes detail requirements and an overall target of the user on a service combination through an ontology description language own-s to form a service request, and the service request is submitted to a cloud service management platform to form a cloud service request sequence;
step 1.2: the cloud service management platform logically divides the service request according to the detail requirement and the whole target of the service combination of the user to form a sub-request stream, and performs service rough selection on the sub-request;
step 1.3: selecting a service composition path for the service request based on the QoS;
step 1.4: the cloud service management platform makes a scheduling plan through a service scheduling model according to a service combination path corresponding to each request in the service request sequence, and enters a scheduling preparation stage;
step 1.5: and the cloud service management platform carries out service resource scheduling.
2. The method for combining and scheduling a batch of cloud service requests according to claim 1, wherein: in the step 1.2, when a user submits a large-scale complex request with diversified functions or a request with personalized customization, the cloud service management platform combines a plurality of single cloud services for the user, and the method comprises the following steps:
step 1.2.1: logically dividing coarse-grained service requests into fine-grained sub-request sets T ═ T1,T2,…TmEach sub-request can use a cloud service with a single function; the sub-requests are sequentially associated and do not form a ring;
step 1.2.2: forming a sub-request T according to the corresponding information of j cloud service resource suppliers registered by the cloud service management platformiCorresponding candidate service resource set
Step 1.2.3: to be provided withRepresents a sub-request TiIs accomplished by cloud service L (i), where L (i) is sub-request TiServices selected in the set of candidate cloud services, i.e.The sub-request set T ═ T1,T2,…TmThe path of the complete set of cloud service combinations used can be expressed as
3. The combination and scheduling method for the cloud service request batch according to claim 2, wherein: in the step 1.2, the service resource interacts with the data reading layer through XML data to read and write data; the data reading layer comprises a plurality of distributed databases.
4. The method for combining and scheduling a batch of cloud service requests according to claim 1, wherein: in step 1.3, selecting a service combination path for the service request based on QoS includes the following steps:
step 1.3.1: establishing a cloud service combination evaluation mechanism based on QoS;
step 1.3.2: and establishing a cloud service combination model based on Qos global optimization.
5. The combination and scheduling method for the cloud service request in batch according to claim 4, wherein: in the step 1.3.1, the cloud service combination evaluation mechanism includes a functional QoS attribute and a non-functional QoS attribute, and the functional QoS attribute includes a response timeAnd costThe non-functional QoS attributes include a reliability indexAnd availability index
6. The combination and scheduling method for the cloud service request batch according to claim 5, wherein: in said step 1.3.1, said response timeAnd costThe reliability index is a reverse Qos attributeAnd availability indexAnd assigning the reverse Qos attribute and the forward Qos attribute for the forward Qos attribute, wherein the total Qos must meet the requirementWhere max and min are the maximum and minimum values in the same QoS attribute.
7. The combination and scheduling method for the cloud service request in batch according to claim 4, wherein: in said step 1.3.2, let the set Q represent a number of global QoS attributesLocal QoS attributes, i.e. Q ═ QoS1,QoS2,…QoSi…,QoSk}；αwIs the weight ratio of each of several QoS attributes in the global QoS, andsub-request TiThe sum of the respective QoS attribute metric values using the corresponding service resources L (i) isWherein,represents a sub-request TjAttribute QoS in Using service resources L (j)wThe measurement value of (a); sub-request set T ═ T1,T2,…TmThe path of a complete set of cloud service combinations currently selected can be represented asWhen not limited by local QoS, a set of global QoS expressions of cloud service combination can be described asWhen the current service combination is limited by local QoS,
8. the method for combining and scheduling a batch of cloud service requests according to claim 7, wherein: the cloud service management platform makes a scheduling plan through a service scheduling model according to a service combination corresponding to each request in the service request sequence, wherein the scheduling model comprises service switching time expenditure and total time expenditure.
9. The method for combining and scheduling a batch of cloud service requests according to claim 8, wherein: the service switching time is a matrix logn×nWherein, logijThe service switching time cost from the virtual service point i to the virtual service point j, the service switching time matrix logn×nThe switching time overhead between services provided by virtual service nodes is recorded, logijRepresenting the service switching time overhead from virtual service point i to virtual service point j.
10. The method for combining and scheduling a batch of cloud service requests according to claim 9, wherein: cloud service S corresponding to jth service point in the service combinationjThe time overhead spent for completing the sub-request corresponding to the service request No. i is TimeTablem×n(ii) a The cloud service management platform sets JQ as { JQ ═ JQ according to the service request sequence1,JQ2,JQ3......JQnScheduling cloud services to complete execution of respective service requests toAndcorresponding service S representing the kth virtual service PointkCompleting execution service request JQiStart time and end time of, for a service request JQi，
JQ executed if cloud service provided by kth virtual service pointiTime 0, i.e. it is not necessary to perform this process, then
JQ executed if cloud service provided by kth virtual service pointiTime is not 0, i.e. it requires the virtual service point to provide cloud service, service request JQiThe starting time of the kth virtual service point is the kth virtual serviceAdding the time of the end of the execution of a virtual service point with the execution time not being 0 before the service point and the switching time of switching the virtual service point to the kth virtual service point;
to be provided withThe virtual service point which represents the k-th virtual service point process and has the previous execution time not equal to 0 has the following constraint relationship:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610747165.4A CN106357739B (en) | 2016-08-27 | 2016-08-27 | It is a kind of for batch cloud service request two stages combination and dispatching method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610747165.4A CN106357739B (en) | 2016-08-27 | 2016-08-27 | It is a kind of for batch cloud service request two stages combination and dispatching method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106357739A true CN106357739A (en) | 2017-01-25 |
| CN106357739B CN106357739B (en) | 2019-10-18 |
Family
ID=57855306
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610747165.4A Active CN106357739B (en) | 2016-08-27 | 2016-08-27 | It is a kind of for batch cloud service request two stages combination and dispatching method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106357739B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107169727A (en) * | 2017-04-20 | 2017-09-15 | 青岛科技大学 | Towards the cloud service combined method and platform of Making mold |
| CN110120888A (en) * | 2019-04-26 | 2019-08-13 | 井冈山大学 | A kind of cloud crowdsourcing platform magnanimity Service Source combined optimization method and system |
| CN110188128A (en) * | 2019-05-29 | 2019-08-30 | 南京邮电大学 | It is a kind of with energy conservation for guiding cloud data center system service method for guaranteeing quality |
| CN110650032A (en) * | 2018-06-27 | 2020-01-03 | 复旦大学 | Method for constructing QoS-based application optimization deployment scheme in multi-cloud environment |
| CN110753083A (en) * | 2019-09-06 | 2020-02-04 | 江苏中云科技有限公司 | Cloud service resource uniform distribution system for multiple service providers |
| CN111047170A (en) * | 2019-12-04 | 2020-04-21 | 北京航空航天大学 | Manufacturing service cooperative scheduling method based on long-term and short-term utility |
| CN111629053A (en) * | 2020-05-27 | 2020-09-04 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Credible geographic information service self-adaptive combination method and system |
| CN111913456A (en) * | 2020-08-13 | 2020-11-10 | 成都信息工程大学 | Cloud manufacturing virtual workshop construction method facing manufacturing task |
| CN112104697A (en) * | 2018-05-31 | 2020-12-18 | 华为技术有限公司 | Data processing method, multi-cloud management system and related equipment |
| CN112600936A (en) * | 2020-12-29 | 2021-04-02 | 山东省计算中心（国家超级计算济南中心） | Quantification and evaluation method for cloud computing system service |
| CN112765407A (en) * | 2020-12-30 | 2021-05-07 | 重庆邮电大学 | QoS service combination method based on user preference in Internet of things environment |
| CN112989558A (en) * | 2021-01-14 | 2021-06-18 | 北京中电飞华通信有限公司 | Power service quality matching optimization method based on multi-objective optimization and related equipment |
| US20210281610A1 (en) * | 2020-02-26 | 2021-09-09 | CloudKnox Security, Inc. | Method and System for Quantifying and Improving Conformance to Least Privilege Security Policies |
| CN113743646A (en) * | 2021-07-19 | 2021-12-03 | 北京工业大学 | Enterprise production resource reasonable configuration cloud platform scheduling method based on processing and manufacturing granularity |
| CN114580911A (en) * | 2022-03-04 | 2022-06-03 | 重庆大学 | Site-factory hybrid service and resource scheduling method |
| CN114629959A (en) * | 2022-03-22 | 2022-06-14 | 北方工业大学 | Context-aware IoT service scheduling policy generation method in cloud environment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101969475A (en) * | 2010-11-15 | 2011-02-09 | 张军 | Business data controllable distribution and fusion application system based on cloud computing |
| CN105471844A (en) * | 2015-11-15 | 2016-04-06 | 北京工业大学 | Cloud service dynamic combination method based on trust synthesis |
-
2016
- 2016-08-27 CN CN201610747165.4A patent/CN106357739B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101969475A (en) * | 2010-11-15 | 2011-02-09 | 张军 | Business data controllable distribution and fusion application system based on cloud computing |
| CN105471844A (en) * | 2015-11-15 | 2016-04-06 | 北京工业大学 | Cloud service dynamic combination method based on trust synthesis |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107169727A (en) * | 2017-04-20 | 2017-09-15 | 青岛科技大学 | Towards the cloud service combined method and platform of Making mold |
| CN107169727B (en) * | 2017-04-20 | 2024-03-29 | 青岛科技大学 | Cloud service combination method and system for mold manufacturing |
| CN112104697B (en) * | 2018-05-31 | 2022-03-04 | 华为技术有限公司 | Data processing method, multi-cloud management system and related equipment |
| US11765243B2 (en) | 2018-05-31 | 2023-09-19 | Huawei Technologies Co., Ltd. | Data processing method, multi-cloud management system, and related device |
| CN112104697A (en) * | 2018-05-31 | 2020-12-18 | 华为技术有限公司 | Data processing method, multi-cloud management system and related equipment |
| CN110650032A (en) * | 2018-06-27 | 2020-01-03 | 复旦大学 | Method for constructing QoS-based application optimization deployment scheme in multi-cloud environment |
| CN110120888A (en) * | 2019-04-26 | 2019-08-13 | 井冈山大学 | A kind of cloud crowdsourcing platform magnanimity Service Source combined optimization method and system |
| CN110188128A (en) * | 2019-05-29 | 2019-08-30 | 南京邮电大学 | It is a kind of with energy conservation for guiding cloud data center system service method for guaranteeing quality |
| CN110753083A (en) * | 2019-09-06 | 2020-02-04 | 江苏中云科技有限公司 | Cloud service resource uniform distribution system for multiple service providers |
| CN111047170A (en) * | 2019-12-04 | 2020-04-21 | 北京航空航天大学 | Manufacturing service cooperative scheduling method based on long-term and short-term utility |
| CN111047170B (en) * | 2019-12-04 | 2021-09-10 | 北京航空航天大学 | Manufacturing service cooperative scheduling method based on long-term and short-term utility |
| US11818175B2 (en) * | 2020-02-26 | 2023-11-14 | Microsoft Technology Licensing, Llc | Method and system for quantifying and improving conformance to least privilege security policies |
| US20210281610A1 (en) * | 2020-02-26 | 2021-09-09 | CloudKnox Security, Inc. | Method and System for Quantifying and Improving Conformance to Least Privilege Security Policies |
| CN111629053A (en) * | 2020-05-27 | 2020-09-04 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Credible geographic information service self-adaptive combination method and system |
| CN111629053B (en) * | 2020-05-27 | 2023-10-20 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Trusted geographic information service self-adaptive combination method and system |
| CN111913456B (en) * | 2020-08-13 | 2023-06-16 | 成都信息工程大学 | Manufacturing task-oriented cloud manufacturing virtual workshop construction method |
| CN111913456A (en) * | 2020-08-13 | 2020-11-10 | 成都信息工程大学 | Cloud manufacturing virtual workshop construction method facing manufacturing task |
| CN112600936A (en) * | 2020-12-29 | 2021-04-02 | 山东省计算中心（国家超级计算济南中心） | Quantification and evaluation method for cloud computing system service |
| CN112765407B (en) * | 2020-12-30 | 2022-11-11 | 重庆邮电大学 | QoS service combination method based on user preference in Internet of things environment |
| CN112765407A (en) * | 2020-12-30 | 2021-05-07 | 重庆邮电大学 | QoS service combination method based on user preference in Internet of things environment |
| CN112989558A (en) * | 2021-01-14 | 2021-06-18 | 北京中电飞华通信有限公司 | Power service quality matching optimization method based on multi-objective optimization and related equipment |
| CN113743646A (en) * | 2021-07-19 | 2021-12-03 | 北京工业大学 | Enterprise production resource reasonable configuration cloud platform scheduling method based on processing and manufacturing granularity |
| CN114580911A (en) * | 2022-03-04 | 2022-06-03 | 重庆大学 | Site-factory hybrid service and resource scheduling method |
| CN114629959A (en) * | 2022-03-22 | 2022-06-14 | 北方工业大学 | Context-aware IoT service scheduling policy generation method in cloud environment |
| CN114629959B (en) * | 2022-03-22 | 2023-11-17 | 北方工业大学 | Cloud environment context-aware internet traffic (IoT) service scheduling policy method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106357739B (en) | 2019-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106357739B (en) | It is a kind of for batch cloud service request two stages combination and dispatching method | |
| Zhu et al. | Task scheduling for multi-cloud computing subject to security and reliability constraints | |
| CN111756812B (en) | Energy consumption perception edge cloud cooperation dynamic unloading scheduling method | |
| Masdari et al. | Efficient task and workflow scheduling in inter-cloud environments: challenges and opportunities | |
| Shi et al. | Location-aware and budget-constrained service deployment for composite applications in multi-cloud environment | |
| WO2019179250A1 (en) | Scheduling method, scheduler, storage medium, and system | |
| CN108880663A (en) | Incorporate network resource allocation method based on improved adaptive GA-IAGA | |
| CN107656799B (en) | Workflow scheduling method considering communication and computing cost under multi-cloud environment | |
| CN108616394B (en) | Virtual network function backup and deployment method | |
| CN103092683A (en) | Scheduling used for analyzing data and based on elicitation method | |
| CN105119751B (en) | A kind of service valuation and choosing method based on environment real-time perception | |
| CN112559135B (en) | Container cloud resource scheduling method based on QoS | |
| CN106779283A (en) | A kind of SLA ensures the desired cloud computing service income methods of QoS | |
| CN107203421A (en) | A kind of adaptive work in combination stream method in cloud computing environment | |
| Qian et al. | A workflow-aided Internet of things paradigm with intelligent edge computing | |
| CN111782627A (en) | Task and data cooperative scheduling method for wide-area high-performance computing environment | |
| CN110119399A (en) | Work Flow Optimizing method based on machine learning | |
| CN117596246B (en) | Method and system for scheduling workflow of computing power network based on heterogeneous resource measurement characteristics | |
| CN111770020A (en) | Method for network perception service combination algorithm based on optimal path selection | |
| Yang et al. | Optimizing aggregation frequency for hierarchical model training in heterogeneous edge computing | |
| CN118396294A (en) | Cloud manufacturing scheduling method based on quantum multi-agent reinforcement learning | |
| Li et al. | Cost-aware scheduling for ensuring software performance and reliability under heterogeneous workloads of hybrid cloud | |
| CN111654396B (en) | Aggregation cooperative configuration method for manufacturing service oriented to aggregation task multidimensional decomposition | |
| Di Cicco et al. | DRL-FORCH: A Scalable Deep Reinforcement Learning-based Fog Computing Orchestrator | |
| CN113010296A (en) | Task analysis and resource allocation method and system based on formalized model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |