CN103400065A - License dynamic predicting and scheduling method based on data statistics - Google Patents
License dynamic predicting and scheduling method based on data statistics Download PDFInfo
- Publication number
- CN103400065A CN103400065A CN2013102765681A CN201310276568A CN103400065A CN 103400065 A CN103400065 A CN 103400065A CN 2013102765681 A CN2013102765681 A CN 2013102765681A CN 201310276568 A CN201310276568 A CN 201310276568A CN 103400065 A CN103400065 A CN 103400065A
- Authority
- CN
- China
- Prior art keywords
- user
- license
- parallel check
- job
- priority
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Debugging And Monitoring (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a License dynamic predicting and scheduling method based on data statistics. The method comprises the steps that (a) users apply for registration, (b) operation priority is divided according to parallel nuclear numbers, (c) queuing for jobs is carried out, (d) whether idle License resources are available or not is judged, (e) whether jobs exceeding tolerant thresholds exist or not is judged, (f) jobs are selected from a job queue, and (g) software License resources and hardware resources are allocated. A dynamic prediction configuration method is adopted by the step (a) and dynamic scheduling of the jobs is achieved through the steps from (b) to (g). According to the License dynamic predicting and scheduling method, the dynamic prediction of the License resource applications under a public computing center, especially under a high performance computing platform can be carried out, floating authorized points of commercial software License can be allocated scientifically and efficiently, reasonable combination and using of the floating license points are guaranteed, and computing service quality of the public computing center can be improved effectively.
Description
Technical field
The present invention relates to a kind of License performance prediction and dispatching method of based on data statistics, in particular, relate in particular to License performance prediction and the dispatching method that a kind of sequence to the wait operation and resource are distributed the based on data statistics that has more rationality, science.
Background technology
In engineering, science, calculate many industries in field, the user need to use business software One's name is legion, expensive large-scale design and analysis.The cost mint of money is bought each software only for oneself, and this mode can spend huge cost price undoubtedly.Statistics shows, these users are not very frequent to the use of most software.
In recent years, along with the development of high-performance calculation, campus, enterprise, each province, so that the public computing center in the whole nation continues to bring out.The software and hardware resources of public computing platform all can be shared in the mode of renting by increasing enterprise, colleges and universities, scientific research institutions, eliminates buying expenses, reduces construction and the R&D costs of self.In public computing center, along with the user who rents software is more and more, a major issue that need to solve is exactly how better business software License resource reasonably to be predicted, to distribute and to manage, in order to provide calculation services efficiently and effectively for the user.
At present, the License of business software manages by FlexLM mostly.In the FlexLM administrative mechanism, Li-cense is the digital license that is created by software vendor, and it has stipulated which user can which software of legal use, and which partial function that uses software.The License Management mechanism of FlexLM, can only realize the unsteady mandate of License in LAN (Local Area Network), the performance prediction and the distribution method that lack software License, therefore can't carry out the scientific and effective distribution to user job group's License resource bid, public computing center depends merely on this mechanism and can't carry out optimal scheduling to the operation that the user submits to.
Summary of the invention
The present invention, in order to overcome the shortcoming of above-mentioned technical matters, provides a kind of sequence and the resource of waiting for operation has been distributed to License performance prediction and the dispatching method of the based on data statistics that has more rationality, science.
License performance prediction and the dispatching method of based on data statistics of the present invention, its special feature is, comprising the following steps: a). the user proposes request for utilization, the user carries out request for utilization by the parallel check figure information of filling in user name, the title of using software, software calculating, with the business software on application high-performance calculation platform; B). job priority is divided, and for user's wait operation, according to the parallel check figure of filling in user's application process what, the priority of operation is divided; More than parallel check figure, be in high priority; Parallel check figure is few, is in low priority; C). wait for the queuing of operation, by user's wait job request according to step b) in the priority determined rank, form job queue; D). judge whether available free License resource, by checking the License resource idle condition of software on computing platform, judge whether enough License resources, if there is no enough License resources, re-start judgement after waiting for one-period; If have enough License resources, perform step e); E). judge whether to have the operation that surpasses the tolerance threshold value, judge in the All Jobs formation and whether have the operation that surpasses the tolerance threshold value, if the operation that surpasses the tolerance threshold value is arranged, for this operation distributes required License resource, execution step g); If do not surpass the operation of tolerance threshold value, perform step f); F). from job queue, choosing operation, at first according to stand-by period of each operation in job queue, calculate the average latency, then choose in job queue the first stand-by period greater than the operation of average latency, and for this operation distributes needed License resource, execution step g); G). distribute hardware resource, for the operation that distributes the License resource distributes corresponding hardware resource, so that the user uses corresponding business software.
License performance prediction and the dispatching method of based on data statistics of the present invention, step a) described user applies for adopting performance prediction configuration information method, it comprises the following steps: a-1). form original template, after first user proposes the License application, take its parallel check figure of filling in as default configuration, form original template, so that after guiding, user's application information fills in; A-2). judge whether default configuration has change, when other users after first user apply for, judge that whether it have change to the default configuration of parallel check figure, if change is arranged, performs step a-3); If do not change, perform step a-4); A-3). revise template, the parallel check figure that the most repeatedly occurs in the historical application of all users of current software is assigned to current masterplate, if the parallel check figure that exists two or more occurrence numbers to equate is arranged, the parallel check figure with maximum carries out assignment to working as front template, execution step a-4); A-4). submit applications information is committed to Cluster-Based Job Management System by user's application information.
License performance prediction and the dispatching method of based on data statistics of the present invention, step b) the User Priority sequence described in, adopt following methods to realize: in user's application information, parallel check figure, greater than the user of parallel check figure in current default configuration, has high one-level job priority; In user's application information, parallel check figure equals the user of parallel check figure in current default configuration, has the intergrade job priority; In user's application information, parallel check figure, less than the user of parallel check figure in current default configuration, has low one-level job priority; Step c) wait in the queuing of operation, job queue sequentially sorts according to the job priority of high one-level, intergrade, low one-level, for the wait operation that is in same one-level, the stand-by period, long operation was arranged in the front end of short operation of stand-by period.
License performance prediction and the dispatching method of based on data of the present invention statistics, step f) in, the calculating of average latency comprises the following steps: f-1). the calculating of average latency first, and establish the number of waiting for operation and be
Individual, be respectively
,
,,
, its stand-by period is respectively
,
,,
Utilize formula (1) to calculate the average latency first
F-2). in job queue, if operation
Be assigned to software License resource, relatively this operation
Stand-by period
Whether more than or equal to the tolerance threshold value, if
More than or equal to the tolerance threshold value, perform step f-3); If
Less than the tolerance threshold value, perform step f-4); F-3). keep the average latency constant, will be over the stand-by period of tolerance threshold value
Cast out, the average latency that keeps current is constant, namely
F-4). upgrade the average latency,
Utilize formula (2) to upgrade the average latency.
The invention has the beneficial effects as follows: License performance prediction and the dispatching method of a kind of based on data statistics that the present invention proposes, can be to the application of the License resource of the user job group under public computing center, particularly high-performance calculation platform, carrying out performance prediction and scientific and efficient distributes, guarantee the unsteady reasonable combination use of counting, the calculation services quality of the public computing center of raising of authorizing of business software License.
Utilize the parallel check figure that the most repeatedly occurs in first user and historical application, configure by default, realized utilizing the performance prediction of business software resource, be conducive to the user software is chosen to best configuration.In the forming process of job queue, according to parallel check figure have many to less, stand-by period order from long to short arranges, realized waiting for operation rationally, science sorts.In the assigning process to the License resource, at first distribute to the operation that surpasses the tolerance threshold value, reallocation greater than the operation of average latency, has realized that the License resource is reasonable, science is distributed to the first stand-by period in job queue.
The accompanying drawing explanation
Fig. 1 is the task management flow process of License performance prediction of the present invention and dispatching method;
Fig. 2 is the process flow diagram of License performance prediction of the present invention and dispatching method;
Fig. 3 is the process flow diagram of performance prediction configuration in the present invention;
Fig. 4 is the calculation flow chart of averaging time in the present invention.
Embodiment
The invention will be further described below in conjunction with accompanying drawing and embodiment.
As shown in Figure 1, provided the task management flow process of License performance prediction of the present invention and dispatching method, the user uses the business software of public computing center group system deploy, the task management flow process of group system realizes by following flow process: 1. at first, the user applies for and files an application, and the content of application comprises the tricks of using what software, software, the parallel check figure that each software calculates.2. then, Cluster-Based Job Management System is submitted to software license administration module by user's application, and this module manages software license resource, and the license idling-resource is checked, distributes.3. Cluster-Based Job Management System is the required hardware resource of user assignment.4. obtained the user job that software license distributes and hardware resource distributes, waited in line operation, after end of run, by the user, downloaded result and check.
As shown in Figure 2, provided the License performance prediction of based on data statistics of the present invention and the process flow diagram of dispatching method, it is realized by following steps:
A). the user files an application, and the user carries out request for utilization by the parallel check figure information of filling in user name, the title of using software, software calculating, with the business software on application high-performance calculation platform;
In this step, for every money business software, after first user proposes the License application, using the parallel check figure applied for as initial masterplate; When the user of this software of use proposes the License application later, the default configuration using current masterplate as the user.Next a minute situation is processed: if 1. the user is dissatisfied to default configuration, can be modified by the user, according to the amended actual pairing check figure of user, submit to the License scheduler module, and revise simultaneously current masterplate.The method of based on data statistics is adopted in the modification of masterplate, while revising, the parallel check figure that the most repeatedly occurs in the historical application of all users of this software is assigned to current masterplate at every turn.
As shown in Figure 3, provided in this step the process flow diagram that adopts the performance prediction configuration to carry out user's application, it can be realized by following steps:
A-1). form original template, after first user proposes the License application, take its parallel check figure of filling in as default configuration, form original template, so that after guiding, user's application information fills in;
A-2). judge whether default configuration has change, when other users after first user apply for, judge that whether it have change to the default configuration of parallel check figure, if change is arranged, performs step a-3); If do not change, perform step a-4);
A-3). revise template, the parallel check figure that the most repeatedly occurs in the historical application of all users of current software is assigned to current masterplate, if the parallel check figure that exists two or more occurrence numbers to equate is arranged, the parallel check figure with maximum carries out assignment to working as front template, execution step a-4);
A-4). submit applications information is committed to Cluster-Based Job Management System by user's application information.
B). job priority is divided, and for user's wait operation, according to the parallel check figure of filling in user's application process what, the priority of operation is divided; More than parallel check figure, be in high priority; Parallel check figure is few, is in low priority;
At first adjust job priority, will in the performance prediction submodule, greater than the job priority of default configuration, heighten one-level by the amended parallel check figure of user, the amended parallel check figure of user is turned down to one-level lower than the job priority of default configuration.Then, all are waited for to operations are according to the renewal of ranking of priority order from high to low.
User Priority sequence described in this step, can realize by the following method: in user's application information, parallel check figure, greater than the user of parallel check figure in current default configuration, has high one-level job priority; In user's application information, parallel check figure equals the user of parallel check figure in current default configuration, has the intergrade job priority; In user's application information, parallel check figure, less than the user of parallel check figure in current default configuration, has low one-level job priority; Step c) wait in the queuing of operation, job queue sequentially sorts according to the job priority of high one-level, intergrade, low one-level, for the wait operation that is in same one-level, the stand-by period, long operation was arranged in the front end of short operation of stand-by period.
C). wait for the queuing of operation, by user's wait job request according to step b) in the priority determined rank, form job queue;
Wherein, step b) the User Priority sequence described in, adopt following methods to realize: in user's application information, parallel check figure, greater than the user of parallel check figure in current default configuration, has high one-level job priority; In user's application information, parallel check figure equals the user of parallel check figure in current default configuration, has the intergrade job priority; In user's application information, parallel check figure, less than the user of parallel check figure in current default configuration, has low one-level job priority; Step c) wait in the queuing of operation, job queue sequentially sorts according to the job priority of high one-level, intergrade, low one-level, for the wait operation that is in same one-level, the stand-by period, long operation was arranged in the front end of short operation of stand-by period.
D). judge whether available free License resource, by checking the License resource idle condition of software on computing platform, judge whether enough License resources, if there is no enough License resources, re-start judgement after waiting for one-period; If have enough License resources, perform step e);
E). judge whether to have the operation that surpasses the tolerance threshold value, judge in the All Jobs formation and whether have the operation that surpasses the tolerance threshold value, if the operation that surpasses the tolerance threshold value is arranged, for this operation distributes required License resource, execution step g); If do not surpass the operation of tolerance threshold value, perform step f);
The License resource idle condition of computing platform systems inspection software, if there is no enough idle License resources, after waiting for a clock period, reexamine the License resource idle condition of this software; If enough idle License resources are arranged, enter next step.Step then is, check the stand-by period of All Jobs, judge whether to exist the operation that surpasses the tolerance threshold value, if have directly (this step is in order to guarantee that the operation stand-by period that priority level is very low can be not long for this operation distributes required software License resource, the tolerance limit that surpasses the user, the tolerance threshold value is a time constant); If no, perform step f), by the first stand-by period in above-mentioned Queue sequence taking-up job queue, greater than the operation of average latency, be that this operation distributes required software License resource.Finally, step g) in, group system is distributed corresponding hardware resource for this operation, the submit job operation.
F). from job queue, choosing operation, at first according to stand-by period of each operation in job queue, calculate the average latency, then choose in job queue the first stand-by period greater than the operation of average latency, and for this operation distributes needed License resource, execution step g);
As shown in Figure 4, provided the process flow diagram of average latency acquiring method, it can be realized by following steps:
F-1). the calculating of average latency first, the number of establishing the wait operation is
Individual, be respectively
,
,,
, its stand-by period is respectively
,
,,
Utilize formula (1) to calculate the average latency first
F-2). in job queue, if operation
Be assigned to software License resource, relatively this operation
Stand-by period
Whether more than or equal to the tolerance threshold value, if
More than or equal to the tolerance threshold value, perform step f-3); If
Less than the tolerance threshold value, perform step f-4);
F-3). keep the average latency constant, will be over the stand-by period of tolerance threshold value
Cast out, the average latency that keeps current is constant, namely
F-4). upgrade the average latency,
Utilize formula (2) to upgrade the average latency.
G). distribute hardware resource, for the operation that distributes the License resource distributes corresponding hardware resource, so that the user uses corresponding business software.
Because the License resource pool distribution requirements of most of business softwares is counted, the non-linear minimizing along with the raising of job parallelism check figure, so License dynamic dispatching method of this material invention, appropriateness is preferentially moved to the user job that degree of parallelism is high, to reach, take full advantage of software License resource pool and count, reduce the purpose of resource fragmentation, thereby make the software License resource utilization of public computing center reach maximization.
Claims (4)
1. License performance prediction and the dispatching method of a based on data statistics, is characterized in that, comprises the following steps:
A). the user proposes request for utilization, and the user carries out request for utilization by the parallel check figure information of filling in user name, the title of using software, software calculating, with the business software on application high-performance calculation platform;
B). job priority is divided, and for user's wait operation, according to the parallel check figure of filling in user's application process what, the priority of operation is divided; More than parallel check figure, be in high priority; Parallel check figure is few, is in low priority;
C). wait for the queuing of operation, by user's wait job request according to step b) in the priority determined rank, form job queue;
D). judge whether available free License resource, by checking the License resource idle condition of software on computing platform, judge whether enough License resources, if there is no enough License resources, re-start judgement after waiting for one-period; If have enough License resources, perform step e);
E). judge whether to have the operation that surpasses the tolerance threshold value, judge in the All Jobs formation and whether have the operation that surpasses the tolerance threshold value, if the operation that surpasses the tolerance threshold value is arranged, for this operation distributes required License resource, execution step g); If do not surpass the operation of tolerance threshold value, perform step f);
F). from job queue, choosing operation, at first according to stand-by period of each operation in job queue, calculate the average latency, then choose in job queue the first stand-by period greater than the operation of average latency, and for this operation distributes needed License resource, execution step g);
G). distribute hardware resource, for the operation that distributes the License resource distributes corresponding hardware resource, so that the user uses corresponding business software.
2. License performance prediction and the dispatching method of based on data statistics according to claim 1, is characterized in that, step a) described user's request for utilization adopts performance prediction configuration information method, and it comprises the following steps:
A-1). form original template, after first user proposes the License application, take its parallel check figure of filling in as default configuration, form original template, so that after guiding, user's application information fills in;
A-2). judge whether default configuration has change, when other users after first user apply for, judge that whether it have change to the default configuration of parallel check figure, if change is arranged, performs step a-3); If do not change, perform step a-4);
A-3). revise template, the parallel check figure that the most repeatedly occurs in the historical application of all users of current software is assigned to current masterplate, if the parallel check figure that exists two or more occurrence numbers to equate is arranged, the parallel check figure with maximum carries out assignment to working as front template, execution step a-4);
A-4). submit applications information is committed to Cluster-Based Job Management System by user's application information.
3. License performance prediction and the dispatching method of based on data according to claim 2 statistics, it is characterized in that, step b) the User Priority sequence described in, adopt following methods to realize: in user's application information, parallel check figure, greater than the user of parallel check figure in current default configuration, has high one-level job priority; In user's application information, parallel check figure equals the user of parallel check figure in current default configuration, has the intergrade job priority; In user's application information, parallel check figure, less than the user of parallel check figure in current default configuration, has low one-level job priority; Step c) wait in the queuing of operation, job queue sequentially sorts according to the job priority of high one-level, intergrade, low one-level, for the wait operation that is in same one-level, the stand-by period, long operation was arranged in the front end of short operation of stand-by period.
4. License performance prediction and the dispatching method of based on data according to claim 1 and 2 statistics, is characterized in that step f) in the calculating of average latency comprise the following steps:
F-1). the calculating of average latency first, the number of establishing the wait operation is
Individual, be respectively
,
,,
, its stand-by period is respectively
,
,,
Utilize formula (1) to calculate the average latency first
F-2). in job queue, if operation
Be assigned to software License resource, relatively this operation
Stand-by period
Whether more than or equal to the tolerance threshold value, if
More than or equal to the tolerance threshold value, perform step f-3); If
Less than the tolerance threshold value, perform step f-4);
F-3). keep the average latency constant, will be over the stand-by period of tolerance threshold value
Cast out, the average latency that keeps current is constant, namely
F-4). upgrade the average latency,
Utilize formula (2) to upgrade the average latency.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310276568.1A CN103400065B (en) | 2013-07-03 | 2013-07-03 | A kind of License performance prediction based on data statistics and dispatching method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310276568.1A CN103400065B (en) | 2013-07-03 | 2013-07-03 | A kind of License performance prediction based on data statistics and dispatching method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103400065A true CN103400065A (en) | 2013-11-20 |
| CN103400065B CN103400065B (en) | 2016-01-20 |
Family
ID=49563689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310276568.1A Expired - Fee Related CN103400065B (en) | 2013-07-03 | 2013-07-03 | A kind of License performance prediction based on data statistics and dispatching method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103400065B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391262A (en) * | 2017-07-31 | 2017-11-24 | 李黎黎 | A kind of job scheduling method and device |
| CN107391254A (en) * | 2017-06-14 | 2017-11-24 | 深圳天珑无线科技有限公司 | Intelligent terminal and its resource allocation method, computer-readable recording medium |
| CN113032143A (en) * | 2021-03-09 | 2021-06-25 | 西北工业大学 | Intelligent tuning method for high-performance calculation of user quota |
| CN113282381A (en) * | 2020-02-19 | 2021-08-20 | 中科寒武纪科技股份有限公司 | Task scheduling method and device, computer equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101694631A (en) * | 2009-09-30 | 2010-04-14 | 曙光信息产业(北京)有限公司 | Real-time operation dispatching system and method thereof |
| CN102130837A (en) * | 2011-03-31 | 2011-07-20 | 中兴通讯股份有限公司 | Method and system for dynamically regulating service license |
| CN102402461A (en) * | 2011-12-08 | 2012-04-04 | 曙光信息产业(北京)有限公司 | Balanced scheduling method based on operation scale |
| CN102541650A (en) * | 2011-12-02 | 2012-07-04 | 曙光信息产业(北京)有限公司 | Dynamic license scheduling method based on historical statistics |
| US20130074189A1 (en) * | 2011-09-17 | 2013-03-21 | International Business Machines Corporation | Software license reconciliation within a cloud computing infrastructure |
| US20130091282A1 (en) * | 2011-10-06 | 2013-04-11 | Channarong Tontiruttananon | On-demand integrated capacity and reliability service level agreement licensing |
-
2013
- 2013-07-03 CN CN201310276568.1A patent/CN103400065B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101694631A (en) * | 2009-09-30 | 2010-04-14 | 曙光信息产业(北京)有限公司 | Real-time operation dispatching system and method thereof |
| CN102130837A (en) * | 2011-03-31 | 2011-07-20 | 中兴通讯股份有限公司 | Method and system for dynamically regulating service license |
| US20130074189A1 (en) * | 2011-09-17 | 2013-03-21 | International Business Machines Corporation | Software license reconciliation within a cloud computing infrastructure |
| US20130091282A1 (en) * | 2011-10-06 | 2013-04-11 | Channarong Tontiruttananon | On-demand integrated capacity and reliability service level agreement licensing |
| CN102541650A (en) * | 2011-12-02 | 2012-07-04 | 曙光信息产业(北京)有限公司 | Dynamic license scheduling method based on historical statistics |
| CN102402461A (en) * | 2011-12-08 | 2012-04-04 | 曙光信息产业(北京)有限公司 | Balanced scheduling method based on operation scale |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391254A (en) * | 2017-06-14 | 2017-11-24 | 深圳天珑无线科技有限公司 | Intelligent terminal and its resource allocation method, computer-readable recording medium |
| CN107391262A (en) * | 2017-07-31 | 2017-11-24 | 李黎黎 | A kind of job scheduling method and device |
| CN107391262B (en) * | 2017-07-31 | 2020-05-15 | 李黎黎 | Job scheduling method and device |
| CN113282381A (en) * | 2020-02-19 | 2021-08-20 | 中科寒武纪科技股份有限公司 | Task scheduling method and device, computer equipment and storage medium |
| CN113032143A (en) * | 2021-03-09 | 2021-06-25 | 西北工业大学 | Intelligent tuning method for high-performance calculation of user quota |
| CN113032143B (en) * | 2021-03-09 | 2022-10-18 | 西北工业大学 | Intelligent tuning method for high-performance calculation of user quota |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103400065B (en) | 2016-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8583799B2 (en) | Dynamic cost model based resource scheduling in distributed compute farms | |
| US20190324819A1 (en) | Distributed-system task assignment method and apparatus | |
| Jin et al. | Towards optimized fine-grained pricing of IaaS cloud platform | |
| US20190095249A1 (en) | System, method, and medium for facilitating auction-based resource sharing for message queues in an on-demand services environment | |
| Masdari et al. | Efficient task and workflow scheduling in inter-cloud environments: challenges and opportunities | |
| Li et al. | Pricing and repurchasing for big data processing in multi-clouds | |
| CN1306754C (en) | Method and system for balancing working load in network computing environment | |
| Zeng et al. | An integrated task computation and data management scheduling strategy for workflow applications in cloud environments | |
| CN106020927B (en) | Task scheduling and the universal method of resource distribution in a kind of cloud computing system | |
| Huang et al. | When backpressure meets predictive scheduling | |
| US20150046279A1 (en) | Auction-based resource sharing for message queues in an on-demand services environment | |
| CN111480145A (en) | System and method for scheduling workloads according to a credit-based mechanism | |
| US10474502B2 (en) | Multi-tenant license enforcement across job requests | |
| Zhao et al. | Multi-resource interleaving for deep learning training | |
| US20160127382A1 (en) | Determining variable wait time in an asynchronous call-back system based on calculated average sub-queue wait time | |
| CN107317836A (en) | One kind mixing cloud environment lower time appreciable request scheduling method | |
| CN103400065A (en) | License dynamic predicting and scheduling method based on data statistics | |
| Xu et al. | Fault tolerance and quality of service aware virtual machine scheduling algorithm in cloud data centers | |
| US20190377596A1 (en) | Flexible batch job scheduling in virtualization environments | |
| Ohmura et al. | Toward building a digital twin of job scheduling and power management on an HPC system | |
| Peng et al. | Joint operator scaling and placement for distributed stream processing applications in edge computing | |
| Mishra et al. | A memory-aware dynamic job scheduling model in Grid computing | |
| Awasare et al. | Survey and comparative study on resource allocation strategies in cloud computing environment | |
| Zhang et al. | Towards reliable (and efficient) job executions in a practical geo-distributed data analytics system | |
| CN103294535A (en) | Apparatus and method for maximizing disk cache effect for workflow job scheduling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160120 Termination date: 20190703 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |