CN106656841B - Nova component message queue dispatching method, system and exchange optimizer based on Openstack - Google Patents
Nova component message queue dispatching method, system and exchange optimizer based on Openstack Download PDFInfo
- Publication number
- CN106656841B CN106656841B CN201710108634.2A CN201710108634A CN106656841B CN 106656841 B CN106656841 B CN 106656841B CN 201710108634 A CN201710108634 A CN 201710108634A CN 106656841 B CN106656841 B CN 106656841B
- Authority
- CN
- China
- Prior art keywords
- message queue
- nova
- message
- nova component
- queue
- 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 21
- 238000012545 processing Methods 0.000 claims abstract description 44
- 238000011156 evaluation Methods 0.000 claims description 42
- 239000002131 composite material Substances 0.000 claims description 37
- 230000001105 regulatory effect Effects 0.000 claims description 12
- 238000001514 detection method Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/50—Queue scheduling
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/50—Queue scheduling
- H04L47/62—Queue scheduling characterised by scheduling criteria
- H04L47/625—Queue scheduling characterised by scheduling criteria for service slots or service orders
- H04L47/6255—Queue scheduling characterised by scheduling criteria for service slots or service orders queue load conditions, e.g. longest queue first
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Nova component message queue dispatching method, system and the exchange optimizer that the invention discloses a kind of based on Openstack, comprising: obtain the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount；Current optimal N ova component message queue is determined according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount；By new Nova scheduling message to current optimal N ova component message queue.Technical solution of the present invention selects current optimal N ova component message queue according to the real-time condition of each Nova component message queue, and Nova message is sent to current optimal N ova component message queue, so as to effectively avoid the occurrence of network congestion.
Description
Technical field
The present invention relates to field of communication technology, in particular to a kind of Nova component message queue tune based on Openstack
Degree method and scheduling system.
Background technique
Cloud computing has become the important trend of global IT industry development, and cloud infrastructure administrative skill is based on virtual
Change technology realize for government/enterprise/personal use cloud resource platform, it is by a series of virtual moneys such as network, storage, calculating
Source is combined, and is used as needed for government/enterprise/personal user.Currently, cloud infrastructure management have become world government/
The generally acknowledged development priority of enterprise's cloud industrial circle；The cloud industry giant of international and national accelerates technical research, enterprise transformation and alliance
Cooperation is to seize the dominant right and emerging market space that cloud computing develops.
Existing cloud computing application system with cloud computing application business rapid growth, with generation message queue block
The problems such as become increasingly conspicuous.Fig. 1 is the structural schematic diagram that Nova component message queue in the prior art dispatches system, such as Fig. 1 institute
Show, comprising: exchange optimizer 8 and several Nova component message queues 9 independent of each other, existing exchange optimizer 8 are main
Nova message is forwarded using pass-through modes such as broadcast type, direct-type, existing this forwarding Nova message is easy to appear Nova
The high problem of 9 degree of congestion of component message queue.
In this case, in face of the cloud computing growth requirement being increasingly urgent to, the Nova component message based on Openstack
Queue Optimal Scheduling is of great significance for the rapid sustainable development of cloud computing.
Summary of the invention
The present invention is directed at least solve one of the technical problems existing in the prior art, proposes one kind and be based on
Nova component message queue dispatching method, system and the exchange optimizer of Openstack.
To achieve the above object, the Nova component message queue dispatching party based on Openstack that the present invention provides a kind of
Method, comprising:
Step S1, each Nova component message queue current Message Payload ratio and unit time Message Processing amount are obtained；
Step S2, true according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
Settled preceding optimal N ova component message queue；
Step S3, by new Nova scheduling message to the current optimal N ova component message queue, for it is described most
Excellent Nova component message queue is stored and processed.
Optionally, the step S2 includes:
Step S201, according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
Generate the initial message queue composite vector of each Nova component message queue
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue,For i-th of Nova group
The current Message Payload ratio of part message queue,For the current unit time message of i-th of Nova component message queue
Treating capacity；
Step S202, iterative model is established, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；N is Nova component message
The sum of queue,Calculated message queue Optimized Operation evaluation of estimate when iteration secondary for kth；Vi kFor corresponding to kth time
When iteration calculated i-th of Nova component message queue Nova component message queue composite vector, The Message Payload ratio of calculated i-th of Nova component message queue, β when for corresponding to kth time iterationi kFor correspondence
When kth time iteration calculated i-th of Nova component message queue unit time Message Processing amount；xi kIt indicates in kth
Whether i-th of Nova component message queue be in processing newcomer's Nova message when secondary iteration, if i-th of Nova when kth time iteration
Component message queue is handling newcomer's Nova message, then xi kValue is 1, otherwise, xi kValue is 0；θ and σ is respectively that message is negative
Carry ratio and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,1)；When k value is 1, When k value is greater than 1,It is compound for Nova component message queue
Vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For
Nova component message queue composite vector Vi k-1Second-order moment around mean,
Step S203, it is directed to above-mentioned iterative model, inputs initial value k=0, and calculate f0(xi k) value；
Step S204, each Nova component message when being carried out to k plus 1 processing, and calculating kth time iteration according to iterative equations
Queue composite vector Vi k；
Step S205, according to calculated each Nova component message queue composite vector Vi k, when calculating kth time iteration pair
The message queue Optimized Operation evaluation of estimate f answeredk(xi k)；
Step S206, judge corresponding message queue Optimized Operation evaluation of estimate f when kth time iterationk(xi k) whether meet such as
Lower condition:
Wherein, (0,1) ε ∈；
If satisfied, thening follow the steps S208；Otherwise, step S207 is executed.
Step S207, judge whether k is less than d；
If judging, k is less than d, continues to execute above-mentioned steps S204；Otherwise, step S208 is executed.
Step S208, the value of current k is exported, and determines that selected Nova component message queue is when kth time iteration
The optimal N ova component message queue.
Optionally, further includes:
Step S4, whether the memory space for detecting each Nova component message queue has expired；
The full Nova component message queue of memory space if it exists, thens follow the steps S5；
Step S5, part Nova message to be processed is sent to by the full Nova component message queue of control memory space
In pre-set shared message queue, so that shared message queue handles the Nova message to be processed.
Optionally, further includes:
Step S6, whether the memory space for detecting the shared message queue has expired；
If it is detected that the memory space of the shared message queue has been expired, S7 is thened follow the steps；
Step S7, it controls the shared message queue and part Nova message to be processed is sent to exchange optimizer, with
Scheduling is re-started by Nova message to be processed is received for the exchange optimizer.
To achieve the above object, the present invention also provides a kind of exchange optimizers, comprising:
First obtains module, disappears for obtaining the current Message Payload ratio of each Nova component message queue and unit time
Cease treating capacity；
Determining module, at according to each Nova component message queue current Message Payload ratio and unit time message
Reason amount determines current optimal N ova component message queue；
Scheduler module, for by new Nova scheduling message to the current optimal N ova component message queue, for
The optimal N ova component message queue is stored and processed.
Optionally, the determining module includes:
First generation unit, for being disappeared according to the current Message Payload ratio of each Nova component message queue and unit time
Breath treating capacity generates the initial message queue composite vector of each Nova component message queue
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue,For i-th of Nova group
The current Message Payload ratio of part message queue,For the current unit time message of i-th of Nova component message queue
Treating capacity；
Modeling unit, for establishing iterative model, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；N is Nova component message
The sum of queue, fk(xi k) calculated message queue Optimized Operation evaluation of estimate when being kth time iteration；Vi kFor corresponding to kth time
When iteration calculated i-th of Nova component message queue Nova component message queue composite vector, The Message Payload ratio of calculated i-th of Nova component message queue, β when for corresponding to kth time iterationi kFor correspondence
When kth time iteration calculated i-th of Nova component message queue unit time Message Processing amount；xi kIt indicates in kth
Whether i-th of Nova component message queue be in processing newcomer's Nova message when secondary iteration, if i-th of Nova when kth time iteration
Component message queue is handling newcomer's Nova message, then xi kValue is 1, otherwise, xi kValue is 0；θ and σ is respectively that message is negative
Carry ratio and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,1)；When k value is 1, When k value is greater than 1,It is multiple for Nova component message queue
Resultant vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For
Nova component message queue composite vector Vi k-1Second-order moment around mean,
Initial value input unit inputs initial value k=0, and calculate f for being directed to above-mentioned iterative model0(xi k)
Value；
Vector calculation unit, for being carried out to k plus 1 processing, and while calculating kth time iteration according to iterative equations each Nova
Component message queue composite vector Vi k；
Evaluation of estimate computing unit, for according to calculated each Nova component message queue composite vector Vi k, calculate kth
Corresponding message queue Optimized Operation evaluation of estimate f when secondary iterationk(xi k)；
First judging unit, corresponding message queue Optimized Operation evaluation of estimate f when for judging kth time iterationk(xi k) be
It is no to meet following condition:
Wherein, (0,1) ε ∈；
Second judgment unit, if judging that corresponding message queue optimization is adjusted when kth time iteration for the first judging unit
Spend evaluation of estimate fk(xi k) when being unsatisfactory for corresponding conditions, judge whether k is less than d；
Output unit is adjusted for message queue optimization corresponding when the first judging unit judges kth time iteration
Spend evaluation of estimate fk(xi k) when meeting corresponding conditions or when the second judgment unit judges that k is equal to d, export current k's
Value, and determine that selected Nova component message queue is the optimal N ova component message queue when kth time iteration.
Optionally, further includes:
Whether first detection module, the memory space for detecting each Nova component message queue have expired；
First control module, if detecting that there are the full Nova components of memory space to disappear for the first detection module
When ceasing queue, part Nova message to be processed is sent to by the full Nova component message queue of control memory space to be set in advance
In the shared message queue set, so that shared message queue handles the Nova message to be processed.
Optionally, further includes:
Whether the second detection module, the memory space for detecting the shared message queue have expired；
Second control module, for if it is detected that when the memory space of the shared message queue has been expired, described in control
Part Nova message to be processed is sent to exchange optimizer by shared message queue, so that the exchange optimizer will receive
Nova message to be processed re-starts scheduling.
To achieve the above object, the Nova component message queue scheduling based on Openstack that the present invention also provides a kind of
System, including above-mentioned exchange optimizer.
Optionally, further includes: several Nova component message queues and at least one shared message queue.
The invention has the following advantages:
Nova component message queue dispatching method, system and the exchange that the present invention provides a kind of based on Openstack are excellent
Change device, by each Nova component message queue of acquisition current Message Payload ratio and unit time Message Processing amount, and according to
The current Message Payload ratio of each Nova component message queue and unit time Message Processing amount determine current optimal N ova component
Message queue；Finally by new Nova scheduling message to current optimal N ova component message queue, for optimal N ova component
Message queue is stored and processed, and can effectively avoid and network congestion occurs.In addition, the setting of shared message queue, it can be effective
Avoid loss of data.
Detailed description of the invention
Fig. 1 is the structural schematic diagram that Nova component message queue in the prior art dispatches system；
Fig. 2 is a kind of Nova component message queue dispatching method based on Openstack that the embodiment of the present invention one provides
Flow chart；
Fig. 3 is the flow chart of interative computation in the present invention；
Fig. 4 is a kind of Nova component message queue dispatching method based on Openstack provided by Embodiment 2 of the present invention
Flow chart；
Fig. 5 is a kind of structural schematic diagram for exchange optimizer that the embodiment of the present invention three provides；
Fig. 6 is that a kind of Nova component message queue based on Openstack that the embodiment of the present invention four provides dispatches system
Structural schematic diagram.
Specific embodiment
To make those skilled in the art more fully understand technical solution of the present invention, the present invention is mentioned with reference to the accompanying drawing
A kind of Nova component message queue dispatching method based on Openstack, system and the exchange optimizer supplied is described in detail.
Fig. 2 is a kind of Nova component message queue dispatching method based on Openstack that the embodiment of the present invention one provides
Flow chart, as shown in Fig. 2, the dispatching method includes:
Step S1, each Nova component message queue current Message Payload ratio and unit time Message Processing amount are obtained.
Wherein, the current Message Payload ratio of i-th of Nova component message queue is denoted asI-th of Nova component disappears
The current unit time Message Processing amount of breath queue is denoted as
Step S2, true according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
Settled preceding optimal N ova component message queue.
Fig. 3 is the flow chart of interative computation in the present invention, as shown in figure 3, optionally, step S2 includes:
Step S201, according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
Generate the initial message queue composite vector of each Nova component message queue
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue.
Step S202, iterative model is established, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；
N is the sum of Nova component message queue；
fk(xi k) calculated message queue Optimized Operation evaluation of estimate when being kth time iteration；
Vi kThe Nova component message team of calculated i-th of Nova component message queue when for corresponding to kth time iteration
Column composite vector,
The Message Payload ratio of calculated i-th of Nova component message queue when for corresponding to kth time iteration；
βi kWhen for corresponding to kth time iteration at the unit time message of calculated i-th of Nova component message queue
Reason amount；
xi kIndicate whether i-th of Nova component message queue is handling newcomer's Nova message at the kth iteration, if the
I-th of Nova component message queue is handling newcomer's Nova message when k iteration, then xi kValue is 1, otherwise, xi kValue is
0；
θ and σ is respectively Message Payload ratio and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,
1)；
When k value is 1,
When k value is greater than 1,For Nova component message queue composite vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For Nova component message queue composite vector Vi k -1Second-order moment around mean,
Step S203, it is directed to above-mentioned iterative model, inputs initial value k=0, and calculate f0(xi k) value.
Detailed process is as follows by step S203:
For n Nova component message queue, the forwarding strategy of the Nova information has n kind, x1 0、x2 0......xn 0Tax
It is as shown in table 1 below to be worth situation；
1. forwarding strategy table of table
| x1 0 | x2 0 | ...... | xn 0 | |
| Strategy 1 | 1 | 0 | ...... | 0 |
| Strategy 2 | 0 | 1 | ...... | 0 |
| ...... | ...... | ...... | ...... | ...... |
| Tactful n | 0 | 0 | ...... | 1 |
By above-mentioned table 1 as it can be seen that for each strategy, x1 0、x2 0......xn 0In only one value be 1 (corresponding
Nova component message queue handles Nova message), and the x that value is 1 in Different Strategiesi kIt is different.
f0(xi 0) indicate in the 0th iteration, its evaluation of estimate is calculated separately for n kind strategy, and choose evaluation of estimate most
Preference policy when a kind of small strategy is as the 0th iteration, meanwhile, based on the calculated evaluation of estimate conduct of preference policy institute
Calculated message queue Optimized Operation evaluation of estimate when the 0th iteration.
Step S204, each Nova component message when being carried out to k plus 1 processing, and calculating kth time iteration according to iterative equations
Queue composite vector Vi k。
Step S205, according to calculated each Nova component message queue composite vector Vi k, when calculating kth time iteration pair
The message queue Optimized Operation evaluation of estimate f answeredk(xi k)。
It should be noted that similar when being 0 with k value, the corresponding message queue Optimized Operation when calculating kth time iteration
When evaluation of estimate, n kind strategy (as shown in table 1) is equally existed.In step S205, at the kth iteration, for n kind strategy point
Preference policy when not calculating its evaluation of estimate, and choosing a kind of secondary as the kth iteration of the smallest strategy of evaluation of estimate, meanwhile, it is based on
Calculated message queue Optimized Operation evaluation of estimate f when the calculated evaluation of estimate of preference policy institute is as kth time iterationk
(xi k)。
By above content as it can be seen that every progress an iteration calculating, can determine this time iterative calculation it is selected go out
Preference policy (the Nova component message queue of selection).
Step S206, judge corresponding message queue Optimized Operation evaluation of estimate f when kth time iterationk(xi k) whether meet such as
Lower condition:
Wherein, ε is preset value, and ε ∈ (0,1)；
If satisfied, thening follow the steps S208；Otherwise, step S207 is executed.
Step S207, judge whether k is less than d；
If judging, k is less than d, continues to execute above-mentioned steps S204；Otherwise, step S208 is executed.
Step S208, the value of current k is exported, and determines that selected Nova component message queue is when kth time iteration
Optimal N ova component message queue.
Step S3, by new Nova scheduling message to current optimal N ova component message queue, for optimal N ova group
Part message queue is stored and processed.
The embodiment of the present invention one provides a kind of Nova component message queue dispatching method based on Openstack, passes through
The current Message Payload ratio of each Nova component message queue and unit time Message Processing amount are obtained, and according to each Nova component
The current Message Payload ratio of message queue and unit time Message Processing amount determine current optimal N ova component message queue, most
Afterwards by new Nova scheduling message to current optimal N ova component message queue, for the progress of optimal N ova component message queue
Storage and processing.Technical solution of the present invention can be selected current optimal according to the real-time condition of each Nova component message queue
Nova component message queue, so as to effectively avoid the occurrence of network congestion.
Embodiment two
Fig. 4 is a kind of Nova component message queue dispatching method based on Openstack provided by Embodiment 2 of the present invention
Flow chart, as shown in figure 4, the dispatching method further includes other than including the steps that S1 in above-described embodiment one~step S3
Step S4~step S7 is below only described in detail step S4~step S7.
Step S4, whether the memory space for detecting each Nova component message queue has expired.
It finds in the prior art, after the memory space of some Nova component message queue has been expired, if continuing to this
Nova component message queue forwards Nova message, then the Nova component message queue will appear the packet loss (part in queue
Nova message or the Nova message newly forwarded are dropped) the phenomenon that.
In order to solve the above technical problems, the present invention is additionally provided at least one in Nova component message queue scheduling system
Shared message queue.
In step s 4, if it is detected that the Nova component message queue full there are memory space, thens follow the steps S5；If
It detects the Nova component message queue full there is no memory space, then shows to be not in packet loss.
Step S5, part Nova message to be processed is sent to by the full Nova component message queue of control memory space
In pre-set shared message queue, so that shared message queue handles Nova message to be processed.
In the present embodiment, by the way that the part Nova message in the full Nova component message queue of memory space is sent to
It is handled in pre-set shared message queue, can effectively avoid and the phenomenon that packet loss occur.
Step S6, whether the memory space for detecting shared message queue has expired.
If it is detected that the memory space of shared message queue has been expired, the shared message queue is easy to appear the wind of packet loss at this time
Danger, thens follow the steps S7；Otherwise, show that the shared message queue will not packet loss.
Step S7, part Nova message to be processed is sent to exchange optimizer by control shared message queue, for handing over
It changes optimizer and re-starts scheduling for Nova message to be processed is received.
It should be noted that can continue to repeat above-mentioned steps S1 after step S7, for exchanging optimizer pair
Nova message transmitted by shared message queue re-starts scheduling, with distribution to current optimal N ova component message queue
In.
The technical solution of the embodiment of the present invention two not only can effectively avoid the occurrence of network congestion, also can effectively avoid system and go out
The risk of existing packet loss.
Embodiment three
Fig. 5 is a kind of structural schematic diagram for exchange optimizer that the embodiment of the present invention three provides, as shown in figure 5, the exchange
The Nova component message queue scheduling based on Openstack that optimizer is used to execute above-described embodiment one or embodiment two provides
Method, the exchange optimizer include:
First obtains module 1, for obtaining each Nova component message queue 9 current Message Payload ratio and unit time
Message Processing amount.
Determining module 2, for the Message Payload ratio and unit time message current according to each Nova component message queue 9
Treating capacity determines current optimal N ova component message queue 9.
Scheduler module 3 is used for new Nova scheduling message to current optimal N ova component message queue 9, for most
Excellent Nova component message queue 9 is stored and processed.
It should be noted that the first acquisition module 1 in the present embodiment is for executing above-described embodiment one and embodiment two
In step S1, determining module 2 is used to execute step S2 in above-described embodiment one and embodiment two, and scheduler module 3 is for holding
Step S3 in row above-described embodiment one and embodiment two can be found in above-described embodiment for the specific descriptions of above three module
Description in one, details are not described herein again.
Optionally it is determined that module includes:
First generation unit 201, when for according to each Nova component message queue 9 current Message Payload ratio and unit
Between Message Processing amount generate the initial message queue composite vector of each Nova component message queue 9
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue,For i-th of Nova group
The current Message Payload ratio of part message queue,For the current unit time message of i-th of Nova component message queue
Treating capacity；
Modeling unit 202, for establishing iterative model, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；N is Nova component message
The sum of queue, fk(xi k) calculated message queue Optimized Operation evaluation of estimate when being kth time iteration；Vi kFor corresponding to kth time
When iteration calculated i-th of Nova component message queue Nova component message queue composite vector, The Message Payload ratio of calculated i-th of Nova component message queue, β when for corresponding to kth time iterationi kFor correspondence
When kth time iteration calculated i-th of Nova component message queue unit time Message Processing amount；xi kIt indicates in kth
Whether i-th of Nova component message queue be in processing newcomer's Nova message when secondary iteration, if i-th of Nova when kth time iteration
Component message queue is handling newcomer's Nova message, then xi kValue is 1, otherwise, xi kValue is 0；θ and σ is respectively that message is negative
Carry ratio and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,1)；When k value is 1,Bi k=E [Vi k-1-E(Vi k-1)]2；When k value is greater than 1,It is multiple for Nova component message queue
Resultant vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For
Nova component message queue composite vector Vi k-1Second-order moment around mean,
Initial value input unit 203 inputs initial value k=0, and calculate f for being directed to above-mentioned iterative model0(xi k)
Value；
Vector calculation unit 204, for carrying out to k plus 1 processing, and while calculating kth time iteration according to iterative equations, is each
Nova component message queue composite vector Vi k；
Evaluation of estimate computing unit 205, for according to calculated each Nova component message queue composite vector Vi k, calculate
Corresponding message queue Optimized Operation evaluation of estimate f when kth time iterationk(xi k)；
First judging unit 206, corresponding message queue Optimized Operation evaluation of estimate f when for judging kth time iterationk(xi k)
Whether following condition is met:
Wherein, (0,1) ε ∈；
Second judgment unit 207, if judging that corresponding message queue optimizes when kth time iteration for the first judging unit
Dispatch evaluation of estimate fk(xi k) when being unsatisfactory for corresponding conditions, judge whether k is less than d；
Output unit 208, for message queue optimization corresponding when the first judging unit judges kth time iteration
Dispatch evaluation of estimate fk(xi k) when meeting corresponding conditions or when the second judgment unit judges that k is equal to d, export current k
Value, and determine that selected Nova component message queue is the optimal N ova component message queue when kth time iteration.
It should be noted that the first generation unit 201 in the present embodiment is used to execute the step in above-described embodiment one
S201, modeling unit 202 are used to execute the step S202 in above-described embodiment one, and initial value input unit 203 is for executing
The step S203 in embodiment one is stated, vector calculation unit 204 is used to execute the step S204 in above-described embodiment one, evaluation of estimate
Computing unit 205 is used to execute the step S205 in above-described embodiment one, and the first judging unit 206 is for executing above-described embodiment
Step S206 in one, second judgment unit 207 are used to execute the step S207 in above-described embodiment one, and output unit 208 is used
Step S208 in execution above-described embodiment one.The specific descriptions of above-mentioned each unit can be found in above-described embodiment one
Description, details are not described herein again.
To solve the problems, such as that packet loss occurs in system in the prior art, optionally, the exchange optimizer further include:
Whether first detection module 4, the memory space for detecting each Nova component message queue 9 have expired；
First control module 5, if detecting the Nova component message full there are memory space for first detection module 4
When queue 9, part Nova message to be processed is sent to by the full Nova component message queue 9 of control memory space to be set in advance
In the shared message queue 10 (attached drawing is merely exemplary to depict a shared message queue 10) set, for shared message queue 10
Nova message to be processed is handled.
Optionally, further includes:
Whether the second detection module 6, the memory space for detecting shared message queue 10 have expired；
Second control module 7, for if it is detected that control is shared to disappear when the memory space of shared message queue 10 has been expired
Cease queue 10 and part Nova message to be processed be sent to exchange optimizer, for exchange optimizer will receive it is to be processed
Nova message re-starts scheduling.
It should be noted that the first detection module 4 in the present embodiment is used to execute the step S4 in above-described embodiment two,
First control module 5 is used to execute the step S5 in above-described embodiment two, and the second detection module 6 is for executing above-described embodiment two
In step S6, the second control module 7 is used to execute step S7 in above-described embodiment two.For specifically retouching for above-mentioned each module
The description that can be found in above-described embodiment two is stated, details are not described herein again.
Example IV
Fig. 6 is that a kind of Nova component message queue based on Openstack that the embodiment of the present invention four provides dispatches system
Structural schematic diagram, as shown in fig. 6, the scheduling system include exchange optimizer 8, the exchange optimizer use above-described embodiment three
In exchange optimizer 8, particular content can be found in the description in above-described embodiment three, and details are not described herein again.
Optionally, further includes: several Nova component message queues 9 and at least one shared message queue 10.This implementation
In example, system can effectively avoid by setting shared message queue and the phenomenon that packet loss occur.
It is understood that the principle that embodiment of above is intended to be merely illustrative of the present and the exemplary implementation that uses
Mode, however the present invention is not limited thereto.For those skilled in the art, essence of the invention is not being departed from
In the case where mind and essence, various changes and modifications can be made therein, these variations and modifications are also considered as protection scope of the present invention.
Claims (8)
1. a kind of Nova component message queue dispatching method based on Openstack characterized by comprising
Step S1, each Nova component message queue current Message Payload ratio and unit time Message Processing amount are obtained；
Step S2, worked as according to the current Message Payload ratio of each Nova component message queue and the determination of unit time Message Processing amount
Preceding optimal N ova component message queue；
Step S3, by new Nova scheduling message to the current optimal N ova component message queue, for described optimal
Nova component message queue is stored and processed；
Wherein, the step S2 includes:
Step S201, it is generated according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
The initial message queue composite vector of each Nova component message queue
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue,Disappear for i-th of Nova component
The current Message Payload ratio of queue is ceased,For the current unit time Message Processing of i-th of Nova component message queue
Amount；
Step S202, iterative model is established, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；N is Nova component message queue
Sum, fk(xi k) calculated message queue Optimized Operation evaluation of estimate when being kth time iteration；Vi kFor corresponding to kth time iteration
When calculated i-th of Nova component message queue Nova component message queue composite vector,
The Message Payload ratio of calculated i-th of Nova component message queue, β when for corresponding to kth time iterationi kFor corresponding to
When k iteration calculated i-th of Nova component message queue unit time Message Processing amount；xi kIt indicates to change in kth time
For when i-th Nova component message queue whether in processing newcomer Nova message, if i-th of Nova component when the secondary iteration of kth
Message queue is handling newcomer's Nova message, then xi kValue is 1, otherwise, xi kValue is 0；θ and σ is respectively Message Payload ratio
Rate and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,1)；When k value is 1, When k value is greater than 1,It is multiple for Nova component message queue
Resultant vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For
Nova component message queue composite vector Vi k-1Second-order moment around mean,
Step S203, it is directed to above-mentioned iterative model, inputs initial value k=0, and calculate f0(xi k) value；
Step S204, each Nova component message queue when being carried out to k plus 1 processing, and calculating kth time iteration according to iterative equations
Composite vector Vi k；
Step S205, according to calculated each Nova component message queue composite vector Vi k, calculate corresponding when kth time iteration
Message queue Optimized Operation evaluation of estimate fk(xi k)；
Step S206, judge corresponding message queue Optimized Operation evaluation of estimate f when kth time iterationk(xi k) whether meet following item
Part:
Wherein, (0,1) ε ∈；
If satisfied, thening follow the steps S208；Otherwise, step S207 is executed；
Step S207, judge whether k is less than d；
If judging, k is less than d, continues to execute above-mentioned steps S204；Otherwise, step S208 is executed；
Step S208, the value of current k is exported, and determines that selected Nova component message queue is described when kth time iteration
Optimal N ova component message queue.
2. dispatching method according to claim 1, which is characterized in that further include:
Step S4, whether the memory space for detecting each Nova component message queue has expired；
The full Nova component message queue of memory space if it exists, thens follow the steps S5；
Step S5, part Nova message to be processed is sent in advance by the full Nova component message queue of control memory space
In the shared message queue of setting, so that shared message queue handles the Nova message to be processed.
3. dispatching method according to claim 2, which is characterized in that further include:
Step S6, whether the memory space for detecting the shared message queue has expired；
If it is detected that the memory space of the shared message queue has been expired, S7 is thened follow the steps；
Step S7, it controls the shared message queue and part Nova message to be processed is sent to exchange optimizer, for institute
It states exchange optimizer and re-starts scheduling for Nova message to be processed is received.
4. a kind of exchange optimizer characterized by comprising
First obtains module, for obtaining at each Nova component message queue current Message Payload ratio and unit time message
Reason amount；
Determining module, for according to the current Message Payload ratio of each Nova component message queue and unit time Message Processing amount
Determine current optimal N ova component message queue；
Scheduler module is used for new Nova scheduling message to the current optimal N ova component message queue, for described
Optimal N ova component message queue is stored and processed；
Wherein, the determining module includes:
First generation unit, at according to each Nova component message queue current Message Payload ratio and unit time message
Reason amount generates the initial message queue composite vector of each Nova component message queue
Wherein, Vi 0For the initial message queue composite vector of i-th of Nova component message queue,Disappear for i-th of Nova component
The current Message Payload ratio of queue is ceased,For the current unit time Message Processing of i-th of Nova component message queue
Amount；
Modeling unit, for establishing iterative model, wherein
Iteration evaluation function:
Iterative equations:
Wherein, k is iterative steps and k ∈ [1, d], d are pre-set maximum number of iterations；N is Nova component message queue
Sum, fk(xi k) calculated message queue Optimized Operation evaluation of estimate when being kth time iteration；Vi kFor corresponding to kth time iteration
When calculated i-th of Nova component message queue Nova component message queue composite vector,
The Message Payload ratio of calculated i-th of Nova component message queue, β when for corresponding to kth time iterationi kFor corresponding to
When k iteration calculated i-th of Nova component message queue unit time Message Processing amount；xi kIt indicates to change in kth time
For when i-th Nova component message queue whether in processing newcomer Nova message, if i-th of Nova component when the secondary iteration of kth
Message queue is handling newcomer's Nova message, then xi kValue is 1, otherwise, xi kValue is 0；θ and σ is respectively Message Payload ratio
Rate and unit time Message Processing amount regulatory factor, and θ ∈ (0,1), σ ∈ (0,1)；When k value is 1, When k value is greater than 1,It is multiple for Nova component message queue
Resultant vector Vi k-1Variation cube be worth expectation,ζ is regulatory factor, and ζ ∈ (0,1),For
Nova component message queue composite vector Vi k-1Second-order moment around mean,
Initial value input unit inputs initial value k=0, and calculate f for being directed to above-mentioned iterative model0(xi k) value；
Vector calculation unit, for being carried out to k plus 1 processing, and while calculating kth time iteration according to iterative equations each Nova component
Message queue composite vector Vi k；
Evaluation of estimate computing unit, for according to calculated each Nova component message queue composite vector Vi k, calculate kth time repeatedly
For when corresponding message queue Optimized Operation evaluation of estimate fk(xi k)；
First judging unit, corresponding message queue Optimized Operation evaluation of estimate f when for judging kth time iterationk(xi k) whether full
The following condition of foot:
Wherein, (0,1) ε ∈；
Second judgment unit, if judging that corresponding message queue Optimized Operation is commented when kth time iteration for the first judging unit
It is worth fk(xi k) when being unsatisfactory for corresponding conditions, judge whether k is less than d；
Output unit, for message queue Optimized Operation evaluation of estimate f corresponding when the first judging unit judges kth time iterationk
(xi k) when meeting corresponding conditions or when the second judgment unit judges that k is equal to d, the value of current k is exported, and determine
Selected Nova component message queue is the optimal N ova component message queue when kth time iteration.
5. exchange optimizer according to claim 4, which is characterized in that further include:
Whether first detection module, the memory space for detecting each Nova component message queue have expired；
First control module, if detecting the Nova component message team full there are memory space for the first detection module
When column, part Nova message to be processed is sent to pre-set by the full Nova component message queue of control memory space
In shared message queue, so that shared message queue handles the Nova message to be processed.
6. exchange optimizer according to claim 5, which is characterized in that further include:
Whether the second detection module, the memory space for detecting the shared message queue have expired；
Second control module, for if it is detected that being controlled described shared when the memory space of the shared message queue has been expired
Part Nova message to be processed is sent to exchange optimizer by message queue, so that the exchange optimizer will be received wait locate
The Nova message of reason re-starts scheduling.
7. a kind of Nova component message queue based on Openstack dispatches system, which is characterized in that wanted including such as aforesaid right
Seek any exchange optimizer in 4-6.
8. scheduling system according to claim 7, which is characterized in that further include: several Nova component message queues and
At least one shared message queue.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710108634.2A CN106656841B (en) | 2017-02-27 | 2017-02-27 | Nova component message queue dispatching method, system and exchange optimizer based on Openstack |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710108634.2A CN106656841B (en) | 2017-02-27 | 2017-02-27 | Nova component message queue dispatching method, system and exchange optimizer based on Openstack |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106656841A CN106656841A (en) | 2017-05-10 |
| CN106656841B true CN106656841B (en) | 2019-08-30 |
Family
ID=58846872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710108634.2A Active CN106656841B (en) | 2017-02-27 | 2017-02-27 | Nova component message queue dispatching method, system and exchange optimizer based on Openstack |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106656841B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107992394A (en) * | 2017-11-28 | 2018-05-04 | 顺丰科技有限公司 | Monitoring method, device, system, equipment and the storage medium of real time data processing |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105306277A (en) * | 2015-11-19 | 2016-02-03 | 中国建设银行股份有限公司 | Message scheduling method and message scheduling device for message queues |
| CN105975340A (en) * | 2016-03-24 | 2016-09-28 | 国云科技股份有限公司 | Deployment algorithm of application distribution of virtual machine |
| CN106250249A (en) * | 2016-08-09 | 2016-12-21 | 广州唯品会信息科技有限公司 | The control method of message queue and device |
-
2017
- 2017-02-27 CN CN201710108634.2A patent/CN106656841B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105306277A (en) * | 2015-11-19 | 2016-02-03 | 中国建设银行股份有限公司 | Message scheduling method and message scheduling device for message queues |
| CN105975340A (en) * | 2016-03-24 | 2016-09-28 | 国云科技股份有限公司 | Deployment algorithm of application distribution of virtual machine |
| CN106250249A (en) * | 2016-08-09 | 2016-12-21 | 广州唯品会信息科技有限公司 | The control method of message queue and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106656841A (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107766135B (en) | Task allocation method based on particle swarm optimization and simulated annealing optimization in moving cloud | |
| CN103309738B (en) | User job dispatching method and device | |
| CN109561148A (en) | Distributed task dispatching method in edge calculations network based on directed acyclic graph | |
| CN104780213B (en) | A kind of master-salve distributed figure processing system load dynamic optimization method | |
| Xie et al. | Dynamic computation offloading in IoT fog systems with imperfect channel-state information: A POMDP approach | |
| CN111182582A (en) | Multitask distributed unloading method facing mobile edge calculation | |
| CN104754053B (en) | A kind of distributed software defines network and the wherein method of dynamic control controller | |
| CN108566659A (en) | A kind of online mapping method of 5G networks slice based on reliability | |
| CN104317646B (en) | Based on cloud data center dispatching method of virtual machine under OpenFlow frameworks | |
| WO2023245740A1 (en) | Fourth-party logistics transportation edge planning method based on ant colony optimization algorithm | |
| CN105808346B (en) | A kind of method for scheduling task and device | |
| CN102621887B (en) | Garment sewing process analysis and combination method | |
| CN109976915A (en) | The optimization method and system of side cloud coordination requirement based on edge calculations | |
| CN110149401A (en) | It is a kind of for optimizing the method and system of edge calculations task | |
| CN105119279B (en) | A kind of distributed power source planing method and its system | |
| CN106656841B (en) | Nova component message queue dispatching method, system and exchange optimizer based on Openstack | |
| CN110162894A (en) | With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things | |
| CN107809737A (en) | A kind of method and apparatus for the network traffics for determining base station | |
| CN103678000B (en) | Calculating grid schedule equalization tasks method based on reliability and cooperative game | |
| WO2024037560A1 (en) | Multi-user collaboration platform based on edge computing and directed acyclic graph and method using same | |
| CN106936905B (en) | Openstack-based scheduling method and scheduling system of Nova component virtual machine | |
| CN109144690A (en) | task processing method and device | |
| CN111198977B (en) | Graph algorithm load-based graph big data stream type dividing method | |
| CN112486651A (en) | Cloud test platform task scheduling method based on improved genetic algorithm | |
| CN112231117A (en) | Cloud robot service selection method and system based on dynamic vector hybrid genetic algorithm |
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 |