CN108924214A - A kind of load-balancing method of computing cluster, apparatus and system - Google Patents
A kind of load-balancing method of computing cluster, apparatus and system Download PDFInfo
- Publication number
- CN108924214A CN108924214A CN201810685391.3A CN201810685391A CN108924214A CN 108924214 A CN108924214 A CN 108924214A CN 201810685391 A CN201810685391 A CN 201810685391A CN 108924214 A CN108924214 A CN 108924214A
- Authority
- CN
- China
- Prior art keywords
- task
- load
- computing unit
- status information
- node server
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1029—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers using data related to the state of servers by a load balancer
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
- H04L67/63—Routing a service request depending on the request content or context
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Computer And Data Communications (AREA)
Abstract
The present invention provides a kind of load-balancing methods of computing cluster, apparatus and system, are related to computing cluster load-balancing technique field.The load-balancing method of the computing cluster reads the task status information of task in task pool first, based on the task status information, judge whether the computing unit in the node server is capable of handling the task, if, the task is obtained from the task pool, then indicates that the computing unit carries out calculation processing to the task.The load-balancing method of the computing cluster cooperates the load balancing apparatus and the SiteServer LBS, in the case where not introducing extras automatically according to the real time resources situation reasonable distribution information process request of server, the waste of system resources in computation caused by due to key node bottleneck, task unreasonable distribution etc. is avoided.
Description
Technical field
The present invention relates to computing cluster load-balancing technique field, the load in particular to a kind of computing cluster is equal
The method, apparatus that weighs and system.
Background technique
With the development that internet economy makes rapid progress, economic activity process requires also water to the information processing capability of enterprise
Rise height, and traditional single server process mode has been unable to satisfy the demand of business development, and more and more enterprises begin to use
The cloud computing service of clustering is to cope with business scale expansion.Wherein, the computer cluster (abbreviation cluster) of clustering formation is
A kind of computer system, by the way that the computer hardware that one group loosely integrates is connected height, closely cooperation is completed to calculate for it
Work.In some sense, they can be considered as a computer.Single calculating equipment in group system commonly known as saves
Point usually by network connection, but also allows other connection types.Cluster computer is commonly used in improvement single computer
Calculating speed and reliability.Either public cloud or private clound, commercial city is supplied in cloud service to be divided jointly by multiple servers
The mode of load computation requests externally provides large batch of computing capability, therefore, how to be reasonably allocated to a large amount of computation requests
In cluster on each server, equilibrium of the computational load in server group is carried out, becomes what clustering calculating must face
Problem.
Load balancing (Load Balance) is common technology in a kind of computer cluster, for CPU, the net in cluster
Load is distributed in network, storage or other resources, is optimized resource use, maximum throughput rate to reach, is minimized the response time etc.
Purpose.Load balancing service is usually to carry specific software by dedicated hardware or private server to realize.But it is traditional
Application load balancing method need to be added in systems additional equipment for distribute request, increase soft and hardware cost;Simultaneously
Since the distribution of information process request depends on load-balancing device, the equipment easily becomes cluster when request amount becomes larger
Bottleneck node;In addition, different calculating scenes requires the mode of request distribution different, the server of load capacity calculation work
The computing capability that can be provided is also different, and what inappropriate request ways of distribution may cause that server in cluster receives asks
The amount of asking is unbalance, causes the waste of computing resource.
Summary of the invention
In view of this, a kind of load-balancing method for being designed to provide computing cluster of the embodiment of the present invention, device and
System, to solve the problem of in above-mentioned cluster that server receives that request amount is unbalance to cause computing resource waste.
In a first aspect, the embodiment of the invention provides a kind of load-balancing method of computing cluster, the computing cluster
Load-balancing method reads the task status information of task in task pool first, the task status information is based on, described in judgement
Whether the computing unit in node server is capable of handling the task, if so, the task is obtained from the task pool, then
Indicate that the computing unit carries out calculation processing to the task.
It is comprehensive in a first aspect, the task status information includes task scale, it is described to be based on the task status information, sentence
Whether the computing unit to break in the node server is capable of handling the task, including:Judge to calculate shape in the server
State is whether the operational capability of idle computing unit is greater than the operand of the task scale;If so, the node server
In computing unit be capable of handling the task.
It is comprehensive to be based on the task status information described in a first aspect, the task status information includes task type,
Judge whether the computing unit in the node server is capable of handling after the task, is obtained from the task pool described
Before taking the task, the load-balancing method of the computing cluster further includes:Obtain the task type of the task and described
The processed historic task type of node server;Judge whether the task type is the historic task class according to similarity
The priority match type of type;If so, executing step:The task is obtained from the task pool.
It is comprehensive in a first aspect, the instruction computing unit carries out calculation processing to the task, including:Described in acquisition
The caching of the processed historic task of node server;Indicate that the computing unit counts the task using the caching
Calculation processing.
It is comprehensive in a first aspect, described after obtaining the task in the task pool, the load of the computing cluster
Equalization methods further include:The calculating state of the computing unit is switched into work by the free time;It is single in the instruction calculating
After member carries out calculation processing to the task, the load-balancing method of the computing cluster further includes:By the computing unit
State switch to the free time from work.
It is comprehensive in a first aspect, after the instruction computing unit carries out calculation processing to the task, the meter
Calculate cluster load-balancing method further include:Judge whether to receive and exit computing cluster instruction, the computing cluster is to use
Multiple computers, multiple storage equipment and remembers the interconnection line of redundancy and form a single, High Availabitity for users
System;If so, judging whether all computing units of the node server are completely in idle state;If so, the section
Point server is disconnected with the computing cluster.
Second aspect, the embodiment of the invention provides a kind of load balancing apparatus of computing cluster, the load balancing dress
It sets including task status information read module, task judgment module, task acquisition module and task processing module.The task shape
State information reading module is used to read the task status information of task in task pool.The task judgment module is used for based on described
Task status information, judges whether the computing unit in the node server is capable of handling the task.The task obtains
When computing unit of the module in the node server is capable of handling the task, from the task pool described in acquisition
Task.The task processing module is used to indicate the computing unit and carries out calculation processing to the task.
Comprehensive second aspect, the task judgment module includes task type judging unit and similarity judging unit.Institute
Task type judging unit is stated for obtaining the task type and the processed historic task of the node server of the task
Type.The similarity judging unit is used to judge whether the task type is the historic task type according to similarity
Priority match type.
Comprehensive second aspect, the task processing module include caching acquiring unit and caching Multiplexing Unit.The caching
Acquiring unit is used to obtain the caching of the processed historic task of the node server.The caching Multiplexing Unit is used to indicate
The computing unit carries out calculation processing to the task using the caching.
The third aspect, the embodiment of the invention also provides a kind of storage medium, the storage medium is stored in computer,
The storage medium includes a plurality of instruction, and a plurality of instruction is configured such that the computer executes above-mentioned method.
Fourth aspect, the embodiment of the invention also provides a kind of SiteServer LBS of computing cluster, the load balancing
System includes task pool management server, task publisher server and node server.The task pool management server is used for
The task status information of task and the task to be processed in centrally stored and management computing cluster.The task publication clothes
Device be engaged in user, there are into the task pool of the task pool management server, publication includes the task when calculating demand
The task of status information.The node server is used to read the task status information in the task pool, and is based on institute
It states task status information instruction computing unit and calculation processing is carried out to the task.
Beneficial effect provided by the invention is:
The present invention provides a kind of load-balancing method of computing cluster, apparatus and system, the load-balancing method control
Computer node server processed reads the task status information of task in task pool, and letter is actively obtained in the task pool
The movement interference calculation process of active probe computing unit state is also not present without using complicated task allocation algorithms in breath,
It ensures that each of cluster node computing capability all reaches to give full play to.Meanwhile the node server is described in the acquisition
It can be based on the task status information before task, judge whether the computing unit in the node server is capable of handling described appoint
Business realizes that computing resource can dynamically distribute to ensure that the node server can complete the task, will not be calculated due to distribution
Method is unbalance to lead to part of nodes congestion and overstocked situation occurs in certain generic task in the case where other nodes free time.Further,
The load-balancing method can be completed using cluster existing equipment, and without increasing special equipment, non-stop layer node will not be due to
The performance limitation of central node drags down the entirety ability of cluster.
Other features and advantages of the present invention will be illustrated in subsequent specification, also, partly be become from specification
It is clear that by implementing understanding of the embodiment of the present invention.The objectives and other advantages of the invention can be by written theory
Specifically noted structure is achieved and obtained in bright book, claims and attached drawing.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is a kind of flow chart of the load-balancing method for computing cluster that first embodiment of the invention provides;
Fig. 2 is a kind of flow chart for node server operational capability judgment step that first embodiment of the invention provides;
Fig. 3 is a kind of module diagram for load balancing apparatus that second embodiment of the invention provides;
Fig. 4 provides a kind of module diagram of the SiteServer LBS of computing cluster for third embodiment of the invention.
Icon:100- load balancing apparatus;110- task status information read module;120- task judgment module;
130- task acquisition module;140- task processing module;200- SiteServer LBS;210- task pool management server;
220- task publisher server;230- node server.
Specific embodiment
Below in conjunction with attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete
Ground description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Usually exist
The component of the embodiment of the present invention described and illustrated in attached drawing can be arranged and be designed with a variety of different configurations herein.Cause
This, is not intended to limit claimed invention to the detailed description of the embodiment of the present invention provided in the accompanying drawings below
Range, but it is merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
It should be noted that:Similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.Meanwhile of the invention
In description, term " first ", " second " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
First embodiment
Through the applicant the study found that in the conventional load equalization methods that existing computing cluster uses, service request
Distribution forwards the load-balancing device that places one's entire reliance upon, with the increase of calculate node quantity in cluster, the meter of load-balancing device
Calculation amount, information throughput will be also continuously increased.Therefore, when cluster scale expands to a certain extent, load-balancing device will be by
The bottleneck node of system performance is reached capacity and become in computing capability or information handling capacity.Meanwhile equal dependent on load
In the computing cluster for the equipment that weighs, since the distribution of service must be added deduct by load-balancing device when cluster needs to increase
When few calculate node, the change information for registering calculate node in load-balancing device in advance is needed, wants the calculating money for adjusting cluster
Source is more troublesome.Traditional load-balancing device generallys use mean allocation or distributes both by calculate node handling capacity size
Mode distributes calculating task.The computing capability of the former calculate node in the cluster easilys lead to part and calculates section when irregular
Point can not be used effectively, and cause the decline of system overall calculation ability;The latter then requires load-balancing device to grasp and calculates section
On the one hand the handling capacity of point may monitor the calculating process of interfering nodes due to introducing, be on the other hand then further aggravated negative
The balanced computing capability of node and the burden of information handling capacity are carried, occurs system bottleneck in advance.To solve the above-mentioned problems,
First embodiment of the invention provides a kind of load-balancing method of computing cluster.
Referring to FIG. 1, Fig. 1 is a kind of process of the load-balancing method for computing cluster that first embodiment of the invention provides
Figure.Specific step is as follows for the load-balancing method of the computing cluster:
Step S10:Read the task status information of task in task pool.
Step S20:Based on the task status information, judge whether the computing unit in the node server can be located
Manage the task.
Step S30:When computing unit in the node server is capable of handling the task, from the task pool
Obtain the task.
Step S40:Indicate that the computing unit carries out calculation processing to the task.
For step S10, in order to guarantee the consistency of data, exist in computing cluster for carry out each node server it
Between data exchange shared library, the shared library be generally network attached storage (Network Attached Storage,
NAS) or forms, the network attached storage such as database are literally briefly exactly to be connected on network, have data storage
The device of function, thus also referred to as " network memory ", it is a kind of exclusive data storage server, it is data-centered, it will
Storage equipment is completely separated with server, manages data concentratedly, to discharge bandwidth, improve performance, reduce total cost of ownership, protect
Shield investment.The task pool utilizes task to be processed in such data sharing library storage and management computing cluster and described
The task status information of business.Wherein, as an implementation, the task status information may include task type, task
The information such as scale, task status and task definition.
It should be understood that executing the task status letter that step S10 reads task in task pool in the node server
Before breath, user also needs through task release module the release tasks in the task pool.Specifically, it is calculated when user exists
When demand, from tradition directly different, the load-balancing method provided in this embodiment to the cluster initiation mode of service request
Service request is proposed by way of issuing a mission bit stream into task pool, release tasks on NAS usually by creating
File is realized to forms such as database insertion data, one to two sentence is needed at program realization upper, without introducing additionally special
Industry processing module, therefore can be considered virtual module.
For step S20, i.e.,:Based on the task status information, judge that the computing unit in the node server is
It is no to be capable of handling the task.As an implementation, the step S20 of the load-balancing method in the present embodiment further includes son
Step, referring to FIG. 2, Fig. 2 is a kind of stream for node server operational capability judgment step that first embodiment of the invention provides
Cheng Tu.Specific step is as follows for the sub-step of the node server operational capability judgment step, that is, step S20:
Step S21:Judge whether state is calculated in the server is greater than institute for the operational capability of idle computing unit
State the operand of task scale.
Step S22:The operational capability that state is idle computing unit is calculated in the server is greater than the task
When the operand of scale, the task type and the processed historic task type of the node server of the task are obtained.
Step S23:According to similarity judge the task type whether be the historic task type priority match class
Type, if so, executing step S30.
For step S21, the poor node server tendency of computing capability chooses small-scale calculating task, when avoiding calculating
Between it is too long;The stronger node server tendency of computing capability chooses computationally intensive task, reduces repeatedly interaction bring I/O damage
It loses.Therefore, node server needs to be implemented the task scale that step S21 identifies the task before acquisition task, with maximum limit
Degree ground plays the computing capability of each computing unit.Further, the calculate node can preferentially be chosen within the scope of operational capability
The largest task of task, and then choose to itself task is most suitable for, to play each calculate node to the maximum extent
Operational capability.
Further, in an alternate embodiment of the invention, the node server is except through identifying the meter of the computing unit
Calculation state is outside operational capability that is idle or working to judge the computing unit current residual, because of many computing units now
Can parallel processing calculating task, the node server can also be in the current residual directly by reading the computing unit
Deposit the operational capability to judge the computing unit current residual.
For step S22 and S23, node server tends to choose identical or close as the task type done before
Task, when system, which exists, to be cached, this measure can promote cache hit rate, and then improve task treatment effeciency.
For step S30:When computing unit in the node server is capable of handling the task, from the task
The task is obtained in pond.The node server can obtain the task in task pool, each node clothes by task acquisition module
Being engaged in device can be according to the suitable task of own actual situation acquisition, it is ensured that the computing resource of itself is maximally utilized.It should manage
Solution, identical as task release module, the task acquisition module also belongs to virtual module, without introducing additional entity
Professional treatment module.
As an implementation, the node server obtains described appoint in execution step S30 from the task pool
When business, it is understood that there may be the case where failing acquisition task, such situation illustrates without task to be processed in the task pool, described
Node server can enable the temporary suspend mode of affiliated computing unit, with energy saving, and realize more reasonable calculation resources scheduling.
For step S40, i.e.,:Indicate that the computing unit carries out calculation processing to the task.In the computing unit
When processing historic task similar with the task, in order to improve task treatment effeciency and accuracy rate, to the tool of the task
Body calculation processing mode can be:Obtain the caching of the processed historic task of the node server;Indicate that the calculating is single
The member application caching carries out calculation processing to the task.
It should be understood that the operational capability and state for the computing unit preferably to the node server carry out
Immediately monitoring, and then realize that the utilization rate of operational capability maximizes, the present embodiment obtains institute in step S30 from the task pool
It further include step after stating task:The calculating state of the computing unit is switched into work by the free time.Meanwhile in step S40
It further include step after indicating that the computing unit carries out calculation processing to the task:By the state of the computing unit from
Work switches to the free time.
Multiple node servers are frequently included in computing cluster, and the quantity of node server can change, and can deposit
The case where new node server or node server exit is being added.Therefore the computing cluster provided in this embodiment is negative
There is also addition/removing steps of node server in load equalization methods:(1) node adds:When new node server is added
When, without additional operation, task acquisition module actively can pull task from task pool after the starting of new node server, from
Calculating task is shared in computing cluster;(2) node removes:When node server will be exited from computing cluster, in the node serve
Mark is exited in device setting, and state can be switched to the free time after completing existing task by the computing unit of the node server, to this
When all computing units of node server are idle state, the node server and computing cluster are disconnected, and can directly be removed
The computing cluster.
Second embodiment
For the load-balancing method for the computing cluster for preferably cooperating first embodiment of the invention to provide, the present invention
Second embodiment additionally provides a kind of load balancing apparatus 100 of computing cluster.
Referring to FIG. 3, Fig. 3 is a kind of module diagram for load balancing apparatus that second embodiment of the invention provides.
Load balancing apparatus 100 includes task status information read module 110, task judgment module 120, task acquisition mould
Block 130 and task processing module 140.
Task status information read module 110, for reading the task status information of task in task pool.
Task judgment module 120 judges the calculating list in the node server for being based on the task status information
Whether member is capable of handling the task.
Task acquisition module 130, when being capable of handling the task for the computing unit in the node server, from
The task is obtained in the task pool.
Task processing module 140 is used to indicate the computing unit and carries out calculation processing to the task.
As an optional implementation manner, the task judgment module 120 in the present embodiment may include that task type is sentenced
Disconnected unit and similarity judging unit.
Task type judging unit, for obtaining the task type of the task and the node server is processed goes through
History task type.
Similarity judging unit judges whether the task type is the preferential of the historic task type according to similarity
Match-type.
Further, in this embodiment task processing module 140 may include that caching acquiring unit and caching multiplexing are single
Member.
Acquiring unit is cached, for obtaining the caching of the processed historic task of the node server.
Multiplexing Unit is cached, the computing unit is used to indicate using the caching and calculation processing is carried out to the task.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description
Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
3rd embodiment
Entity load-balancing device is all made of in the load-balancing method of traditional calculations cluster to receive at the task of client
Reason request, and true server is selected by dispatching algorithm, then task processing request is transmitted to selected true clothes
Business device has high demand, and the entity load-balancing device to the operational capability of the entity load-balancing device
The bottleneck node of system performance will be reached capacity and become due to computing capability or information handling capacity.In order to cooperate the present invention the
The load-balancing method for the computing cluster that one embodiment provides solves the above problems, and third embodiment of the invention additionally provides
A kind of SiteServer LBS 200 of computing cluster.
Referring to FIG. 4, a kind of module for the SiteServer LBS that Fig. 4 provides computing cluster for third embodiment of the invention is shown
It is intended to.
The SiteServer LBS 200 of the computing cluster includes task pool management server 210, task publisher server
220 and node server 230.
Task pool management server 210, for task to be processed in centrally stored and management computing cluster and described
The task status information of task.
Task publisher server 220, for user there are when calculating demand to the task pool management server 210
Publication includes the task of the task status information in task pool.
Node server 230 is believed for reading the task status information in the task pool, and based on the task status
Breath instruction computing unit carries out calculation processing to the task.
Optionally, task pool management server 210 can be entity storage server, or be distributed in each computer
Or the virtual server in other processing equipments.Further, the task pool management server 210 can also be with the task
Publisher server 220 is arranged on same entity calculation processing apparatus.
It is appreciated that structure shown in Fig. 4 is only to illustrate, the SiteServer LBS 200 may also include than shown in Fig. 4
More perhaps less component or with the configuration different from shown in Fig. 4.Each component shown in Fig. 4 can using hardware,
Software or combinations thereof is realized.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description
Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
In conclusion the embodiment of the invention provides a kind of load-balancing method of computing cluster, apparatus and system, it is described
Load-balancing method controls the task status information that computer node server reads task in task pool, and actively at described
Business obtains information in pond, and without using complicated task allocation algorithms, the movement of active probe computing unit state is also not present
Interference calculation process, it is ensured that each of cluster node computing capability, which all reaches, to be given full play to.Meanwhile the node serve
Device can be based on the task status information before obtaining the task, judge that the computing unit in the node server whether can
The task is enough handled, to ensure that the node server can complete the task, realizes that computing resource can dynamically distribute, no
There are overstocked feelings in certain generic task in the case where causing part of nodes congestion and other nodes idle since allocation algorithm is unbalance
Condition.Further, the load-balancing method can be completed using cluster existing equipment, without increasing special equipment, non-stop layer
Node will not limit the entirety ability for dragging down cluster due to the performance of central node.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through
Other modes are realized.The apparatus embodiments described above are merely exemplary, for example, flow chart and block diagram in attached drawing
Show the device of multiple embodiments according to the present invention, the architectural framework in the cards of method and computer program product,
Function and operation.In this regard, each box in flowchart or block diagram can represent the one of a module, section or code
Part, a part of the module, section or code, which includes that one or more is for implementing the specified logical function, to be held
Row instruction.It should also be noted that function marked in the box can also be to be different from some implementations as replacement
The sequence marked in attached drawing occurs.For example, two continuous boxes can actually be basically executed in parallel, they are sometimes
It can execute in the opposite order, this depends on the function involved.It is also noted that every in block diagram and or flow chart
The combination of box in a box and block diagram and or flow chart can use the dedicated base for executing defined function or movement
It realizes, or can realize using a combination of dedicated hardware and computer instructions in the system of hardware.
In addition, each functional module in each embodiment of the present invention can integrate one independent portion of formation together
Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes:USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access
The various media that can store program code such as memory (RAM, Random Access Memory), magnetic or disk.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.It should be noted that:Similar label and letter exist
Similar terms are indicated in following attached drawing, therefore, once being defined in a certain Xiang Yi attached drawing, are then not required in subsequent attached drawing
It is further defined and explained.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
Claims (10)
1. a kind of load-balancing method of computing cluster, which is characterized in that be applied to node server, the load-balancing method
Including:
Read the task status information of task in task pool;
Based on the task status information, judge whether the computing unit in the node server is capable of handling the task;
If so, obtaining the task from the task pool;
Indicate that the computing unit carries out calculation processing to the task.
2. load-balancing method according to claim 1, which is characterized in that the task status information includes task rule
Mould, it is described to be based on the task status information, judge whether the computing unit in the node server is capable of handling described appoint
Business, including:
Judge whether state is calculated in the server is greater than the task scale for the operational capability of idle computing unit
Operand;
If so, the computing unit in the node server is capable of handling the task.
3. load-balancing method according to claim 1, which is characterized in that the task status information includes task class
Type is based on the task status information described, it is described to judge whether the computing unit in the node server is capable of handling
After task, described before obtaining the task in the task pool, the load-balancing method further includes:
Obtain the task type and the processed historic task type of the node server of the task;
According to similarity judge the task type whether be the historic task type priority match type;
If so, executing step:The task is obtained from the task pool.
4. load-balancing method according to claim 3, which is characterized in that the instruction computing unit is to described
Business carries out calculation processing, including:
Obtain the caching of the processed historic task of the node server;
Indicate that the computing unit carries out calculation processing to the task using the caching.
5. load-balancing method according to claim 3, which is characterized in that it is described from the task pool obtain described in
After task, the load-balancing method further includes:
The calculating state of the computing unit is switched into work by the free time;
After the instruction computing unit carries out calculation processing to the task, the load-balancing method further includes:
The state of the computing unit is switched into the free time from work.
6. load-balancing method according to claim 1, which is characterized in that in the instruction computing unit to described
After task carries out calculation processing, the load-balancing method further includes:
Judge whether to receive and exit computing cluster instruction, the computing cluster is to use multiple computers, multiple storage equipment
Single, High Availabitity a system for users is formed with the interconnection line of memory redundancy;
If so, judging whether all computing units of the node server are completely in idle state;
If so, the node server is disconnected with the computing cluster.
7. a kind of load balancing apparatus of computing cluster, which is characterized in that the load balancing apparatus includes:
Task status information read module, for reading the task status information of task in task pool;
Task judgment module judges that the computing unit in node server whether can for being based on the task status information
Handle the task;
Task acquisition module, when being capable of handling the task for the computing unit in the node server, from described
Business obtains the task in pond;
Task processing module is used to indicate the computing unit and carries out calculation processing to the task.
8. load balancing apparatus according to claim 7, which is characterized in that the task judgment module includes:
Task type judging unit, task type and the processed history of the node server for obtaining the task are appointed
Service type;
Similarity judging unit, for judging whether the task type is the preferential of the historic task type according to similarity
Match-type.
9. load balancing apparatus according to claim 7 or 8, which is characterized in that the task processing module includes:
Acquiring unit is cached, for obtaining the caching of the processed historic task of the node server;
Multiplexing Unit is cached, the computing unit is used to indicate using the caching and calculation processing is carried out to the task.
10. a kind of SiteServer LBS of computing cluster, which is characterized in that the SiteServer LBS includes:
Task pool management server, for appointing for task and the task to be processed in centrally stored and management computing cluster
Business status information;
Task publisher server, for there are send out when calculating demand into the task pool of the task pool management server in user
Cloth includes the task of the task status information;
Node server is indicated for reading the task status information in the task pool, and based on the task status information
Computing unit carries out calculation processing to the task.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810685391.3A CN108924214A (en) | 2018-06-27 | 2018-06-27 | A kind of load-balancing method of computing cluster, apparatus and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810685391.3A CN108924214A (en) | 2018-06-27 | 2018-06-27 | A kind of load-balancing method of computing cluster, apparatus and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108924214A true CN108924214A (en) | 2018-11-30 |
Family
ID=64421893
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810685391.3A Pending CN108924214A (en) | 2018-06-27 | 2018-06-27 | A kind of load-balancing method of computing cluster, apparatus and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108924214A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110096349A (en) * | 2019-04-10 | 2019-08-06 | 山东科技大学 | A kind of job scheduling method based on the prediction of clustered node load condition |
CN111090517A (en) * | 2019-11-26 | 2020-05-01 | 中国建设银行股份有限公司 | Job scheduling method, device, equipment and storage medium |
CN111738446A (en) * | 2020-06-12 | 2020-10-02 | 北京百度网讯科技有限公司 | Scheduling method, device, equipment and medium of deep learning inference engine |
CN111866043A (en) * | 2019-04-29 | 2020-10-30 | 中国移动通信集团河北有限公司 | Task processing method and device, computing equipment and computer storage medium |
CN112035259A (en) * | 2020-09-01 | 2020-12-04 | 中国银行股份有限公司 | Data processing system, method and server group |
CN112395058A (en) * | 2020-11-13 | 2021-02-23 | 北京数码视讯技术有限公司 | Task regulation and control device, method and system |
CN113553171A (en) * | 2021-06-07 | 2021-10-26 | 用友汽车信息科技(上海)股份有限公司 | Load balancing control method, device and computer readable storage medium |
CN115695432A (en) * | 2023-01-04 | 2023-02-03 | 河北华通科技股份有限公司 | Load balancing method and device, electronic equipment and storage medium |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1791113A (en) * | 2005-12-26 | 2006-06-21 | 北京航空航天大学 | Method for balancing gridding load |
US20100122019A1 (en) * | 2008-11-10 | 2010-05-13 | David Flynn | Apparatus, system, and method for managing physical regions in a solid-state storage device |
CN102012840A (en) * | 2010-12-23 | 2011-04-13 | 中国农业银行股份有限公司 | Batch data scheduling method and system |
CN102637186A (en) * | 2011-12-09 | 2012-08-15 | 中兴通讯股份有限公司 | CMS (content management system) content order system and implementation method thereof |
CN102957733A (en) * | 2011-08-31 | 2013-03-06 | 中国科学院声学研究所 | Content publish method and system with dynamic adjusting and optimizing function |
CN103186418A (en) * | 2011-12-30 | 2013-07-03 | 北大方正集团有限公司 | Method and system for distributing tasks |
CN103294728A (en) * | 2012-03-05 | 2013-09-11 | 阿里巴巴集团控股有限公司 | Data processing method and system |
CN103327072A (en) * | 2013-05-22 | 2013-09-25 | 中国科学院微电子研究所 | Method for cluster load balance and system thereof |
CN105597324A (en) * | 2014-11-14 | 2016-05-25 | 一隻怪兽股份有限公司 | System with task growth and screening mechanism and implementation method thereof |
CN106095585A (en) * | 2016-06-22 | 2016-11-09 | 中国建设银行股份有限公司 | Task requests processing method, device and enterprise information system |
WO2017117252A1 (en) * | 2015-12-28 | 2017-07-06 | Amazon Technologies, Inc. | Multi-path transport design |
CN107623732A (en) * | 2017-09-15 | 2018-01-23 | 郑州云海信息技术有限公司 | A kind of date storage method based on cloud platform, device, equipment and storage medium |
-
2018
- 2018-06-27 CN CN201810685391.3A patent/CN108924214A/en active Pending
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1791113A (en) * | 2005-12-26 | 2006-06-21 | 北京航空航天大学 | Method for balancing gridding load |
US20100122019A1 (en) * | 2008-11-10 | 2010-05-13 | David Flynn | Apparatus, system, and method for managing physical regions in a solid-state storage device |
CN102012840A (en) * | 2010-12-23 | 2011-04-13 | 中国农业银行股份有限公司 | Batch data scheduling method and system |
CN102957733A (en) * | 2011-08-31 | 2013-03-06 | 中国科学院声学研究所 | Content publish method and system with dynamic adjusting and optimizing function |
CN102637186A (en) * | 2011-12-09 | 2012-08-15 | 中兴通讯股份有限公司 | CMS (content management system) content order system and implementation method thereof |
CN103186418A (en) * | 2011-12-30 | 2013-07-03 | 北大方正集团有限公司 | Method and system for distributing tasks |
CN103294728A (en) * | 2012-03-05 | 2013-09-11 | 阿里巴巴集团控股有限公司 | Data processing method and system |
CN103327072A (en) * | 2013-05-22 | 2013-09-25 | 中国科学院微电子研究所 | Method for cluster load balance and system thereof |
CN105597324A (en) * | 2014-11-14 | 2016-05-25 | 一隻怪兽股份有限公司 | System with task growth and screening mechanism and implementation method thereof |
WO2017117252A1 (en) * | 2015-12-28 | 2017-07-06 | Amazon Technologies, Inc. | Multi-path transport design |
CN106095585A (en) * | 2016-06-22 | 2016-11-09 | 中国建设银行股份有限公司 | Task requests processing method, device and enterprise information system |
CN107623732A (en) * | 2017-09-15 | 2018-01-23 | 郑州云海信息技术有限公司 | A kind of date storage method based on cloud platform, device, equipment and storage medium |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110096349A (en) * | 2019-04-10 | 2019-08-06 | 山东科技大学 | A kind of job scheduling method based on the prediction of clustered node load condition |
CN111866043A (en) * | 2019-04-29 | 2020-10-30 | 中国移动通信集团河北有限公司 | Task processing method and device, computing equipment and computer storage medium |
CN111866043B (en) * | 2019-04-29 | 2023-04-28 | 中国移动通信集团河北有限公司 | Task processing method, device, computing equipment and computer storage medium |
CN111090517A (en) * | 2019-11-26 | 2020-05-01 | 中国建设银行股份有限公司 | Job scheduling method, device, equipment and storage medium |
CN111738446A (en) * | 2020-06-12 | 2020-10-02 | 北京百度网讯科技有限公司 | Scheduling method, device, equipment and medium of deep learning inference engine |
CN111738446B (en) * | 2020-06-12 | 2023-11-03 | 北京百度网讯科技有限公司 | Scheduling method, device, equipment and medium of deep learning reasoning engine |
CN112035259A (en) * | 2020-09-01 | 2020-12-04 | 中国银行股份有限公司 | Data processing system, method and server group |
CN112395058A (en) * | 2020-11-13 | 2021-02-23 | 北京数码视讯技术有限公司 | Task regulation and control device, method and system |
CN113553171A (en) * | 2021-06-07 | 2021-10-26 | 用友汽车信息科技(上海)股份有限公司 | Load balancing control method, device and computer readable storage medium |
CN115695432A (en) * | 2023-01-04 | 2023-02-03 | 河北华通科技股份有限公司 | Load balancing method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108924214A (en) | A kind of load-balancing method of computing cluster, apparatus and system | |
JP5336094B2 (en) | Multipurpose allocation of computing jobs in client / server or hosting environments | |
WO2017167025A1 (en) | Method and device for realizing task scheduling, and computer storage medium | |
CN110134495B (en) | Container cross-host online migration method, storage medium and terminal equipment | |
TWI244030B (en) | Information processing system, information processing device, distributed information processing method and computer readable recording medium | |
JP5584765B2 (en) | Method and apparatus for data center automation | |
JP4959845B2 (en) | Method, computer program, and data processing system for calculating dispatcher share in workload manager | |
CN104375897B (en) | Cloud computing resource scheduling method based on the unbalanced degree of minimum relative load | |
CN105141541A (en) | Task-based dynamic load balancing scheduling method and device | |
CN107548549A (en) | Equilibrium of stock in DCE | |
CN110537169A (en) | Cluster resource management in distributed computing system | |
TW201214284A (en) | Virtualization and dynamic resource allocation aware storage level reordering | |
TW201140451A (en) | Goal oriented performance management of workload utilizing accelerators | |
CN109684074A (en) | Physical machine resource allocation methods and terminal device | |
CN110362388A (en) | A kind of resource regulating method and device | |
CN107977271A (en) | A kind of data center's total management system load-balancing method | |
Perwej | The ambient scrutinize of scheduling algorithms in big data territory | |
Praveen et al. | An effective multi-faceted cost model for auto-scaling of servers in cloud | |
Liu et al. | Deadline guaranteed service for multi-tenant cloud storage | |
Deiab et al. | Energy efficiency in cloud computing | |
Johora et al. | A load balancing strategy for reducing data loss risk on cloud using remodified throttled algorithm | |
CN107528871A (en) | Data analysis in storage system | |
Anan et al. | SLA-based optimization of energy efficiency for green cloud computing | |
Heidari et al. | A cost-efficient auto-scaling algorithm for large-scale graph processing in cloud environments with heterogeneous resources | |
CN107402812A (en) | Cluster resource dispatching method, device, equipment and storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181130 |