WO2014094495A1 - 云调度方法及其系统 - Google Patents

云调度方法及其系统 Download PDF

Info

Publication number
WO2014094495A1
WO2014094495A1 PCT/CN2013/085748 CN2013085748W WO2014094495A1 WO 2014094495 A1 WO2014094495 A1 WO 2014094495A1 CN 2013085748 W CN2013085748 W CN 2013085748W WO 2014094495 A1 WO2014094495 A1 WO 2014094495A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
cloud
priority
resource
scheduling
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.)
Ceased
Application number
PCT/CN2013/085748
Other languages
English (en)
French (fr)
Inventor
朱定局
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2014094495A1 publication Critical patent/WO2014094495A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5066—Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs

Definitions

  • the present invention relates to the technical field of cloud computing scheduling through a network, and in particular to a cloud scheduling method, and a cloud scheduling system.
  • cloud computing scheduling is based on the task quantity and resource amount information of the cloud scheduling task, and the cloud scheduling of network resources is performed.
  • cloud scheduling is based only on the amount of tasks and resources. Although it can achieve the goal of fully utilizing cloud computing resources and processing cloud scheduling tasks as quickly as possible, it cannot meet the different individual needs of different cloud scheduling tasks, such as some clouds. Processing tasks require high quality of service, but not cost factors, and some cloud processing tasks require low-cost processing, but the processing quality and other service quality requirements are not high, and in the existing cloud scheduling mode, All cloud scheduling tasks are treated equally, and the cloud service scheduling obtained by each cloud scheduling task is the same, and the cloud service effect that meets its own requirements cannot be obtained.
  • an object of the present invention is to provide a cloud scheduling method and system thereof, which can process different hardware and software resources according to different requirements of different cloud scheduling tasks to meet different cloud scheduling tasks. service requirements.
  • a cloud scheduling method includes the following steps:
  • Receiving a cloud scheduling task acquiring user level and priority attribute setting information of the cloud scheduling task;
  • the processing results of the plurality of subtasks are combined to generate a processing result of the cloud scheduling task.
  • a cloud scheduling system comprising:
  • An information acquiring module configured to receive a cloud scheduling task, and obtain user level and priority attribute setting information of the cloud scheduling task;
  • a task division module configured to divide the cloud scheduling task into several subtasks
  • a resource scheduling module configured to separately perform corresponding hardware resources and software resources in the cloud network system to perform parallel processing on the plurality of subtasks according to the user level and priority attribute setting information, and obtain a processing result
  • the result merging module is configured to combine the processing results of the several subtasks to generate a processing result of the cloud scheduling task.
  • the cloud scheduling method and the system of the present invention divide the cloud scheduling task into a plurality of subtasks by acquiring the user level and priority attribute setting information of the cloud scheduling task, and setting the information scheduling corresponding soft according to the user level and the priority attribute.
  • the hardware resources process each sub-task to meet the processing requirements of different cloud scheduling tasks, and finally merge the processing results of the sub-tasks to obtain the processing result of the cloud scheduling task, improve the processing efficiency, and satisfy the personalized service of different cloud scheduling tasks. demand.
  • FIG. 1 is a schematic flow chart of a cloud scheduling method according to the present invention.
  • FIG. 2 is a schematic structural diagram of a cloud scheduling system of the present invention.
  • FIG. 1 is a schematic flowchart of a cloud scheduling method according to the present invention.
  • the cloud scheduling method includes the following steps:
  • S101 Receive a cloud scheduling task, and obtain user level and priority attribute setting information of the cloud scheduling task.
  • the cloud scheduling task is divided into several subtasks.
  • the invention can divide the cloud scheduling task into parallel processing of multiple subtasks, and schedule different hardware and software resources according to the requirements of different cloud scheduling tasks, improve processing efficiency, and meet the personalized service requirements of different cloud scheduling tasks.
  • the cloud scheduling task includes a cloud computing scheduling task issued by various users, including a free user, a paying user, and a scheduling of a video cloud service for scheduling a cloud resource for processing, and a robot cloud service. Scheduling, data review, scheduling tasks for cloud services.
  • the user levels of the cloud scheduling task include free users, paying users, and VIP users.
  • the user level of the user who issues the cloud scheduling task may be determined by acquiring user account information such as login information, registration information, or payment information of the user;
  • the priority attribute setting information of the cloud scheduling task refers to priorities of various cloud scheduling service indicators required by the cloud scheduling task, including speed priority, cost priority, quality priority, and combinations thereof.
  • the priority attribute setting information may prompt the user to input when the cloud scheduling task is issued, that is, provide the user with a selection by using an option provided in the user interface, for example, by using a drop-down box mode, and having multiple options in the drop-down box. Different options include different priority attribute setting information, such as speed priority, cost priority, quality priority, and combinations thereof.
  • the priority attribute setting information of the cloud scheduling task may also be obtained by collecting and filtering various behavior information and input information of the corresponding user.
  • the division of the cloud scheduling task may be determined according to the task amount (the amount of calculation) of the cloud scheduling task.
  • the present invention provides a method for dividing the cloud scheduling task as follows:
  • Y M*X/N; where Y is the default available resource amount of the cloud scheduling task, M is the total resource amount, and N is the total task. the amount, X is the task amount of the cloud scheduling task;
  • the number of subtasks divided by the cloud scheduling task is set.
  • the division of the cloud scheduling task first considers the amount of available resources, that is, the resource share that can be occupied in the cloud network system, and if the available resource amount can be divided into more subtasks. , so that the processing time occupied by each subtask is shortened, and the speed of cloud scheduling is maximized.
  • the number of subtasks divided by the cloud scheduling task may be uniformly set to be equal to the default concurrency.
  • the number of subtasks divided by the cloud scheduling task is set according to the user's personalized requirement, and consider that the user level includes a free user, a paying user, and a VIP user; the priority attribute setting information includes: speed priority The cost priority and the quality priority are set.
  • the user level of the cloud scheduling task and the priority attribute setting information are first determined:
  • the number of subtasks that are set to be divided by the cloud scheduling task is smaller than the default concurrency
  • the number of subtasks that are set to be divided by the cloud scheduling task is greater than the default concurrency.
  • the number of subtasks divided by each cloud scheduling task may be specifically set according to the personalized requirements of the user or different cloud scheduling tasks.
  • the processing speed and processing quality required by the cloud scheduling task are higher, the Increase the number of subtasks that are divided, schedule more resources to process each subtask, and improve the processing speed and processing quality of the entire cloud scheduling task.
  • a resource scheduling control table may be pre-stored in the scheduling terminal, where the resource attributes of each of the hardware resources and software resources are pre-stored, and various user levels and The priority attribute sets the resource attribute corresponding to the information.
  • the resource attributes of the hardware resources and the software resources are first entered and saved in the resource scheduling control table, and the setting of the resource scheduling control table is set. Entering various user levels and resource attributes corresponding to the priority attribute setting information;
  • the resource scheduling control table saved in advance is obtained according to the user level and the priority attribute setting information of the cloud scheduling task, and the corresponding resource attribute is obtained, and then the corresponding resource attribute is queried according to the resource attribute.
  • the hardware resource and the software resource according to the query result, schedule corresponding hardware resources and software resources in the cloud network system to perform parallel processing on the plurality of subtasks.
  • cloud scheduling can be implemented by calling a scheduling interface provided by the cloud computing system, and existing cloud computing systems (such as hadoop, a distributed system infrastructure) are With the scheduling module API interface, the resources in the existing cloud computing system can be scheduled through the scheduling module API interface.
  • existing cloud computing systems such as hadoop, a distributed system infrastructure
  • the resource attributes of the hardware resource and the software resource include: cost, speed, and stability.
  • the user level includes: a free user, a paying user, and a VIP user;
  • the priority attribute setting information includes: speed priority, cost priority, and quality priority;
  • the scheduling of the corresponding resource is preferably performed by using the mode:
  • the priority attribute setting information is a cost priority
  • a hardware resource and a software resource pair whose cost in the resource attribute is lower than a default value is invoked.
  • the several subtasks are processed in parallel. That is, if it is a free user or a low-priority charging user and a VIP user, the low-priority cloud scheduling mode is adopted, that is, the hardware and software modules that consume less resources than the default scheduling method are used as much as possible. At this time, the quality of cloud services is relatively low and the speed of cloud services is relatively slow;
  • the plurality of subtasks are processed in parallel by calling hardware resources and software resources whose speed is higher than a default value in the resource attribute. That is, if the charging user or the VIP user selects the speed priority mode, try to call a hardware module with a higher performance than the default scheduling mode, and a more functional software module;
  • the priority attribute setting information is quality priority
  • hardware resources and software resources whose stability in the callability resource attribute is higher than a default value are parallel to the several subtasks. deal with. That is, if the charging user or the VIP user selects the quality priority mode, try to call a hardware module that is more stable than the default scheduling mode, and a more complete software service module;
  • the priority attribute setting information is quality priority and speed priority
  • hardware resources and software in the resource attribute that are faster than the default value and whose stability is higher than the default value are called.
  • the resources process the several subtasks in parallel. That is, if the charging user or the VIP user selects the speed priority mode and the quality priority at the same time, try to call a hardware module with more stable and higher performance than the default scheduling mode, and a more complete software service module.
  • the default values corresponding to various resource attributes may be set according to actual needs, and are set to mean the resource attributes averaged when all cloud scheduling tasks are treated equally, regardless of the individual requirements of the cloud scheduling task.
  • the method for scheduling the parallel processing of the plurality of subtasks by the corresponding hardware resources and software resources in the cloud network system according to the query result is:
  • the content of the running process is the number of software resources obtained by the query.
  • the parallel processing of the plurality of subtasks is performed by the parallel cloud processing process, the processing time of the cloud scheduling task is greatly shortened, and the processing efficiency is improved.
  • the resource attributes of the hardware resource and the software resource may also be set to other parameters.
  • the resource attribute may be set to a CPU performance range. Memory performance range and disk performance range.
  • the cloud scheduling task v is divided into p subtasks ⁇ v1, v2, . . . , vp ⁇ in step S102, in this step, according to the query result of the resource scheduling control table, the cloud resource pool is selected.
  • the CPU performance range satisfies (c1, c2)
  • the memory performance range satisfies (m1, m2)
  • the disk performance range satisfies (d1, d2)
  • p hardware resources ⁇ h1, h2, ..., hp ⁇
  • the corresponding p software Resources ⁇ s1, s2,...,sp ⁇ .
  • p parallel cloud processing processes corresponding to the subtask collection are cloned; p parallel cloud processing is processed according to p parallel cloud processing processes, p computer hardware resources, p software resources The process is assigned to run on p corresponding computer hardware resources, and the running process content is p software resources.
  • step S104 the processing result sets of the plurality of subtasks are merged by the p parallel cloud processing processes, and the processing result of the cloud scheduling task is generated, and returned to the user to complete the processing process of the cloud scheduling task.
  • the cloud scheduling method of the present invention divides the cloud scheduling task into a plurality of subtasks, and performs parallel processing on each of the subtasks to merge the processing results into the processing result of the cloud scheduling task. Therefore, the processing time can be greatly shortened.
  • the user level and priority attribute setting information of the cloud scheduling task invoke different software and hardware resources to process each sub-task, which can meet the processing requirements of different cloud scheduling tasks, make cloud scheduling more flexible, and improve resource utilization.
  • FIG. 2 is a schematic structural diagram of a cloud scheduling system according to the present invention.
  • the cloud scheduling system includes:
  • the information obtaining module 11 is configured to receive a cloud scheduling task, and acquire user level and priority attribute setting information of the cloud scheduling task.
  • the task division module 12 is configured to divide the cloud scheduling task into several subtasks
  • the resource scheduling module 13 is configured to separately perform corresponding hardware resources and software resources in the cloud network system to perform parallel processing on the plurality of subtasks according to the user level and priority attribute setting information, and obtain a processing result;
  • the result merging module 14 is configured to combine the processing results of the plurality of subtasks to generate a processing result of the cloud scheduling task.
  • the cloud scheduling task includes cloud computing scheduling tasks issued by various users, including free users, paying users, and scheduling of video cloud services for scheduling cloud resources for processing, processing of robot cloud services, and data review. Scheduling tasks for cloud services.
  • the information acquiring module 11 Upon receiving the cloud scheduling task, the information acquiring module 11 acquires user level and priority attribute setting information of the cloud scheduling task.
  • the user levels of the cloud scheduling task include free users, paying users, and VIP users.
  • the information acquiring module 11 may determine the user level of the user who sends the cloud scheduling task by acquiring user account information such as login information, registration information, or payment information of the user;
  • the priority attribute setting information of the cloud scheduling task refers to priorities of various cloud scheduling service indicators required by the cloud scheduling task, including speed priority, cost priority, quality priority, and combinations thereof.
  • the priority attribute setting information may prompt the user to input when the cloud scheduling task is issued, that is, provide the user with a selection by using an option provided in the user interface, for example, by using a drop-down box mode, and having multiple options in the drop-down box. Different options include different priority attribute setting information, such as speed priority, cost priority, quality priority, and combinations thereof.
  • the priority attribute setting information of the cloud scheduling task may also be obtained by collecting and filtering various behavior information and input information of the corresponding user.
  • the division of the cloud scheduling task by the task division module 12 may be determined according to the task quantity (calculation amount) of the cloud scheduling task.
  • the task division module includes the following sub-modules:
  • the dividing module is configured to set the number of subtasks divided by the cloud scheduling task according to the default concurrency.
  • the task division module 12 first considers the amount of available resources, that is, the resource share that can be occupied in the cloud network system, and may divide the available resource amount. Make more subtasks, shorten the processing time occupied by each subtask, and try to increase the speed of cloud scheduling.
  • the number of subtasks divided by the cloud scheduling task may be uniformly set to be equal to the default concurrency.
  • the number of subtasks that are divided by the cloud scheduling task is set according to the user's personalized requirement, and the user level includes the free user, the paying user, and the VIP user; the priority attribute setting information includes: speed priority, Cost priority and quality priority.
  • the task division module 12 first determines the user level and priority attribute setting information of the cloud scheduling task when the number of subtasks divided by the cloud scheduling task is set according to the default concurrency:
  • the number of subtasks that are set to be divided by the cloud scheduling task is smaller than the default concurrency
  • the number of subtasks that are set to be divided by the cloud scheduling task is greater than the default concurrency.
  • the number of subtasks divided by each cloud scheduling task may be specifically set according to the personalized requirements of the user or different cloud scheduling tasks.
  • the processing speed and processing quality required by the cloud scheduling task are higher, the Increase the number of subtasks that are divided, schedule more resources to process each subtask, and improve the processing speed and processing quality of the entire cloud scheduling task.
  • the resource scheduling module 13 is configured to process, according to the user level and the priority attribute setting information, the corresponding hardware resources and software resources in the cloud network system to process each of the plurality of subtasks.
  • a resource scheduling control table may be pre-stored in the scheduling terminal, where the resource attributes of each of the hardware resources and software resources are pre-stored, and various user levels and The priority attribute sets the resource attribute corresponding to the information.
  • the resource attributes of the hardware resources and the software resources are first entered and saved in the resource scheduling control table, and the setting of the resource scheduling control table is set. Entering various user levels and resource attributes corresponding to the priority attribute setting information;
  • the resource scheduling control table saved in advance is obtained according to the user level and the priority attribute setting information of the cloud scheduling task, and the corresponding resource attribute is obtained, and then the corresponding resource attribute is queried according to the resource attribute.
  • the hardware resource and the software resource according to the query result, schedule corresponding hardware resources and software resources in the cloud network system to perform parallel processing on the plurality of subtasks.
  • the resource scheduling module 13 can implement cloud scheduling by calling a scheduling interface provided by the cloud computing system when scheduling each of the hardware resources and software resources, and the existing cloud computing system (for example, hadoop) has scheduling.
  • the module API interface can schedule resources in the existing cloud computing system through the scheduling module API interface.
  • the resource attributes of the hardware resource and the software resource include: cost, speed, and stability.
  • the user level includes: a free user, a paying user, and a VIP user;
  • the priority attribute setting information includes: speed priority, cost priority, and quality priority;
  • the resource scheduling module 13 preferably includes the following sub-modules:
  • a first scheduling module configured to: when the user level is a free user, or the user level is a paying user or a VIP user, and the priority attribute setting information is a cost priority, the cost in the calling resource attribute is lower than a default value.
  • the hardware resources and software resources process the several subtasks in parallel. That is, if it is a free user or a low-priority charging user and a VIP user, the low-priority cloud scheduling mode is adopted, that is, the hardware and software modules that consume less resources than the default scheduling method are used as much as possible. At this time, the quality of cloud services is relatively low and the speed of cloud services is relatively slow;
  • a second scheduling module configured to: when the user attribute is a paying user or a VIP user, and the priority attribute setting information is speed priority, calling a hardware resource and a software resource whose speed is higher than a default value in the resource attribute Several subtasks are processed in parallel. That is, if the charging user or the VIP user selects the speed priority mode, try to call a hardware module with a higher performance than the default scheduling mode, and a more functional software module;
  • a third scheduling module configured to: when the user level is a paying user or a VIP user, and the priority attribute setting information is quality priority, the stability of the invocation resource attribute is higher than a default value of the hardware resource and the software resource pair
  • the several subtasks are processed in parallel. That is, if the charging user or the VIP user selects the quality priority mode, try to call a hardware module that is more stable than the default scheduling mode, and a more complete software service module;
  • a fourth scheduling module configured to: when the user level is a paying user or a VIP user, and the priority attribute setting information is quality priority and speed priority, the speed in the calling resource attribute is higher than a default value, and the stability is higher than
  • the default hardware and software resources process the several subtasks in parallel. That is, if the charging user or the VIP user selects the speed priority mode and the quality priority at the same time, try to call a hardware module with more stable and higher performance than the default scheduling mode, and a more complete software service module.
  • the default values corresponding to various resource attributes may be set according to actual needs, and are set to mean the resource attributes averaged when all cloud scheduling tasks are treated equally, regardless of the individual requirements of the cloud scheduling task.
  • the method for scheduling the parallel processing of the plurality of subtasks by the corresponding hardware resources and software resources in the cloud network system according to the query result is:
  • the content of the running process is the number of software resources obtained by the query.
  • the parallel processing of the plurality of subtasks is performed by the parallel cloud processing process, the processing time of the cloud scheduling task is greatly shortened, and the processing efficiency is improved.
  • the resource attributes of the hardware resource and the software resource may also be set to other parameters.
  • the resource attribute may be set to a CPU performance range. Properties such as memory performance range and disk performance range.
  • the result merging module 14 merges the processing result sets of the plurality of subtasks by the p parallel cloud processing processes, generates the processing result of the cloud scheduling task, and returns the result to the user.
  • the cloud scheduling method and system thereof of the present invention can be used for cloud scheduling of all types of applications, including but not limited to: scheduling of video cloud services, scheduling of robot cloud services, scheduling of data review cloud services.
  • the storage medium may be a magnetic disk, an optical disk, or a read-only storage memory (Read-Only) Memory, ROM) or Random Access Memory (RAM).

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

本发明提供一种云调度方法及其系统，所述方法包括：接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；将所述云调度任务划分为若干个子任务；根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。本发明能够将云调度任务划分成为多个子任务并行处理，对不同的云调度任务根据其需求调度不同的硬件和软件资源，提高处理效率，满足不同云调度任务的个性化服务需求。

Description

云调度方法及其系统
【技术领域】
本发明涉及通过网络的云计算调度的技术领域，特别是涉及一种云调度方法，以及一种云调度系统。
【背景技术】
随着计算机网络技术的发展，通过网络的云计算调度技术的应用越来越广泛，目前云计算调度都是仅根据云调度任务的任务量、资源量的信息进行网络资源的云调度。
但是，仅根据任务量、资源量的信息进行云调度，虽然能够达到充分利用云计算资源、尽快处理云调度任务的目的，但却不能满足不同云调度任务的不同个性化需求，例如有的云处理任务要求高的服务质量，而对成本因素并不要求，而有的云处理任务要求采取低成本的处理方式，但对处理速度等服务质量要求不高，而在现有云调度模式下，对所有的云调度任务都是同等对待的，各个云调度任务获得的云服务调度是相同的，而无法获得与自身要求符合的云服务效果。
【发明内容】
针对上述背景技术中存在的问题，本发明的目的在于提供一种云调度方法及其系统，能够对不同的云调度任务根据其需求调度不同的硬件和软件资源进行处理，满足不同云调度任务的服务需求。
一种云调度方法，包括以下步骤：
接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
将所述云调度任务划分为若干个子任务；
根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
一种云调度系统，包括：
信息获取模块，用于接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
任务划分模块，用于将所述云调度任务划分为若干个子任务；
资源调度模块，用于根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
结果合并模块，用于将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
本发明的云调度方法及其系统通过获取云调度任务的用户等级和优先属性设置信息，将所述云调度任务划分为若干个子任务，根据所述用户等级和优先属性设置信息调度对应的软、硬件资源对各个子任务进行处理，满足不同的云调度任务的处理要求，最后合并子任务的处理结果，得到所述云调度任务的处理结果，提高处理效率，满足不同云调度任务的个性化服务需求。
【附图说明】
图1是本发明云调度方法的流程示意图；
图2是本发明云调度系统的结构示意图。
【具体实施方式】
请参阅图1，图1是本发明云调度方法的流程示意图。
所述云调度方法包括以下步骤：
S101，接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
S102，将所述云调度任务划分为若干个子任务；
S103，根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
S104，将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
本发明能够将云调度任务划分成为多个子任务并行处理，对不同的云调度任务根据其需求调度不同的硬件和软件资源，提高处理效率，满足不同云调度任务的个性化服务需求。
其中，对于步骤S101，所述云调度任务包括各种用户发出的云计算调度任务，包括免费用户、付费用户以及VIP用户发出的调度云资源进行计算处理的视频云服务的调度、机器人云服务的调度、数据审查云服务的调度任务。
在接收到所述云调度任务时，获取所述云调度任务的用户等级和优先属性设置信息。所述云调度任务的用户等级包括免费用户、付费用户以及VIP用户。可以通过获取用户的登陆信息、注册信息或者付费信息等用户账户信息来判断发出所述云调度任务的用户的用户等级；
所述云调度任务的优先属性设置信息是指所述云调度任务要求的各种云调度服务指标的优先级，包括速度优先、成本优先、质量优先及其组合。所述优先属性设置信息可以在发出所述云调度任务时提示用户输入，即通过用户界面中提供的选项，提供给用户进行选择，例如通过下拉框模式，在所述下拉框中有多个选项，不同的选项中包含有不同的优先属性设置信息，如速度优先、成本优先、质量优先及其组合。所述云调度任务的优先属性设置信息也可以通过对相应的用户的各种行为信息和输入信息的搜集、筛选提取得出。
对于上述步骤S102，对所述云调度任务的划分可以根据所述云调度任务的任务量（运算量）来确定，优选地，本发明提供一种所述云调度任务的划分方式如下：
首先计算所述云调度任务的默认可用资源量，根据以下公式：Y=M*X/N；其中，Y为所述云调度任务的默认可用资源量，M为资源总量，N为任务总量， X为所述云调度任务的任务量；
然后，根据所述默认可用资源量计算所述云调度任务的默认并发度，按照以下公式计算：p=Y/R，其中，p为默认并发度，R为每一个并发进程占用的资源量；
最后根据所述默认并发度，设置所述云调度任务划分的子任务个数。
在本实施方式中，对所述云调度任务的划分首先考虑其可用资源量，即在所述云网络系统中可以占用的资源份额，如果所述可用资源量则可以划分成更多的子任务，使每一个子任务所占用的处理时间缩短，尽量提高云调度的速度。
在获取所述默认并发度之后，可以将所述云调度任务划分的子任务个数统一设置成等于所述默认并发度。
本发明中，优选根据用户个性化需求设置所述云调度任务划分的子任务个数，则考虑到所述用户等级包括免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先、成本优先以及质量优先；在根据所述默认并发度设置所述云调度任务划分的子任务个数时，首先判断所述云调度任务的用户等级以及优先属性设置信息：
当用户等级为免费用户，或者用户等级为付费用户或VIP用户、且优先属性设置信息为成本优先时，设置所述云调度任务划分的子任务个数小于所述默认并发度；
当用户等级为付费用户或VIP用户，并且优先属性设置信息为速度优先，或质量优先时，设置所述云调度任务划分的子任务个数大于所述默认并发度。
通过上述方式，可以根据用户或者不同云调度任务的个性化需求，具体设置每个云调度任务划分的子任务个数，在所述云调度任务要求的处理速度和处理质量更高时，可以将增加其划分的子任务数，调度更多的资源去处理各个子任务，提高整个云调度任务的处理速度和处理质量。
对于上述步骤S103，根据所述用户等级和优先属性设置信息，在云网络系统中调度相应的硬件资源和软件资源对各个所述若干个子任务进行处理。为实现相应的资源的调度，可以在调度端预先保存一个资源调度控制表，所述资源调度控制表中预先保存有各个所述硬件资源和软件资源的资源属性，以及各种所述用户等级以及所述优先属性设置信息所对应的资源属性。
在各个硬件资源和软件资源投入云网络系统中使用时，首先将所述硬件资源和软件资源的资源属性录入并保存在所述资源调度控制表中，并通过对所述资源调度控制表的设置录入各种所述用户等级以及所述优先属性设置信息所对应的资源属性；
则在实际处理云调度任务时，根据所述云调度任务的用户等级和优先属性设置信息，查询预先保存的所述资源调度控制表，获取对应的资源属性，再根据所述资源属性查询对应的硬件资源和软件资源，根据查询结果调度云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理。
在调度各个所述硬件资源和软件资源时，可以通过调用所述云计算系统提供的调度接口衔接来实现云调度，现有的云计算系统（例如hadoop，一种分布式系统基础架构）都是有调度模块API接口的，通过调度模块API接口就可以调度现有云计算系统中的资源。
在所述资源调度控制表中，所述硬件资源和软件资源的资源属性包括：成本、速度、稳定性。
考虑到所述用户等级包括：免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先，成本优先以及质量优先；
在查询所述资源调度控制表，根据查询结果调度对应的硬件资源和软件资源时，优选采用与方式进行相应资源的调度：
当所述用户等级为免费用户，或者所述用户等级为付费用户或者VIP用户、并且所述优先属性设置信息为成本优先时，调用资源属性中的成本低于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。亦即，如果是免费用户或选取成本低优先的收费用户和VIP用户，则采用成本低优先的云调度方式，即尽量调用比默认调度方式耗资源更少的、成本更低廉的硬件和软件模块，此时云服务质量先对较低和云服务速度相对较慢；
当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为速度优先时，调用资源属性中的速度高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户选取速度优先方式，则尽量调用比默认调度方式更高性能的硬件模块、功能更简要的软件模块；
当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先时，调用性资源属性中的稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户选取质量优先方式，则尽量调用比默认调度方式更稳定的硬件模块、功能更完备的软件服务模块；
当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先以及速度优先时，调用资源属性中的速度高于默认值、并且稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户同时选取速度优先方式和质量优先，则尽量调用比默认调度方式更稳定、更高性能的硬件模块、功能更完备的软件服务模块。
通常，各种资源属性所对应的默认值可以根据实际需要设定，也以设定为不区分云调度任务的个性化需求，对所有的云调度任务同等对待时调度的资源属性平均值。
通过上述方式，根据不同的用户等级以及优先属性设置信息来调度不同的硬件资源和软件资源对各个子任务进行处理，满足不同云调度任务的个性化需求。
根据查询结果调度所述云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理的方法为：
根据所述若干个子任务的集合，生成与所述子任务一一对应的若干个并行云处理进程；将所述若干个并行云处理进程分别指派到查询获取的若干个相应的硬件资源上运行，运行的进程内容为查询获取的若干个软件资源。
由于是通过并行云处理进程对多个子任务进行并行处理，因此大大缩短了对所述云调度任务的处理时间，提高了处理效率。
在所述资源调度控制表中，所述硬件资源和软件资源的资源属性还可以设置成其他的参数，例如，对于硬件资源，可设置其资源属性为CPU性能范围， memory性能范围和disk性能范围。
如果在步骤S102中将所述云调度任务v划分为p个子任务{v1,v2,…,vp}，则在本步骤中，根据所述资源调度控制表的查询结果，从云资源池中选取CPU性能范围满足（c1,c2）,memory性能范围满足（m1,m2）,disk性能范围满足（d1,d2）的p个硬件资源{h1,h2,…,hp},以及对应的p个软件资源{s1,s2,…,sp}。
根据p个并发子任务集合，克隆出p个与所述子任务集合对应的并行云处理进程；根据p个并行云处理进程、p个计算机硬件资源、p个软件资源，将p个并行云处理进程指派到p个相应的计算机硬件资源上运行，运行的进程内容为p个软件资源。
对于上述步骤S104，对p个并行云处理进程对所述若干个子任务的处理结果集合进行合并，生成所述云调度任务的处理结果，并返回给用户，完成云调度任务的处理过程。
本发明的云调度方法将云调度任务划分为若干个子任务，对各个所述子任务进行并行处理后将处理结果合并成所述云调度任务的处理结果，因此，能够大大缩短处理时间。通过云调度任务的用户等级和优先属性设置信息调用不同的软、硬件资源对各个子任务进行处理，能够满足不同的云调度任务的处理要求，使云调度更灵活，提高资源利用率。
请参阅图2，图2是本发明云调度系统的结构示意图。
所述云调度系统包括：
信息获取模块11，用于接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
任务划分模块12，用于将所述云调度任务划分为若干个子任务；
资源调度模块13，用于根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
结果合并模块14，用于将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
其中，所述云调度任务包括各种用户发出的云计算调度任务，包括免费用户、付费用户以及VIP用户发出的调度云资源进行计算处理的视频云服务的调度、机器人云服务的调度、数据审查云服务的调度任务。
在接收到所述云调度任务时，所述信息获取模块11获取所述云调度任务的用户等级和优先属性设置信息。所述云调度任务的用户等级包括免费用户、付费用户以及VIP用户。所述信息获取模块11可以通过获取用户的登陆信息、注册信息或者付费信息等用户账户信息来判断发出所述云调度任务的用户的用户等级；
所述云调度任务的优先属性设置信息是指所述云调度任务要求的各种云调度服务指标的优先级，包括速度优先、成本优先、质量优先及其组合。所述优先属性设置信息可以在发出所述云调度任务时提示用户输入，即通过用户界面中提供的选项，提供给用户进行选择，例如通过下拉框模式，在所述下拉框中有多个选项，不同的选项中包含有不同的优先属性设置信息，如速度优先、成本优先、质量优先及其组合。所述云调度任务的优先属性设置信息也可以通过对相应的用户的各种行为信息和输入信息的搜集、筛选提取得出。
所述任务划分模块12对所述云调度任务的划分可以根据所述云调度任务的任务量（运算量）来确定，优选地，所述任务划分模块包括以下子模块：
可用资源计算模块，用于根据以下公式计算所述云调度任务的默认可用资源量：Y=M*X/N，其中，Y为所述云调度任务的默认可用资源量，M为资源总量，N为任务总量， X为所述云调度任务的任务量；
并发度计算模块，用于根据以下公式计算所述云调度任务的默认并发度：p=Y/R，其中，p为默认并发度，R为每一个并发进程占用的资源量；
划分模块，用于根据所述默认并发度，设置所述云调度任务划分的子任务个数。
在本实施方式中，所述任务划分模块12对所述云调度任务的划分首先考虑其可用资源量，即在所述云网络系统中可以占用的资源份额，如果所述可用资源量则可以划分成更多的子任务，使每一个子任务所占用的处理时间缩短，尽量提高云调度的速度。
在获取所述默认并发度之后，可以将所述云调度任务划分的子任务个数统一设置成等于所述默认并发度。
本发明中，优选根据用户个性化需求设置所述云调度任务划分的子任务个数，考虑到所述用户等级包括免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先、成本优先以及质量优先。则所述任务划分模块12在根据所述默认并发度设置所述云调度任务划分的子任务个数时，首先判断所述云调度任务的用户等级以及优先属性设置信息：
当用户等级为免费用户，或者用户等级为付费用户或VIP用户、且优先属性设置信息为成本优先时，设置所述云调度任务划分的子任务个数小于所述默认并发度；
当用户等级为付费用户或VIP用户，并且优先属性设置信息为速度优先，或质量优先时，设置所述云调度任务划分的子任务个数大于所述默认并发度。
通过上述方式，可以根据用户或者不同云调度任务的个性化需求，具体设置每个云调度任务划分的子任务个数，在所述云调度任务要求的处理速度和处理质量更高时，可以将增加其划分的子任务数，调度更多的资源去处理各个子任务，提高整个云调度任务的处理速度和处理质量。
所述资源调度模块13用于根据所述用户等级和优先属性设置信息，在云网络系统中调度相应的硬件资源和软件资源对各个所述若干个子任务进行处理。为实现相应的资源的调度，可以在调度端预先保存一个资源调度控制表，所述资源调度控制表中预先保存有各个所述硬件资源和软件资源的资源属性，以及各种所述用户等级以及所述优先属性设置信息所对应的资源属性。
在各个硬件资源和软件资源投入云网络系统中使用时，首先将所述硬件资源和软件资源的资源属性录入并保存在所述资源调度控制表中，并通过对所述资源调度控制表的设置录入各种所述用户等级以及所述优先属性设置信息所对应的资源属性；
则在实际处理云调度任务时，根据所述云调度任务的用户等级和优先属性设置信息，查询预先保存的所述资源调度控制表，获取对应的资源属性，再根据所述资源属性查询对应的硬件资源和软件资源，根据查询结果调度云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理。
所述资源调度模块13在调度各个所述硬件资源和软件资源时，可以通过调用所述云计算系统提供的调度接口衔接来实现云调度，现有的云计算系统（例如hadoop）都是有调度模块API接口的，通过调度模块API接口就可以调度现有云计算系统中的资源。
在所述资源调度控制表中，所述硬件资源和软件资源的资源属性包括：成本、速度、稳定性。
考虑到所述用户等级包括：免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先，成本优先以及质量优先；
则，所述资源调度模块13优选包括以下子模块：
第一调度模块，用于在所述用户等级为免费用户，或者所述用户等级为付费用户或者VIP用户、并且所述优先属性设置信息为成本优先时，调用资源属性中的成本低于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。亦即，如果是免费用户或选取成本低优先的收费用户和VIP用户，则采用成本低优先的云调度方式，即尽量调用比默认调度方式耗资源更少的、成本更低廉的硬件和软件模块，此时云服务质量先对较低和云服务速度相对较慢；
第二调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为速度优先时，调用资源属性中的速度高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户选取速度优先方式，则尽量调用比默认调度方式更高性能的硬件模块、功能更简要的软件模块；
第三调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先时，调用性资源属性中的稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户选取质量优先方式，则尽量调用比默认调度方式更稳定的硬件模块、功能更完备的软件服务模块；
第四调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先以及速度优先时，调用资源属性中的速度高于默认值、并且稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。即，如果是收费用户或VIP用户同时选取速度优先方式和质量优先，则尽量调用比默认调度方式更稳定、更高性能的硬件模块、功能更完备的软件服务模块。
通常，各种资源属性所对应的默认值可以根据实际需要设定，也以设定为不区分云调度任务的个性化需求，对所有的云调度任务同等对待时调度的资源属性平均值。
通过上述方式，根据不同的用户等级以及优先属性设置信息来调度不同的硬件资源和软件资源对各个子任务进行处理，满足不同云调度任务的个性化需求。
根据查询结果调度所述云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理的方法为：
根据所述若干个子任务的集合，生成与所述子任务一一对应的若干个并行云处理进程；将所述若干个并行云处理进程分别指派到查询获取的若干个相应的硬件资源上运行，运行的进程内容为查询获取的若干个软件资源。
由于是通过并行云处理进程对多个子任务进行并行处理，因此大大缩短了对所述云调度任务的处理时间，提高了处理效率。
在所述资源调度控制表中，所述硬件资源和软件资源的资源属性还可以设置成其他的参数，例如，对于硬件资源，可设置其资源属性为CPU性能范围， memory性能范围和disk性能范围等属性。
所述结果合并模块14对p个并行云处理进程对所述若干个子任务的处理结果集合进行合并，生成所述云调度任务的处理结果，并返回给用户。
本发明的云调度方法及其系统可以用于所有各种类型应用的云调度，包括但不限于：视频云服务的调度、机器人云服务的调度、数据审查云服务的调度。
本领域普通技术人员可以理解实现上述实施方式中的全部或部分流程以及对应的系统，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各实施方式的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体（Read-Only Memory，ROM）或随机存储记忆体（Random Access Memory，RAM）等。
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。

Claims (10)

  1. 一种云调度方法，其特征在于，包括以下步骤：
    接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
    将所述云调度任务划分为若干个子任务；
    根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
    将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
  2. 如权利要求1所述的云调度方法，其特征在于，将所述云调度任务划分为若干个子任务的步骤包括：
    根据以下公式计算所述云调度任务的默认可用资源量：Y=M*X/N，其中，Y为所述云调度任务的默认可用资源量，M为资源总量，N为任务总量， X为所述云调度任务的任务量；
    根据以下公式计算所述云调度任务的默认并发度：p=Y/R，其中，p为默认并发度，R为每一个并发进程占用的资源量；
    根据所述默认并发度，设置所述云调度任务划分的子任务个数。
  3. 如权利要求2所述的云调度方法，其特征在于，所述用户等级包括免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先、成本优先以及质量优先；
    根据所述默认并发度，设置所述云调度任务划分的子任务个数的步骤包括：
    当用户等级为免费用户，或者用户等级为付费用户或VIP用户、且优先属性设置信息为成本优先时，设置所述云调度任务划分的子任务个数小于所述默认并发度；
    当用户等级为付费用户或VIP用户，并且优先属性设置信息为速度优先，或质量优先时，设置所述云调度任务划分的子任务个数大于所述默认并发度。
  4. 如权利要求1所述的云调度方法，其特征在于，根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理的步骤包括：
    根据所述用户等级和优先属性设置信息，查询预先保存的资源调度控制表，根据查询结果调度所述云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理；其中，所述资源调度控制表中预先保存有各个所述硬件资源和软件资源的资源属性，以及各种所述用户等级以及所述优先属性设置信息所对应的资源属性。
  5. 如权利要求4所述的云调度方法，其特征在于，根据查询结果调度所述云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理的步骤包括：
    根据所述若干个子任务的集合，生成与所述子任务一一对应的若干个并行云处理进程；
    将所述若干个并行云处理进程分别指派到查询获取的若干个相应的硬件资源上运行，运行的进程内容为查询获取的若干个软件资源。
  6. 如权利要求4所述的云调度方法，其特征在于，所述用户等级包括：免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先，成本优先以及质量优先；所述资源属性包括：成本、速度、稳定性；
    查询预先保存的资源调度控制表，根据查询结果调度所述云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理的步骤包括：
    当所述用户等级为免费用户，或者所述用户等级为付费用户或者VIP用户、并且所述优先属性设置信息为成本优先时，调用资源属性中的成本低于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为速度优先时，调用资源属性中的速度高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先时，调用性资源属性中的稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    当所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先以及速度优先时，调用资源属性中的速度高于默认值、并且稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。
  7. 一种云调度系统，其特征在于，包括：
    信息获取模块，用于接收云调度任务，获取所述云调度任务的用户等级和优先属性设置信息；
    任务划分模块，用于将所述云调度任务划分为若干个子任务；
    资源调度模块，用于根据所述用户等级和优先属性设置信息，分别调度云网络系统中相应的硬件资源和软件资源对所述若干个子任务进行并行处理，并获取处理结果；
    结果合并模块，用于将所述若干个子任务的处理结果合并，生成所述云调度任务的处理结果。
  8. 如权利要求7所述的云调度系统，其特征在于，所述任务划分模块包括：
    可用资源计算模块，用于根据以下公式计算所述云调度任务的默认可用资源量：Y=M*X/N，其中，Y为所述云调度任务的默认可用资源量，M为资源总量，N为任务总量， X为所述云调度任务的任务量；
    并发度计算模块，用于根据以下公式计算所述云调度任务的默认并发度：p=Y/R，其中，p为默认并发度，R为每一个并发进程占用的资源量；
    划分模块，用于根据所述默认并发度，设置所述云调度任务划分的子任务个数；在用户等级为免费用户，或者用户等级为付费用户或VIP用户、且优先属性设置信息为成本优先时，设置所述云调度任务划分的子任务个数小于所述默认并发度；在用户等级为付费用户或VIP用户，并且优先属性设置信息为速度优先，或质量优先时，设置所述云调度任务划分的子任务个数大于所述默认并发度。
  9. 如权利要求7所述的云调度系统，其特征在于，所述资源调度模块根据所述用户等级和优先属性设置信息，查询预先保存的资源调度控制表，根据查询结果调度云网络系统中对应的硬件资源和软件资源对所述若干个子任务进行并行处理；其中，所述资源调度控制表中预先保存有各个所述硬件资源和软件资源的资源属性，以及各种所述用户等级以及所述优先属性设置信息所对应的资源属性。
  10. 如权利要求9所述的云调度系统，其特征在于，所述用户等级包括：免费用户、付费用户以及VIP用户；所述优先属性设置信息包括：速度优先，成本优先以及质量优先；所述资源属性包括：成本、速度、稳定性；
    所述资源调度模块包括：
    第一调度模块，用于在所述用户等级为免费用户，或者所述用户等级为付费用户或者VIP用户、并且所述优先属性设置信息为成本优先时，调用资源属性中的成本低于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    第二调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为速度优先时，调用资源属性中的速度高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    第三调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先时，调用性资源属性中的稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理；
    第四调度模块，用于在所述用户等级为付费用户或者VIP用户，并且所述优先属性设置信息为质量优先以及速度优先时，调用资源属性中的速度高于默认值、并且稳定性高于默认值的硬件资源和软件资源对所述若干个子任务进行并行处理。
PCT/CN2013/085748 2012-12-22 2013-10-23 云调度方法及其系统 Ceased WO2014094495A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210563107.8A CN103067468B (zh) 2012-12-22 2012-12-22 云调度方法及其系统
CN201210563107.8 2012-12-22

Publications (1)

Publication Number Publication Date
WO2014094495A1 true WO2014094495A1 (zh) 2014-06-26

Family

ID=48109922

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/085748 Ceased WO2014094495A1 (zh) 2012-12-22 2013-10-23 云调度方法及其系统

Country Status (2)

Country Link
CN (1) CN103067468B (zh)
WO (1) WO2014094495A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114995050A (zh) * 2022-06-27 2022-09-02 上海华力集成电路制造有限公司 提高opc运算效率的方法
CN116016711A (zh) * 2022-12-05 2023-04-25 江苏鑫辰通信有限公司 通信网络调度系统及方法及系统
CN116366355A (zh) * 2023-04-14 2023-06-30 北京智享嘉网络信息技术有限公司 一种网络设备硬件资源智能调度方法及系统

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103067468B (zh) * 2012-12-22 2016-03-09 深圳先进技术研究院 云调度方法及其系统
CN103699441B (zh) * 2013-12-05 2017-07-18 深圳先进技术研究院 基于任务粒度的MapReduce报表任务执行方法
CN103701886A (zh) * 2013-12-19 2014-04-02 中国信息安全测评中心 一种云计算环境下的服务及资源分层调度方法
CN104765640B (zh) * 2014-01-02 2018-02-16 中国科学院声学研究所 一种智能服务调度方法
US20170212791A1 (en) * 2014-08-15 2017-07-27 Intel Corporation Facilitating dynamic thread-safe operations for variable bit-length transactions on computing devices
CN104850576B (zh) * 2015-03-02 2018-07-24 武汉烽火众智数字技术有限责任公司 一种基于海量视频的快速特征提取系统
CN107239327A (zh) * 2016-03-29 2017-10-10 平安科技（深圳）有限公司 保单处理的优化方法和装置
CN107315409A (zh) * 2017-05-27 2017-11-03 芜湖星途机器人科技有限公司 银行服务机器人调度跟随系统的硬件平台
CN108227654A (zh) * 2017-12-28 2018-06-29 顺丰科技有限公司 一种调度服务端、调度装置、机器人系统及调度方法
CN109992403B (zh) * 2017-12-30 2021-06-01 中国移动通信集团福建有限公司 多租户资源调度的优化方法、装置、终端设备及存储介质
CN109669773B (zh) * 2018-11-12 2024-03-08 平安科技（深圳）有限公司 金融数据处理方法、装置、设备和存储介质
CN112540841B (zh) * 2020-12-28 2021-11-12 智慧神州(北京)科技有限公司 任务调度的方法、装置、处理器与电子设备
US11880363B2 (en) * 2021-01-20 2024-01-23 Palo Alto Networks, Inc. Joining JavaScript object notation (JSON) queries across cloud resources
CN113076198A (zh) * 2021-04-27 2021-07-06 上海商汤临港智能科技有限公司 数据处理方法及装置
CN115495206A (zh) * 2021-06-17 2022-12-20 北京小米移动软件有限公司 云端服务器、终端、信息处理方法及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110154350A1 (en) * 2009-12-18 2011-06-23 International Business Machines Corporation Automated cloud workload management in a map-reduce environment
CN102222174A (zh) * 2011-02-22 2011-10-19 深圳华大基因科技有限公司 一种基因计算系统和方法
CN102402423A (zh) * 2010-09-19 2012-04-04 百度在线网络技术（北京）有限公司 一种在网络设备中进行多任务处理的方法和设备
CN103067468A (zh) * 2012-12-22 2013-04-24 深圳先进技术研究院 云调度方法及其系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110154350A1 (en) * 2009-12-18 2011-06-23 International Business Machines Corporation Automated cloud workload management in a map-reduce environment
CN102402423A (zh) * 2010-09-19 2012-04-04 百度在线网络技术（北京）有限公司 一种在网络设备中进行多任务处理的方法和设备
CN102222174A (zh) * 2011-02-22 2011-10-19 深圳华大基因科技有限公司 一种基因计算系统和方法
CN103067468A (zh) * 2012-12-22 2013-04-24 深圳先进技术研究院 云调度方法及其系统

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114995050A (zh) * 2022-06-27 2022-09-02 上海华力集成电路制造有限公司 提高opc运算效率的方法
CN116016711A (zh) * 2022-12-05 2023-04-25 江苏鑫辰通信有限公司 通信网络调度系统及方法及系统
CN116366355A (zh) * 2023-04-14 2023-06-30 北京智享嘉网络信息技术有限公司 一种网络设备硬件资源智能调度方法及系统

Also Published As

Publication number Publication date
CN103067468B (zh) 2016-03-09
CN103067468A (zh) 2013-04-24

Similar Documents

Publication Publication Date Title
WO2014094495A1 (zh) 云调度方法及其系统
CN114020470B (zh) 资源分配方法、装置、可读介质及电子设备
Wang et al. Dynamic resource scheduling in mobile edge cloud with cloud radio access network
TWI541725B (zh) 用於管理多核心架構之資源的方法和設備
CN109766189B (zh) 集群调度方法和装置
CN101499019B (zh) 电信级以太网系统及用于该系统的实时任务调度方法
WO2012044015A2 (en) Method and apparatus for dynamic resource allocation of processing units
WO2019001092A1 (zh) 负载均衡引擎，客户端，分布式计算系统以及负载均衡方法
JP2015537307A (ja) コンポーネント指向ハイブリッドクラウドオペレーティングシステムのアーキテクチャ及びその通信方法
CA2367039A1 (en) Interstream control and communications for multi-streaming digital processors
Convolbo et al. GEODIS: towards the optimization of data locality-aware job scheduling in geo-distributed data centers
CN114780227A (zh) 一种基于芯粒化网络处理器架构的任务调度映射方法及系统
Subramanian et al. An adaptive algorithm for dynamic priority based virtual machine scheduling in cloud
WO2025025498A1 (zh) 一种负载均衡方法、装置、设备及可读存储介质
Ma et al. vLocality: Revisiting data locality for MapReduce in virtualized clouds
CN104346220B (zh) 一种任务调度方法与系统
CN118051331B (zh) 资源分配方法及系统
Meskar et al. Fair multi-resource allocation in mobile edge computing with multiple access points
Hać et al. Dynamic load balancing in a distributed system using a sender-initiated algorithm
CN110955526A (zh) 一种用于在分布式异构环境下实现多gpu调度的方法和系统
WO2013165188A1 (ko) 컴퓨팅 장치에서 수행되는 자원 모니터링 방법 및 컴퓨팅 장치
CN103297511A (zh) 高度动态环境下的客户端/服务器的调度方法和系统
CN106506594B (zh) 一种并行计算资源分配方法及装置
CN120068054A (zh) 一种安全设备管理、安全业务编排方法及系统
CN109412973B (zh) 一种音频处理方法、装置及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13864197

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13864197

Country of ref document: EP

Kind code of ref document: A1