CN109086127B - Resource scheduling method based on FSM control and framework system - Google Patents
Resource scheduling method based on FSM control and framework system Download PDFInfo
- Publication number
- CN109086127B CN109086127B CN201810829551.7A CN201810829551A CN109086127B CN 109086127 B CN109086127 B CN 109086127B CN 201810829551 A CN201810829551 A CN 201810829551A CN 109086127 B CN109086127 B CN 109086127B
- Authority
- CN
- China
- Prior art keywords
- scheduling
- fsm
- scheduler
- finite state
- resource
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 230000007246 mechanism Effects 0.000 claims description 81
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 235000019580 granularity Nutrition 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/485—Task life-cycle, e.g. stopping, restarting, resuming execution
- G06F9/4856—Task life-cycle, e.g. stopping, restarting, resuming execution resumption being on a different machine, e.g. task migration, virtual machine migration
- G06F9/4862—Task life-cycle, e.g. stopping, restarting, resuming execution resumption being on a different machine, e.g. task migration, virtual machine migration the task being a mobile agent, i.e. specifically designed to migrate
- G06F9/4875—Task life-cycle, e.g. stopping, restarting, resuming execution resumption being on a different machine, e.g. task migration, virtual machine migration the task being a mobile agent, i.e. specifically designed to migrate with migration policy, e.g. auction, contract negotiation
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 Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a resource scheduling method and a frame system based on FSM control, which respectively receive scheduling events of scheduling strategies sent from the outside in real time, based on the corresponding relation between each preset scheduling strategy scheduling event and each FSM state of the finite state automaton, the finite state automaton correspondingly transfers the current FSM state to each corresponding FSM state corresponding to the scheduling event of the scheduling policy received from the outside, based on the output result of each FSM state of the finite state automaton which is established in advance and the corresponding relation of each scheduling strategy which is preset in the multi-mechanism scheduler, correspondingly determining a target scheduling strategy corresponding to a corresponding FSM state to which the finite state automaton is currently migrated through the multi-mechanism scheduler, and scheduling the resources in the resource pool respectively through the multi-mechanism scheduler according to the corresponding determined target scheduling strategies. The invention can meet the requirements of various scheduling scenes of enterprises in real time and increase the flexibility of resource scheduling.
Description
Technical Field
The invention relates to the field of resource scheduling, in particular to a resource scheduling method and a frame system based on FSM control, which are mainly used for resource scheduling of a big data computing platform.
Background
In the era of today's explosion of big data technology, the management of large amounts of storage and computing resources has become especially important. In order to achieve efficient utilization of resources, or to achieve low latency of time-sensitive computation, various resource scheduling schemes have been proposed in the prior art, such as the YARN resource scheduling framework, the facilities resource scheduling framework, and the kubernets resource scheduling framework, which are commonly used in the prior art, and these scheduling schemes provide relatively optimized solutions in most scenarios, but these scheduling schemes either have a problem that the scheduling scheme of the scheduler is single, or have a problem that the scheduler is not flexible due to small dynamic change space.
For example, the YARN resource scheduling framework is based on a Hadoop platform, and is used for a resource management framework of a big data computing platform, and the basic idea of the framework is as follows: the resources can be divided into containers one by one according to hardware such as memory/cpu/hard disk/network card, etc., or divided according to tags; and then the application obtains the use right of the resource by applying for the container of the fixed unit, and returns the resource after the application execution is completed. To fit different scenarios, the YARN framework provides the following mechanisms:
1) a reservation mechanism: reserving a container for the application with large resource demand, wherein the container is used for avoiding that the application with large resource demand cannot obtain resources all the time;
2) the preemption mechanism is as follows: the method comprises the steps of enabling a time-sensitive application to preempt resources from other applications when available resources are insufficient;
3) the relative fairness mechanism: a systematic weighting algorithm is employed to ensure relatively fair utilization of resources by applications.
4) The sharing mechanism comprises the following steps: the calling interface is used for providing idle resources, and a user can conveniently call each idle resource.
It can be seen that the YARN resource scheduling framework actually provides a series of mechanisms to meet the requirements in different scenarios, and through these mechanisms and their combination, a variety of scheduling strategies can be implemented, but still has the following disadvantages:
the preemption mechanism is good for time-sensitive application, but the mechanism mode is global and can only be started when the system is started, and cannot be dynamically switched;
under the condition that a plurality of departments in an enterprise use resources in a mixed mode, different scheduling strategies are often needed among different departments, the granularity of YARN scheduling is often particularly fine, and the use flexibility is relatively weak;
③ the plug-in scheduler of the YARN resource scheduling framework, although it increases the flexibility of the YARN framework to some extent, it also needs to be enabled at system start-up and cannot be changed dynamically.
Moreover, the Mesos resource scheduling framework and the kubernets resource scheduling framework are independent, both can manage virtual resources, and do not provide a complex scheduling mechanism, and the Mesos resource scheduling framework and the kubernets resource scheduling framework only divide resources into fine granularities and then externally provide an interface of 'how many XX resources are available', and directly divide the resource requirements of applications into applications for management. It can be seen that both the mess resource scheduling framework and the kubernets resource scheduling framework have fewer scheduling mechanisms provided and are essentially laborious for complex scheduling management.
Therefore, the invention provides a resource scheduling method based on FSM control and a frame system, which are used for solving the technical problems in the prior art.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a resource scheduling method and a frame system based on FSM control aiming at the defects of the prior art, which are used for meeting the requirements of various scheduling scenes of enterprises and increasing the flexibility of resource scheduling.
In order to solve the technical problem, the invention provides a resource scheduling method based on FSM control, which comprises the following steps:
s1, receiving scheduling policy scheduling events sent from outside in real time, and sending the received scheduling policy scheduling events to a finite state automaton (FSM) in real time;
s2, storing the current FSM state in real time through the finite state automata, receiving each scheduling policy scheduling event sent in the step S1 in real time, and correspondingly transferring the current FSM state to each corresponding FSM state corresponding to the scheduling policy scheduling event sent in the step S1 based on the corresponding relation between each scheduling policy scheduling event and each FSM state of the finite state automata which is preset;
s3, acquiring output results corresponding to the corresponding FSM states to which the finite state automaton is currently migrated in the step S2 in real time through the multi-mechanism scheduler, correspondingly determining target scheduling strategies corresponding to the corresponding FSM states to which the finite state automaton is currently migrated in the step S2 according to the corresponding relationship between the output results of the FSM states of the finite state automaton established in advance and the scheduling strategies set in the multi-mechanism scheduler in advance, and scheduling resources in the resource pool according to the determined target scheduling strategies.
Wherein, each scheduling policy preset in the multi-mechanism scheduler in step S3 is formed by the multi-mechanism scheduler correspondingly according to at least one resource scheduling mechanism in each resource scheduling mechanism preset in the multi-mechanism scheduler; the resource scheduling mechanism at least comprises any two mechanisms of a sharing mechanism, a preempting mechanism and a reserving mechanism.
The resource scheduling mechanism comprises a sharing mechanism, a preempting mechanism and a reserving mechanism.
In step S1, each scheduling policy scheduling event received in real time from the outside is defined by an xml file.
In step S3, each scheduling policy preset in the multi-mechanism scheduler is defined by an xml file.
In addition, the invention also provides a resource scheduling framework system based on FSM control, which comprises an event receiver, a finite state automaton and a multi-mechanism scheduler, wherein:
the event receiver is used for receiving the scheduling strategy scheduling event sent from the outside in real time and sending the received scheduling strategy scheduling event to the finite state automaton in real time;
the finite state automaton is used for storing the current FSM state of the finite state automaton in real time, receiving scheduling strategy scheduling events sent by the event receiver in real time, and correspondingly transferring the FSM state of the finite state automaton to corresponding FSM states corresponding to the received scheduling strategy scheduling events based on the preset corresponding relation between the scheduling strategy scheduling events and the FSM states of the finite state automaton;
and the multi-mechanism scheduler is connected with the finite state automata and the resource pool and is used for correspondingly determining a target scheduling strategy corresponding to the corresponding FSM state to which the finite state automata is currently migrated based on the output result corresponding to the corresponding FSM state to which the finite state automata is currently migrated and the corresponding relationship between the output result of each FSM state of the finite state automata pre-established in the multi-mechanism scheduler and each scheduling strategy pre-set in the multi-mechanism scheduler, and scheduling the resources in the resource pool according to the determined target scheduling strategy.
The scheduling strategies preset in the multi-mechanism scheduler are correspondingly formed by the multi-mechanism scheduler according to at least one resource scheduling mechanism in the preset resource scheduling mechanisms; the resource scheduling mechanism at least comprises any two mechanisms of a sharing mechanism, a preempting mechanism and a reserving mechanism.
The resource scheduling mechanism comprises a sharing mechanism, a preempting mechanism and a reserving mechanism.
And each related scheduling strategy scheduling event received by the event receiver and sent from the outside is defined by an xml file.
Wherein, each scheduling strategy preset in the multi-mechanism scheduler is defined by an xml file.
Compared with the prior art, the invention has the advantages that:
(1) the resource scheduling method and the frame system based on FSM control respectively receive scheduling events of scheduling strategies sent from the outside in real time, correspondingly migrate the current FSM state of the finite state automata to the corresponding FSM states of the scheduling strategies scheduling events sent from the outside received by the finite state automata based on the corresponding relation between the scheduling events of each scheduling strategy preset by the finite state automata and the FSM states of the finite state automata, correspondingly determine target scheduling strategies corresponding to the corresponding FSM states to which the finite state automata is currently migrated by the multi-mechanism scheduler based on the corresponding relation between the output results of the FSM states of the finite state automata and the scheduling strategies preset in the multi-mechanism scheduler, respectively schedule resources in a resource pool by the multi-mechanism scheduler according to the determined target scheduling strategies, the invention is based on the strategy management of the finite state automata, the change of the target scheduling strategy is driven by the scheduling event of the scheduling strategy, and the scheduling events of different scheduling strategies correspond to different target scheduling strategies, so that the invention can meet the requirements of various scheduling scenes of an enterprise in real time, can solve the problem that the dynamic change space of the existing scheduler is small to a certain extent, can solve the problem that the scheduling mode of the existing scheduler is single to a certain extent, and increases the flexibility of resource scheduling.
(2) According to the resource scheduling method and the frame system based on FSM control, each scheduling strategy preset in the multi-mechanism scheduler is defined by the xml file, so that a user can conveniently change the scheduling strategies according to actual needs.
Therefore, compared with the prior art, the invention has prominent substantive features and remarkable progress, and the beneficial effects of the implementation are also obvious.
Drawings
Fig. 1 is a schematic flow chart of a resource scheduling method based on FSM control according to the present invention.
Fig. 2 is a schematic diagram of an architecture of a resource scheduling framework system based on FSM control according to the present invention.
Fig. 3 is an electrical schematic block diagram of a resource scheduling framework system based on FSM control according to the present invention.
Wherein: 1. event receiver, 2 finite state automata, 3 multi-mechanism scheduler, 4 resource pool.
Detailed Description
In order to make the technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings.
See fig. 1. Fig. 1 is a specific embodiment of a resource scheduling method based on FSM control according to the present invention. In this embodiment, the FSM control-based resource scheduling method includes the following steps S1-S3, which are used to enable the resource scheduling method to meet the needs of various scheduling scenarios of an enterprise in real time, to solve the problem that the dynamic change space of the existing scheduler is small to a certain extent, and to solve the problem that the scheduling mode of the existing scheduler is single to a certain extent, thereby increasing the flexibility of the resource scheduling method used in the present invention.
And S1, receiving the scheduling policy scheduling event sent from the outside in real time, and sending the received scheduling policy scheduling event to the finite state automaton 2 in real time.
In this embodiment, each scheduling policy scheduling event received in real time in step S1 and transmitted from the outside is defined by an xml file.
S2, storing the current FSM state of the finite state automata 2 in real time, receiving each scheduling policy scheduling event sent in step S1 in real time, and correspondingly migrating the current FSM state to each corresponding FSM state corresponding to the scheduling policy scheduling event sent in step S1, based on the preset corresponding relationship between each scheduling policy scheduling event and each FSM state of the finite state automata 2.
S3, obtaining in real time, by the multi-mechanism scheduler 3, the output result corresponding to the corresponding FSM state to which the finite state automaton 2 currently migrates in step S2, and correspondingly determining the target scheduling policy corresponding to the corresponding FSM state to which the finite state automaton 2 currently migrates in step S2 according to the correspondence between the output result of each FSM state of the finite state automaton 2 and each scheduling policy preset in the multi-mechanism scheduler 3, and scheduling the resources in the resource pool 4 according to the determined target scheduling policy.
In this embodiment, each scheduling policy preset in the multi-mechanism scheduler 3 in step S3 is formed by the multi-mechanism scheduler 3 according to at least one of the preset resource scheduling mechanisms; the resource scheduling mechanism comprises a sharing mechanism, a preempting mechanism and a reserving mechanism.
In the present embodiment, each scheduling policy set in advance in the multi-mechanism scheduler 3 described in step S3 is defined by an xml file.
Before use, pre-defining each required scheduling policy by an xml file in the multi-mechanism scheduler 3, for example, pre-defining each preset resource scheduling mechanism in the multi-mechanism scheduler 3 into an independent scheduling policy by an xml file in the multi-mechanism scheduler 3, and/or defining any combination of each preset resource scheduling mechanism in the multi-mechanism scheduler 3 into a corresponding scheduling policy (pre-defining the above xml file in the multi-mechanism scheduler 3); defining character strings corresponding to scheduling events of each scheduling strategy related by the invention in advance through another xml file; then, the conversion rules of the FSM are preset, that is, the corresponding relationship between each scheduling policy scheduling event and each FSM state of the finite state automata 2 is preset, and the corresponding relationship between the output result of each FSM state of the finite state automata 2 and each scheduling policy preset in the multi-mechanism scheduler 3 through the corresponding XML file is preset.
When in use, the scheduling policy scheduling event sent from the outside is received in real time (the scheduling policy scheduling event is the same as the corresponding event in each scheduling policy scheduling event defined by the corresponding xml file in advance), and the received scheduling policy scheduling event is sent to the finite state automaton 2 in real time; the finite state automaton 2 stores the current FSM state of the finite state automaton in real time, receives each scheduling policy scheduling event sent from the outside in real time, and correspondingly transfers each current FSM state to each corresponding FSM state corresponding to each scheduling policy scheduling event received by the finite state automaton 2 based on the preset corresponding relation between each scheduling policy scheduling event and each FSM state of the finite state automaton; the multi-mechanism scheduler 3 acquires output results corresponding to the corresponding FSM states to which the finite state automata 2 is currently migrated in real time, correspondingly determines a target scheduling policy corresponding to the corresponding FSM states to which the finite state automata 2 is currently migrated according to a pre-established correspondence between the output results of the FSM states of the finite state automata 2 and scheduling policies preset in the multi-mechanism scheduler 3, and schedules resources in the resource pool 4 according to the determined target scheduling policy.
It should be noted that, in this embodiment, the multi-mechanism scheduler 3 is based on the scheduler built in the YARN resource scheduling framework (but this is not essential, and those skilled in the art can also implement this by other manners in the prior art), and the scheduling mechanisms (including the preemption mechanism, the sharing mechanism, and the reservation mechanism, which are original scheduling mechanisms in the existing YARN resource scheduling framework) preset in the scheduler are mutually independent and can be assembled and combined for use, so that a user can customize, through an XML file built in the user, each corresponding scheduling policy required by the user (for example, each scheduling mechanism is assembled and combined into one scheduling policy), and the user can meet the requirements of various scheduling scenarios in real time. .
In summary, the FSM control-based resource scheduling method of the present invention facilitates a user to change a scheduling policy according to actual needs, and thus the present invention can meet the needs of various scheduling scenarios of an enterprise in real time, and solves the problem of single scheduling mode of the existing scheduler to a certain extent, thereby solving the problem of small dynamic change space of the existing scheduler to a certain extent, and increasing the flexibility of the present invention in use.
Corresponding to the resource scheduling method based on FSM control, the present invention further provides a resource scheduling framework system based on FSM control, and fig. 2 and 3 are specific embodiments of the resource scheduling framework system based on FSM control.
Referring to fig. 2 and fig. 3, the FSM control-based resource scheduling framework system in this embodiment includes an event receiver 1, a finite state automaton 2 and a multi-mechanism scheduler 3, wherein the event receiver 1 is connected to a resource pool 4 through the finite state automaton 2 and the multi-mechanism scheduler 3 in sequence, specifically:
the event receiver 1 is used for receiving scheduling strategy scheduling events sent from the outside in real time and sending the received scheduling strategy scheduling events to the finite state automaton 2 in real time;
the finite state automaton 2 is used for storing the current FSM state of the finite state automaton 2 in real time, receiving scheduling strategy scheduling events sent by the event receiver 1 in real time, and correspondingly transferring the FSM state of the finite state automaton to corresponding FSM states corresponding to the received scheduling strategy scheduling events based on the preset corresponding relation between the scheduling strategy scheduling events and the FSM states of the finite state automaton 2;
the multi-mechanism scheduler 3 is connected with the finite state automata 2 and the resource pool 4, and is used for correspondingly determining a target scheduling policy corresponding to the corresponding FSM state to which the finite state automata 2 is currently migrated based on an output result corresponding to the corresponding FSM state to which the finite state automata 2 is currently migrated and a correspondence between the output result of each FSM state of the finite state automata 2 pre-established in the output result and each scheduling policy pre-set in the multi-mechanism scheduler 3, and scheduling resources in the resource pool 4 according to the determined target scheduling policy.
Each scheduling policy preset in the multi-mechanism scheduler 3 is formed by the multi-mechanism scheduler 3 correspondingly according to at least one resource scheduling mechanism in each resource scheduling mechanism preset in the multi-mechanism scheduler 3; the resource scheduling mechanism comprises a sharing mechanism, a preemption mechanism and a reservation mechanism.
In order to increase the flexibility of the present invention, in the present embodiment, each scheduling event of the related scheduling policies received by the event receiver 1 and sent from the outside is defined by the corresponding xml file therein, and each scheduling policy preset in the multi-mechanism scheduler 3 is defined by the corresponding xml file therein, so that the user can change the scheduling policy according to the actual need.
Corresponding to the resource scheduling method based on FSM control, the resource scheduling framework system based on FSM control is characterized in that before use: predefining each required scheduling policy by an xml file in the multi-mechanism scheduler 3, for example, predefining each preset resource scheduling mechanism in the multi-mechanism scheduler 3 as an independent scheduling policy by the xml file in the multi-mechanism scheduler 3, and/or defining any combination of each preset resource scheduling mechanism in the multi-mechanism scheduler 3 as a corresponding scheduling policy (by the xml file in the multi-mechanism scheduler 3); defining character strings corresponding to scheduling events of scheduling strategies related by the invention in advance through an xml file in the event receiver 1; then, the conversion rules of the FSM are preset, that is, the corresponding relationship between each scheduling policy scheduling event and each FSM state of the finite state automata 2 is preset, and the corresponding relationship between the output result of each FSM state of the finite state automata 2 and each scheduling policy preset in the multi-mechanism scheduler 3 through the corresponding XML file is preset.
When the resource scheduling framework system based on FSM control is used, the event receiver 1 receives and sends the received scheduling policy scheduling events (the same as the corresponding events in the scheduling events of the scheduling policies defined by the corresponding xml files in advance) sent from the outside to the finite state automaton 2 in real time; the finite state automaton 2 stores the current FSM state of the finite state automaton in real time, receives each scheduling policy scheduling event sent from the outside in real time, and correspondingly transfers each current FSM state to each corresponding FSM state corresponding to each scheduling policy scheduling event received by the finite state automaton 2 based on the preset corresponding relation between each scheduling policy scheduling event and each FSM state of the finite state automaton; the multi-mechanism scheduler 3 acquires output results corresponding to the corresponding FSM states to which the finite state automata 2 is currently migrated in real time, correspondingly determines a target scheduling policy corresponding to the corresponding FSM states to which the finite state automata 2 is currently migrated according to a pre-established correspondence between the output results of the FSM states of the finite state automata 2 and scheduling policies preset in the multi-mechanism scheduler 3, and schedules resources in the resource pool 4 according to the determined target scheduling policy. Therefore, the invention is based on the strategy management of the finite state automaton 2, the change of the target scheduling strategy is driven by the scheduling event of the scheduling strategy, and the scheduling events of different scheduling strategies correspond to different target scheduling strategies, so that the invention can meet the requirements of various scheduling scenes of an enterprise in real time, the problem of single scheduling mode of the existing scheduler is solved to a certain extent, the problem of small dynamic change space of the existing scheduler is solved to a certain extent, and the use flexibility of the invention is increased.
It should also be noted that the multi-mechanism scheduler 3 in the present system is based on a scheduler built in the existing YARN resource scheduling framework (but this is not essential, and those skilled in the art can also implement this by other means in the prior art), and each scheduling mechanism (including the preemption mechanism, the sharing mechanism, and the reservation mechanism, which are the original scheduling mechanisms in the existing YARN resource scheduling framework on which the present invention is based) preset in the scheduler are mutually independent and can be assembled and combined for use, so that a user can customize each required corresponding scheduling policy (for example, assemble each scheduling mechanism into one scheduling policy) by using its built-in XML file, and the user can meet the requirements of various scheduling scenarios in real time.
The above embodiments are only for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may be modified or some technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A resource scheduling method based on FSM control is characterized by comprising the following steps:
s1, receiving scheduling strategy scheduling events sent from the outside in real time, and sending the received scheduling strategy scheduling events to the finite state automaton (2) in real time;
s2, storing the current FSM state in real time through the finite state automaton (2), receiving each scheduling policy scheduling event sent in the step S1 in real time, and correspondingly migrating the current FSM state to each corresponding FSM state corresponding to the scheduling policy scheduling event sent in the step S1 received by the finite state automaton (2) on the basis of the preset corresponding relation between each scheduling policy scheduling event and each FSM state of the finite state automaton (2);
s3, acquiring output results corresponding to the corresponding FSM states to which the finite state automaton (2) is currently migrated in the step S2 in real time through the multi-mechanism scheduler (3), correspondingly determining target scheduling strategies corresponding to the corresponding FSM states to which the finite state automaton (2) is currently migrated in the step S2 according to the corresponding relationship between the output results of the FSM states of the finite state automaton (2) established in advance and the scheduling strategies preset in the multi-mechanism scheduler (3), and scheduling resources in the resource pool (4) according to the determined target scheduling strategies.
2. The FSM control-based resource scheduling method of claim 1, wherein the scheduling policies pre-configured in the multi-mechanism scheduler (3) in step S3 are formed by the multi-mechanism scheduler (3) according to at least one of the pre-configured resource scheduling mechanisms therein; the resource scheduling mechanism at least comprises any two mechanisms of a sharing mechanism, a preempting mechanism and a reserving mechanism.
3. The FSM control-based resource scheduling method of claim 2, wherein the resource scheduling mechanisms comprise a sharing mechanism, a preemption mechanism, and a reservation mechanism.
4. A FSM control-based resource scheduling method according to any of claims 1-3, wherein each scheduling policy scheduling event received in real time in step S1 from outside is defined by an xml file.
5. The FSM control-based resource scheduling method of claim 4, wherein each scheduling policy preset in the multi-mechanism scheduler (3) in step S3 is defined by an xml file.
6. A FSM control based resource scheduling framework system comprising an event receiver (1), a finite state automaton (2) and a multi-mechanism scheduler (3), wherein:
the event receiver (1) is used for receiving scheduling strategy scheduling events sent from the outside in real time and sending the received scheduling strategy scheduling events to the finite state automaton (2) in real time;
the finite state automaton (2) is used for storing the current FSM state of the finite state automaton in real time, receiving scheduling strategy scheduling events sent by the event receiver (1) in real time, and correspondingly transferring the FSM state of the finite state automaton to corresponding FSM states corresponding to the received scheduling strategy scheduling events based on the preset corresponding relation between the scheduling strategy scheduling events and the FSM states of the finite state automaton (2);
and the multi-mechanism scheduler (3) is connected with the finite state automata (2) and the resource pool (4) and is used for correspondingly determining a target scheduling strategy corresponding to the corresponding FSM state to which the finite state automata (2) is currently migrated based on the output result corresponding to the corresponding FSM state to which the finite state automata (2) is currently migrated, the output result of each FSM state of the finite state automata (2) pre-established in the multi-mechanism scheduler (3) and the corresponding relation of each scheduling strategy pre-set in the multi-mechanism scheduler (3), and scheduling the resources in the resource pool (4) according to the determined target scheduling strategy.
7. The resource scheduling framework system based on FSM control of claim 6, wherein each scheduling policy preset in the multi-mechanism scheduler (3) is formed by the multi-mechanism scheduler (3) according to at least one resource scheduling mechanism of the resource scheduling mechanisms preset in the multi-mechanism scheduler; the resource scheduling mechanism at least comprises any two mechanisms of a sharing mechanism, a preempting mechanism and a reserving mechanism.
8. The FSM control-based resource scheduling framework system of claim 7, wherein the resource scheduling mechanisms comprise a sharing mechanism, a preemption mechanism, and a reservation mechanism.
9. A FSM control-based resource scheduling framework system according to any of claims 6-8, characterized in that each externally-sent relevant scheduling policy scheduling event received by the event receiver (1) is defined by an xml file.
10. A resource scheduling framework system based on FSM control according to any of claims 6-8, characterized in that each scheduling policy preset in the multi-mechanism scheduler (3) is defined by an xml-file.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810829551.7A CN109086127B (en) | 2018-07-25 | 2018-07-25 | Resource scheduling method based on FSM control and framework system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810829551.7A CN109086127B (en) | 2018-07-25 | 2018-07-25 | Resource scheduling method based on FSM control and framework system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109086127A CN109086127A (en) | 2018-12-25 |
| CN109086127B true CN109086127B (en) | 2021-11-26 |
Family
ID=64838660
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810829551.7A Active CN109086127B (en) | 2018-07-25 | 2018-07-25 | Resource scheduling method based on FSM control and framework system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109086127B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986272A (en) * | 2010-11-05 | 2011-03-16 | 北京大学 | Task scheduling method under cloud computing environment |
| CN103019808A (en) * | 2011-09-28 | 2013-04-03 | 中国移动通信集团公司 | Business state regulating device and method |
| US8424003B2 (en) * | 2006-05-31 | 2013-04-16 | International Business Machines Corporation | Unified job processing of interdependent heterogeneous tasks using finite state machine job control flow based on identified job type |
| CN103309799A (en) * | 2012-03-13 | 2013-09-18 | 百度在线网络技术（北京）有限公司 | Finite-state machine-based scheduling test methods, systems and devices |
| US8917738B2 (en) * | 2009-04-27 | 2014-12-23 | Lsi Corporation | Multicasting traffic manager in a network communications processor architecture |
-
2018
- 2018-07-25 CN CN201810829551.7A patent/CN109086127B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8424003B2 (en) * | 2006-05-31 | 2013-04-16 | International Business Machines Corporation | Unified job processing of interdependent heterogeneous tasks using finite state machine job control flow based on identified job type |
| US8917738B2 (en) * | 2009-04-27 | 2014-12-23 | Lsi Corporation | Multicasting traffic manager in a network communications processor architecture |
| CN101986272A (en) * | 2010-11-05 | 2011-03-16 | 北京大学 | Task scheduling method under cloud computing environment |
| CN103019808A (en) * | 2011-09-28 | 2013-04-03 | 中国移动通信集团公司 | Business state regulating device and method |
| CN103309799A (en) * | 2012-03-13 | 2013-09-18 | 百度在线网络技术（北京）有限公司 | Finite-state machine-based scheduling test methods, systems and devices |
Non-Patent Citations (2)
| Title |
|---|
| "Scheduler implementation in MP SoC design";Cho等;《Proceedings of the 2005 Asia and South Pacific Design Automation Conference》;20050131;全文 * |
| 有限状态机的一种实现框架;徐小良等;《工程设计学报》;20031031;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109086127A (en) | 2018-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3270289B1 (en) | Container-based multi-tenant computing infrastructure | |
| US9720727B1 (en) | Managing virtual machine migration | |
| Rajkumar et al. | Resource kernels: A resource-centric approach to real-time and multimedia systems | |
| US9250863B1 (en) | Managing virtual machine migration | |
| KR100628492B1 (en) | Method and system for performing real-time operation | |
| US9098214B1 (en) | Managing virtual machine migration | |
| US8332862B2 (en) | Scheduling ready tasks by generating network flow graph using information receive from root task having affinities between ready task and computers for execution | |
| US7774781B2 (en) | Storage subsystem access prioritization by system process and foreground application identification | |
| US20110154350A1 (en) | Automated cloud workload management in a map-reduce environment | |
| US9973512B2 (en) | Determining variable wait time in an asynchronous call-back system based on calculated average sub-queue wait time | |
| JP2005534116A (en) | A method for dynamically allocating and managing resources in a multi-consumer computer system. | |
| CN111796908A (en) | System and method for automatic elastic expansion and contraction of resources and cloud platform | |
| CN110069341A (en) | What binding function configured on demand has the dispatching method of dependence task in edge calculations | |
| WO2023045467A1 (en) | Container cpu resource scheduling and isolation method and apparatus, and storage medium and electronic device | |
| CN114327843A (en) | Task scheduling method and device | |
| US10037225B2 (en) | Method and system for scheduling computing | |
| CN111488210A (en) | Task scheduling method and device based on cloud computing and computer equipment | |
| US20230275976A1 (en) | Data processing method and apparatus, and computer-readable storage medium | |
| JP2023526883A (en) | Scheduling methods, computing devices, and storage media for tasks | |
| US10749813B1 (en) | Spatial-temporal cloud resource scheduling | |
| CN114598665A (en) | Resource scheduling method and device, computer readable storage medium and electronic equipment | |
| Bartolini et al. | Proactive workload dispatching on the EURORA supercomputer | |
| KR101890046B1 (en) | Concurrent network application scheduling for reduced power consumption | |
| CN109086127B (en) | Resource scheduling method based on FSM control and framework system | |
| CN111190731A (en) | Cluster task scheduling system based on weight |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |