CN105528283B - A kind of method that load value is calculated in mobile application detection load-balancing algorithm - Google Patents
A kind of method that load value is calculated in mobile application detection load-balancing algorithm Download PDFInfo
- Publication number
- CN105528283B CN105528283B CN201510908946.2A CN201510908946A CN105528283B CN 105528283 B CN105528283 B CN 105528283B CN 201510908946 A CN201510908946 A CN 201510908946A CN 105528283 B CN105528283 B CN 105528283B
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- complexity
- detection
- detection node
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3409—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment for performance assessment
- G06F11/3433—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment for performance assessment for load management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/302—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a software system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3452—Performance evaluation by statistical analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The method that load value is calculated in load-balancing algorithm is detected the invention discloses a kind of mobile application, is realized based on the experiment porch being made of client, central administration node module and detection node；The heartbeat of detection node includes：The CPU weighted load values of current detection node, memory weighted load value, queue weight load value and control stream complexity weighted load value；Central administration node module receives mobile application Detection task and detects the heartbeat of child node；The loading condition of current each detection node is obtained by the heartbeat of each detection node of real-time reception, and pass through load dispatcher and select wherein optimal detection node and issue mobile application Detection task, then the load fraction of each detection node, the load information of more new record are recalculated.The advantage is that greatly improving the execution efficiency of magnanimity mobile application Detection task, can meet the needs of magnanimity application now quickly detects；More rationally effectively task is distributed to detection node.
Description
Technical field
The invention belongs to information security fields, are related to the optimization method of mobile application detection load balancing, are specifically one kind
The method that load value is calculated in mobile application detection load-balancing algorithm.
Background technology
Due to mobile application Detection task have particularity, it is difficult to by its cutting be more fine-grained subfile, more can not
Simple and crude mean size cutting can be carried out to single mobile application Detection task, is caused in single mobile application Detection task
Mobile application size difference it is totally different, influence the execution efficiency of mobile application Detection task.Appoint in addition, influencing mobile application detection
Another factor of the execution efficiency of business is the control stream complexity in single mobile application Detection task.
Control stream complexity refers to by the execution flow inside mobile application Detection task and performs complexity, is flowed by control
Side number and number of nodes in figure embody；
Controlling stream graph (CFG, Control flow graph, i.e. control flow chart) is the abstract of a process or program
Performance.The execution stream of sentence, statement block and process is abstracted in the controlling stream graph of mobile application Detection task, i.e. mobile application.Control
Flow graph processed is a digraph, includes N number of node node and M side edge.
In the prior art using McCabe complexity metric standards, the flow chart of software is converted into digraph, control stream
Figure is the basis of McCabe complicated dynamic behaviours, loop complexity degree of the McCabe complexity metrics as mobile application Detection task
Amount standard is generally described with cyclomatic complexity V (G).It is demonstrated experimentally that the mobile application that cyclomatic complexity is bigger, the static detection time gets over
It is long.
The computational methods of cyclomatic complexity are as follows：
V (G)=e-n+2
Wherein e is that side number, n are controlling stream graph in mobile application Detection task in controlling stream graph in mobile application Detection task
In number of nodes.
The purpose of load balancing (Load Balancing) algorithm is to improve tasks carrying efficiency, improves system throughput
Amount improves system execution performance using distributed frame.By balancing the loading condition of each detection node, each detection section is strengthened
The data-handling capacity of point improves availability.The load of each detection node includes：All mobile application inspections in each detection node
Network throughput, cpu load rate, memory usage of survey task etc.；
In conventional load equalization algorithm, generally gulped down using CPU usage, memory usage, hard disk service condition, network
The measurement standard as server load such as the amount of spitting.
But under the application scenarios of magnanimity mobile application Detection task, since mobile application file is small, in detection process
In, the service condition of server hard disc will not vary widely, and hard disk usage amount influences the detection of mobile application smaller；And
Since the storage and detection of mobile application are all completed under LAN environment, network is all right in LAN, and small documents are to net
Network bandwidth requirement is not high, therefore network throughput also has no significant effect the detection efficiency of mobile application.
Under the application scenarios of magnanimity mobile application Detection task, mobile application Detection task is stored using ActiveMQ
On queue server, the API provided by ActiveMQ obtains the queue size of mobile application Detection task, and configures queue
Maximum length, and the size of Detection task queue is to weigh time and the expected time of return that a Detection task needs wait
Most direct factor.
To sum up, for magnanimity mobile application Detection task, in conventional load equalization algorithm, it only considered influence load balancing
Following a few class factors：1), the cpu load of detection node uses percentage including CPU core number, cpu frequency, CPU；2), detect
The memory load of node, including free memory percentage；3), the task queue load of detection node, including task queue length
With task queue maximum length.
If flow complexity without considering control, it will entire detection efficiency is made to decline more than 50%.Have multinode and
Magnanimity application is needed in the case of detecting, and detection time will be delayed over 1 minute, can not meet magnanimity mobile application at this stage
The demand of detection.
The content of the invention
The present invention is asked for the processing of magnanimity mobile application Detection task, in order to improve multi-threading parallel process efficiency,
With reference to existing load balancing computational methods, load information of the control stream complexity as detection node is introduced, it is proposed that a kind of
The method that load value is calculated in mobile application detection load-balancing algorithm, can detect a large amount of mobile applications in short-term.
It is as follows：
Step 1: for each detection node, the control of each mobile application Detection task in the detection node is calculated respectively
System stream complexity Complexity；
Using McCabe complexity metric standards, the controlling stream graph of each mobile application Detection task is obtained, so as to obtain
The side number and number of nodes of the controlling stream graph calculate control stream complexity Complexity；
Complexityj=ej-nj+2
ComplexityjFor the control stream complexity of j-th of mobile application Detection task；J is mobile application to be detected
The sequence number of Detection task, j are integer；ejFor the side number of the controlling stream graph of j-th of mobile application Detection task, njFor j-th of movement
Using the number of nodes of the controlling stream graph of Detection task.
Step 2: the CPU weighted loads of all mobile application Detection tasks in each detection node are calculated, memory weighting
Load, the sum of queue weight load and control stream complexity weighted load value LoadGrade：
LoadGradei=Cw+Mw+Qw+Comw
Wherein, LoadGradeiThe sum of weighted load value for detection node i；I is the sequence number of detection node, and i is integer；
Cw is the CPU weighted load values of detection node i：Cw=wcpu*CPUGradei；wcpuIt is the weight of CPU weighted load values；
CPUGradeiFor the cpu load value of detection node i：
CPUCoresiFor the CPU core number of detection node i, unit is a；CPUGHziIt is single for the CPU frequency of detection node i
Position is GHz；CPUUsedPerciPercentage, unit % are used for the CPU of detection node i.
Mw is the memory weighted load value of detection node i：Mw=wmemory*MemoryGradei；wmemoryIt is memory weighting
The weight of load value, MemoryGradeiFor the memory load value of detection node i：
FreeMemoryiFor the current idle memory of detection node i, unit %.
Qw is the queue weight load value of detection node i：Qw=wqueue*QueueGradei；wqueueIt is queue weight load
The weight of value, QueueGradeiFor the task queue load value of detection node i：
QueueCapacityiFor the task queue maximum length of task queue capacity, that is, detection node i of detection node i,
Unit is a；QueueNumiFor task queue current task quantity, that is, queue size of detection node i, unit is a.
Comw is the control stream complexity weighted load value of detection node i：
Comw=wcomplexity*ComplexityGradei
wcomplexityFor the weight of the control stream complexity of mobile application detection node, ComplexityGradeiFor movement
Using the control stream complexity load value of detection node i:
ComplexitymaxiFor detection node i in time slot T the medium task to be detected of task queue Complexity it
The maximum of sum；ComplexityaverageiFor the task queue medium task to be detected of detection node i in time slot T
Complexity average values：
∑TComplexityjThe wait of expression detection node i in time slot T all Detection tasks in task queue
The sum of Complexity, n represent n mobile application in time slot T.
All kinds of weighted values of detection node i meet：wcpu+wmemory+wqueue+wComplexity=100
wcpu,wmemory,wqueue,wComplexity∈[0,100]
Step 3: the information encapsulation of the sum of weighted load value by each detection node LoadGrade is in nodes heart beat；
Step 4: nodes heart beat is sent every time slot T once to load dispatcher, load dispatcher calculates each detection section
Next mobile application Detection task is distributed to the detection node of LoadGrade minimums by point.
The advantage of the invention is that：
1) a kind of method that load value, is calculated in mobile application detection load-balancing algorithm, greatly improves magnanimity movement
Using the execution efficiency of Detection task, can meet the needs of magnanimity application now quickly detects.
2) a kind of method that load value, is calculated in mobile application detection load-balancing algorithm, use are more rationally effective square
Formula distributes task to node, and ensure that node prevents its overload again while fully running, is conducive to safeguard the node longevity
Life.
Description of the drawings
Fig. 1 is mobile application control stream complexity of the present invention and the relational graph of static detection time；
Fig. 2 is the method flow diagram that load value is calculated in a kind of mobile application detection load-balancing algorithm of the present invention；
Fig. 3 is the present invention and polling dispatching method, and the time of traditional mobile application Detection task load-balancing algorithm counts
The figure compared with.
Specific embodiment
Below in conjunction with attached drawing, the present invention is described in further detail.
As shown in Figure 1, as mobile application controls the increase of stream complexity, the static detection time of mobile application is also gradual
Increase, mobile application detection efficiency continuously decreases, and therefore, can be calculated the control stream complexity of mobile application as load balancing
One parameter of method, allow each detection node task queue on the sum of control stream complexity of task is waited to tend to balance,
Improve detection performance when batch performs mobile application detection.
The method that load value is calculated in a kind of mobile application detection load-balancing algorithm of the present invention, is examined in magnanimity mobile application
Dynamic feedback of load equalization algorithm is introduced in the scene of survey task, the real time load and response condition of server is considered, constantly adjusts
The ratio of number of tasks is handled between whole node, task node to be avoided still to receive a large amount of requests, equiblibrium mass distribution formula system in overload
The loading condition of task node in system, so as to improve the throughput of whole system and tasks carrying efficiency.
This method is realized based on the experiment porch being made of client, central administration node module and detection child node；
The heartbeat of detection node includes：The CPU weighted load values of current detection node i, memory weighted load value, queue add
Weigh load value and control stream complexity weighted load value；
Central administration node module receives mobile application Detection task and detects the heartbeat of child node；It is each by real-time reception
The heartbeat of detection node obtains the loading condition of current each detection node, and passes through load dispatcher and select wherein optimal inspection
It surveys node and issues mobile application Detection task, then recalculate the load fraction of each detection node, the load letter of more new record
Breath.
Central administration node module is divided into heartbeat and receives engine, load value computing engines and detection child node resource queue three
Part.Heartbeat receives that after engine receives the heartbeat message sent of detection child node, load information therein taking-up is passed to
Load value computing engines；After load value computing engines calculate the real time load value of egress according to node load information, according to negative
The size of load value resets detection child node resource queue.It, can be direct after load dispatcher receives mobile application Detection task
It is maximum to be assigned to load fraction, loads the detection child node of lowest section；
Load dispatcher effect is monitoring and the load information for collecting each server, and one is calculated according to multiple load informations
A integrated load value.When integrated load value represents that server is busy, the priority of the detection node is smaller, so new distribution
Number of request to the mobile application Detection task of the server will be lacked.When integrated load value represents that server is in low profit
During with rate, the priority of detection node is larger, increases the number of request for being newly assigned to the server with this.
As shown in Fig. 2, it is as follows：
Step 1: for each detection node, the control of each mobile application Detection task in the detection node is calculated respectively
System stream complexity Complexity；
Using McCabe complexity metric standards, the controlling stream graph of each mobile application Detection task is obtained, so as to obtain
The side number and number of nodes of the controlling stream graph calculate control stream complexity Complexity；
Complexityj=ej-nj+2
ComplexityjFor the control stream complexity of j-th of mobile application Detection task；J is mobile application to be detected
The sequence number of Detection task, j are integer；ejFor the side number of the controlling stream graph of j-th of mobile application Detection task, njFor j-th of movement
Using the number of nodes of the controlling stream graph of Detection task.
Wherein, using McCabe complexity metric standards, the side of the controlling stream graph of each mobile application Detection task is obtained
Number and number of nodes, are that mobile application is pre-processed by using the reverse instrument Androguard of Android：
Androguard is an Android mobile application static analysis tools write with python, comprising multiple modules, wherein
Androgexf.py can generate the controlling stream graph of Android mobile applications.Since McCabe algorithms only need to obtain mobile application
The side number and number of nodes of controlling stream graph, while in order to accelerate execution efficiency, it is returned only to the side number and number of nodes of controlling stream graph.It is right
Android mobile application seconds grade is completed as above to pre-process, and is considerably less than the static detection time.
Step 2: the CPU weighted loads of all mobile application Detection tasks in each detection node are calculated, memory weighting
Load, the sum of queue weight load and control stream complexity weighted load value LoadGrade：
LoadGradei=Cw+Mw+Qw+Comw
Wherein, LoadGradeiThe sum of weighted load value for detection node i；I is the sequence number of detection node, is integer；
The present embodiment selects 10 mobile application static detection child nodes, and Cw is the CPU weighted load values of detection node i：Cw=wcpu*
CPUGradei；wcpuIt is the weight of CPU weighted load values；CPUGradeiFor the cpu load value of detection node i：
CPUCoresiFor the CPU core number of detection node i, unit is a；CPUGHziIt is single for the CPU frequency of detection node i
Position is GHz；CPUUsedPerciPercentage, unit % are used for the CPU of detection node i.
Mw is the memory weighted load value of detection node i：Mw=wmemory*MemoryGradei；wmemoryIt is memory weighting
The weight of load value, MemoryGradeiFor the memory load value of detection node i：
FreeMemoryiFor the current idle memory of detection node i, unit %.
Qw is the queue weight load value of detection node i：Qw=wqueue*QueueGradei；wqueueIt is queue weight load
The weight of value, QueueGradeiFor the task queue load value of detection node i：
QueueCapacityiFor detection node i task queues capacity, that is, maximum length, unit is a；QueueNumiFor inspection
Task queue current task quantity, that is, queue size of node i is surveyed, unit is a.
Comw is the control stream complexity weighted load value of detection node i：
Comw=wcomplexity*ComplexityGradei
wcomplexityFor the weight of the control stream complexity of mobile application detection node, ComplexityGradeiFor movement
Using the control stream complexity load value of detection node i:
ComplexitymaxiThe sum of the medium Complexity's to be detected of task queue for detection node i in time slot T
Maximum；ComplexityaverageiIt is averaged for the medium Complexity to be detected of task queue of detection node i in time slot T
Value, average control stream complexity is bigger, and the control stream complexity load value of the detection node is lower.
∑TComplexityjRepresent in time slot T detection node i wait in task queue all Complexity it
With n mobile application in n expression time slots T.
Wherein
All kinds of weighted values of detection node i meet：wcpu+wmemory+wqueue+wComplexity=100
wcpu,wmemory,wqueue,wComplexity∈[0,100]
Step 3: the sum of weighted load value by each detection node LoadGradeiInformation encapsulation in nodes heart beat
In；
Above-mentioned load information is wrapped in nodes heart beat by each detection node, and nodes heart beat selects JAVA classes, is provided outer
Portion's interface is set and the load information of accessed node.
Step 4: nodes heart beat is sent every time slot T once to load dispatcher, load dispatcher calculates each detection section
Next mobile application Detection task is distributed to the detection node of LoadGrade minimums by point；
Time slot T is selected 15 seconds, and central administration node module is updated the load fraction of each detection node in every 15 seconds, obtains
The load value of each detection node into magnanimity mobile application Detection task.The higher detection node priority of load value is relatively low, bears
The relatively low detection node priority of load value is higher, and the minimum detection node of load value will receive next mobile application detection and appoint
Business.
It is respectively 2,4,6,8,10 to select detection node, counts and is issuing 500 different sizes and control stream in batches again
During the application Detection task of miscellaneous degree, pass through traditional polling dispatching method, mobile application Detection task load-balancing algorithm, Yi Jiben
The load-balancing algorithm of the introducing mobile application control stream complexity of invention is compared.
As shown in figure 3, the present invention and polling dispatching method, traditional mobile application Detection task load-balancing algorithm is in difference
Under quantity engine, time statistics during carry out task distribution and compare figure, it can be seen that in inventive algorithm, each detection section
The execution efficiency of point is above other two kinds of algorithms, moreover, being improved than mobile application Detection task load-balancing algorithm
54.1% efficiency.The load balancing of distributed system is more advantageous to using the present invention and improves the execution efficiency of system.
By using the present invention, the average time of 57691 mobile application Detection tasks of statistics is 31.413s, is tested
It as a result can be with the improved mobile application Detection task load-balancing algorithm of effecting reaction to the effect of magnanimity mobile application Detection task
Rate is promoted.
Claims (1)
1. the method for load value is calculated in a kind of mobile application detection load-balancing algorithm, which is characterized in that comprise the following steps：
Step 1: being directed to each detection node, the control stream of each mobile application Detection task in the detection node is calculated respectively
Complexity Complexity；
By the controlling stream graph of each mobile application Detection task, the side number and number of nodes of the controlling stream graph are obtained, calculates control
Flow complexity Complexity；
Complexityj=ej-nj+2
ComplexityjFor the control stream complexity of j-th of mobile application Detection task；J is that mobile application detection to be detected is appointed
The sequence number of business；ejFor the side number of the controlling stream graph of j-th of mobile application Detection task, njFor j-th mobile application Detection task
The number of nodes of controlling stream graph；
Step 2: calculating the CPU weighted loads of all mobile application Detection tasks in each detection node, memory weighting is negative
It carries, the sum of queue weight load and control stream complexity weighted load value LoadGrade：
LoadGradei=Cw+Mw+Qw+Comw
Wherein, LoadGradeiThe sum of weighted load value for detection node i；I is the sequence number of detection node；Cw is detection node i
CPU weighted load values：Cw=wcpu*CPUGradei；wcpuIt is the weight of CPU weighted load values；CPUGradeiFor detection node
The cpu load value of i：
<mrow>
<msub>
<mi>CPUGrade</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>CPUCores</mi>
<mi>i</mi>
</msub>
<mo>*</mo>
<msub>
<mi>CPUGHz</mi>
<mi>i</mi>
</msub>
<mo>*</mo>
<mrow>
<mo>(</mo>
<mn>100</mn>
<mo>-</mo>
<msub>
<mi>CPUUsedPerc</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mn>100</mn>
</mfrac>
</mrow>
CPUCoresiFor the CPU core number of detection node i；CPUGHziFor the CPU frequency of detection node i；CPUUsedPerciFor inspection
The CPU for surveying node i uses percentage；
Mw is the memory weighted load value of detection node i：Mw=wmemory*MemoryGradei；wmemoryIt is memory weighted load value
Weight, MemoryGradeiFor the memory load value of detection node i：
<mrow>
<msub>
<mi>MemoryGrade</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>FreeMemory</mi>
<mi>i</mi>
</msub>
</mrow>
<mn>100</mn>
</mfrac>
</mrow>
FreeMemoryiFor the current idle memory of detection node i；
Qw is the queue weight load value of detection node i：Qw=wqueue*QueueGradei；wqueueIt is queue weight load value
Weight, QueueGradeiFor the task queue load value of detection node i：
<mrow>
<msub>
<mi>QueueGrade</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>QueueCapacity</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>QueueNum</mi>
<mi>i</mi>
</msub>
</mrow>
<mrow>
<msub>
<mi>QueueCapacity</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
</mrow>
QueueCapacityiFor the task queue maximum length of task queue capacity, that is, detection node i of detection node i；
QueueNumiFor task queue current task quantity, that is, queue size of detection node i；
Comw is the control stream complexity weighted load value of detection node i：
Comw=wcomplexity*ComplexityGradei
wcomplexityFor the weight of the control stream complexity of mobile application detection node, ComplexityGradeiFor mobile application
The control stream complexity load value of detection node i:
<mrow>
<msub>
<mi>ComplexityGrade</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>Complexity</mi>
<mrow>
<mi>max</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>Complexity</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>e</mi>
<mi>r</mi>
<mi>a</mi>
<mi>g</mi>
<mi>e</mi>
<mi>i</mi>
</mrow>
</msub>
</mrow>
<mrow>
<msub>
<mi>Complexity</mi>
<mrow>
<mi>max</mi>
<mi>i</mi>
</mrow>
</msub>
</mrow>
</mfrac>
</mrow>
ComplexitymaxiThe sum of the Complexity's of the medium task to be detected of task queue for detection node i in time slot T
Maximum；ComplexityaverageiFor the Complexity of the medium task to be detected of task queue of detection node i in time slot T
Average value：
<mrow>
<msub>
<mi>Complexity</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>e</mi>
<mi>r</mi>
<mi>a</mi>
<mi>g</mi>
<mi>e</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mi>T</mi>
</msub>
<msub>
<mi>Complexity</mi>
<mi>j</mi>
</msub>
</mrow>
<mi>n</mi>
</mfrac>
</mrow>
∑TComplexityjThe wait of expression detection node i in time slot T all Detection tasks in task queue
The sum of Complexity, n represent n mobile application in time slot T；
All kinds of weighted values of detection node i meet：wcpu+wmemory+wqueue+wComplexity=100
wcpu,wmemory,wqueue,wComplexity∈[0,100]
Step 3: the information encapsulation of the sum of weighted load value by each detection node LoadGrade is in nodes heart beat；
Step 4: nodes heart beat is sent every time slot T once to load dispatcher, load dispatcher calculates each detection node,
Next mobile application Detection task is distributed to the detection node of LoadGrade minimums.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510908946.2A CN105528283B (en) | 2015-12-10 | 2015-12-10 | A kind of method that load value is calculated in mobile application detection load-balancing algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510908946.2A CN105528283B (en) | 2015-12-10 | 2015-12-10 | A kind of method that load value is calculated in mobile application detection load-balancing algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105528283A CN105528283A (en) | 2016-04-27 |
| CN105528283B true CN105528283B (en) | 2018-05-18 |
Family
ID=55770526
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510908946.2A Active CN105528283B (en) | 2015-12-10 | 2015-12-10 | A kind of method that load value is calculated in mobile application detection load-balancing algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105528283B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109062739B (en) * | 2018-08-28 | 2022-04-01 | 深圳市网心科技有限公司 | Scheduling server, load balancing method, system and readable storage medium |
| CN109151041B (en) * | 2018-09-06 | 2021-02-26 | 网宿科技股份有限公司 | Method and device for adjusting monitoring node |
| CN110472526A (en) * | 2019-07-26 | 2019-11-19 | 南京熊猫电子股份有限公司 | A kind of edge processing apparatus and method based on recognition of face |
| CN111581068A (en) * | 2020-04-22 | 2020-08-25 | 北京华宇信息技术有限公司 | Terminal workload calculation method and device, storage medium, terminal and cloud service system |
| CN112596902A (en) * | 2020-12-25 | 2021-04-02 | 中科星通(廊坊)信息技术有限公司 | Task scheduling method and device based on CPU-GPU cooperative computing |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104364767A (en) * | 2013-03-15 | 2015-02-18 | 莫基移动公司 | Device and settings management platform |
-
2015
- 2015-12-10 CN CN201510908946.2A patent/CN105528283B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104364767A (en) * | 2013-03-15 | 2015-02-18 | 莫基移动公司 | Device and settings management platform |
Non-Patent Citations (4)
| Title |
|---|
| Detection of Repackaged Mobile Applications through a Collaborative Approach;Alessandro Aldini ET AL;《Wiley InterScience》;20141231;全文 * |
| Mobile malware detection through analysis of deviations in application network behavior;A. Shabtai ET AL;《COMPUTER & SECURITY》;20141231;全文 * |
| 一种针对流水线任务的云计算模型基于MapReduce的改进;郑宇瀚等;《中国信息通信研究新进展》;20141231;全文 * |
| 异构网络中基于载波聚合的负载均衡研究;李祺;《中国优秀硕士学位论文全文数据库（电子期刊）》;20150831;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105528283A (en) | 2016-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105528283B (en) | A kind of method that load value is calculated in mobile application detection load-balancing algorithm | |
| CN105718479B (en) | Execution strategy generation method and device under cross-IDC big data processing architecture | |
| US20170329643A1 (en) | Distributed node intra-group task scheduling method and system | |
| CN102708011B (en) | Adaptive task scheduling method of cloud computing platform | |
| US8332873B2 (en) | Dynamic application instance placement in data center environments | |
| US8387059B2 (en) | Black-box performance control for high-volume throughput-centric systems | |
| EP2278502A1 (en) | Deleting data stream overload | |
| CN104978236A (en) | HDFS load source and sink node selection method based on multiple measurement indexes | |
| Alsadie | A metaheuristic framework for dynamic virtual machine allocation with optimized task scheduling in cloud data centers | |
| US9063750B2 (en) | Mapping high-performance computing applications to platforms | |
| CN110413389A (en) | A kind of task schedule optimization method under the unbalanced Spark environment of resource | |
| Mekala et al. | A DRL-based service offloading approach using DAG for edge computational orchestration | |
| US8813087B2 (en) | Managing a workload in a cluster of computing systems with multi-type operational resources | |
| CN108881432A (en) | Cloud computing cluster load dispatching method based on GA algorithm | |
| GB2507038A (en) | Scheduling jobs weighted according to the memory usage using a knapsack problem. | |
| Hu et al. | Multi-objective container deployment on heterogeneous clusters | |
| CN112261120B (en) | Cloud-side cooperative task unloading method and device for power distribution internet of things | |
| CN101634953A (en) | Method and device for calculating search space, and method and system for self-adaptive thread scheduling | |
| CN106878389B (en) | Method and device for resource scheduling in cloud system | |
| US8180823B2 (en) | Method of routing messages to multiple consumers | |
| CN114519006A (en) | Test method, device, equipment and storage medium | |
| Khomonenko et al. | Stochastic models for cloud computing performance evaluation | |
| CN110266757A (en) | A kind of edge calculations method for optimizing resources of oriented mission collaboration scene | |
| Kumar et al. | Load balancing algorithm to minimize the makespan time in cloud environment | |
| Feng et al. | A dynamic load balancing algorithm based on distributed database system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |