CN106385381B - A kind of the scheduling of resource distribution method and its system of matching primitives - Google Patents
A kind of the scheduling of resource distribution method and its system of matching primitives Download PDFInfo
- Publication number
- CN106385381B CN106385381B CN201610710107.4A CN201610710107A CN106385381B CN 106385381 B CN106385381 B CN 106385381B CN 201610710107 A CN201610710107 A CN 201610710107A CN 106385381 B CN106385381 B CN 106385381B
- Authority
- CN
- China
- Prior art keywords
- node
- data
- engine
- checked
- detecting
- 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
-
- 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
-
- 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
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/12—Applying verification of the received information
- H04L63/123—Applying verification of the received information received data contents, e.g. message integrity
-
- 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
- H04L67/14—Session management
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention discloses a kind of scheduling of resource distribution method of matching primitives, is ranked up including the computational load at interval of each engine node of △ t time real-time monitoring, and to the load weight of each engine node；When there is unloaded or overload node, the data to be tested packet of the most heavy node of present load is scheduled to node most lightly loaded according to a certain percentage as unit of session, and traverse node carries out the adjusting of load balancing.
Description
Technical field
The present invention relates to network invasion monitoring fields, and in particular to a kind of resource regulating method of matching primitives and its is
System.
Background technique
Under high speed network environment, the detecting and alarm of Network Intrusion Detection System (NIDS) is faced with more serious performance
Bottleneck problem is mainly reflected in detection data sample caused by the high speed of network data flow and greatly increases, and attack multiplicity
Matching characteristic mode caused by changing increases significantly.Solve the problems, such as that the performance bottleneck of the NIDS under high speed network environment has become currently
One of the hot spot of information security field research, the prior art mainly includes two kinds to the settling mode of this problem: one is excellent
Change the detection algorithm efficiency of single detecting and alarm, but be limited to the processing speed of single engine node, this method is difficult to comprehensively
Solve the problems, such as the performance bottleneck that NIDS faces.Another kind is the parallel processing technique of multi engine, becomes solve high speed in recent years
The research emphasis of NIDS bottleneck problem.The key of multi engine parallel processing NIDS is the flow load balance of multi engine node, institute
The load-balancing algorithm of design dynamic equalization should shunt as far as possible and minimize the relevance for destroying attack data context.
The load-balancing algorithm of multi engine parallel processing NIDS is broadly divided into state algorithm and dynamic algorithm.Static load is equal
The method of accounting is shunted according to pre-set strategy, and advantage is that the realization of algorithm is simple, does not generate additional operation
Expense, but distribution condition, the effect of load balancing are adjusted in real time since such algorithm is not bound with real network load situation
Fruit is relatively poor.Dynamic Load-balancing Algorithm can dynamically distribute in the process of running according to each node or the loading condition of link
It is sent to the flow of each engine node, maintains the substantial equilibrium of each node flow in real time, therefore, load balancing effect is relative to static state
Load-balancing algorithm can be well very much, but the algorithm will cause the additional operation expense of system.
Therefore load balancing is carried out to system using dynamic load-balancing algorithm mostly in the prior art, in dynamic load
In assigning process, if may destroy the message that original properties of flow belongs to same session is assigned to different processing nodes,
This may not influence the process performance of routing table look-up, but in network security application, this may be fatal: many at present
Attack conceals attack signature, cannot detect its attack based on independent packet check technology, only capture
All messages of session just can detecte out its attack, if SiteServer LBS after message splicing, recombination
Entire session cannot be assigned to a processor node, then will lead to intruding detection system cannot detect specifically to attack, shape
At is failed to report in face of this problem, there is scholar to propose a kind of adaptive load balancing algorithm of session-oriented, the algorithm by IP packet into
Row multi-domain classification, dynamic adjust TCP flow amount, dynamic load leveling, mould can be carried out on the basis of not influencing integrity of sessions
Draft experiment shows that the algorithm has certain load balancing effect, and smaller for the damage degree of integrity of sessions, algorithm
Realization is also relatively easy to, but the proof of the algorithm still deficient in stability.
The larger flow amount of live network tracer is few but its shared flow specific gravity is larger, existing in order to cope with such case
In technology occur based on it is larger stream adjustment safe Diffluence Algorithm, verified by simulated experiment, the algorithm bit stream portfolio effect compared with
Good, stream destructive rate is relatively low, but the complexity of the algorithm is relatively higher.
Summary of the invention
Present invention aim to address the defect of the prior art, provide a kind of matching primitives scheduling of resource distribution method and
Its system, the technical solution adopted is as follows:
A kind of scheduling of resource distribution method of matching primitives characterized by comprising
At interval of the computational load of each engine node of △ t time real-time monitoring, and to the load weight of each engine node into
Row sequence；
When there is unloaded or overload node, by the data to be tested packet of the most heavy node of present load as unit of session
It is scheduled to node most lightly loaded according to a certain percentage, and traverse node carries out the adjusting of load balancing.
Preferably, the method specifically includes following steps:
S1: initialization △ t；
S2: crawl data packet and by allocation of packets to each detecting and alarm node；
The workload situation of each detecting and alarm node is detected after S3: △ t time, and to current time each detecting and alarm section
The loading condition of point to light sequence according to by being ranked up again；
S4: detect whether unloaded or overload node occur；
S5: if then as unit of session by the data to be tested packet of the most heavy engine node of present load by a certain percentage
It is scheduled to node most lightly loaded, and returns to S2, if it is not, data packet dispatching ratio is then calculated, by the meeting in next △ t period
Words are successively distributed to each engine node after sequence in proportion, then return to S3.
It is that main monitoring carrys out detecting and alarm node with the degree of resource consumption of detecting and alarm node and bottom pour ladle rate in the present invention
Loading condition illustrates corresponding detecting and alarm node overload if bottom pour ladle rate is not zero.
Further, the step S3 is specifically included:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiIt is corresponding
Engine node ID insert Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix,
According to niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large is filled in
Data.
Preferably, in the step S5, when there is unloaded node, if unloaded node directly with session if only one
The data to be tested packet of the most heavy engine node of present load is scheduled to the zero load node by a certain percentage for unit, if unloaded
The quantity of node is greater than 1, then randomly selects one of unloaded node as unit of session for engine section that present load is most heavy
The data to be tested packet of point is scheduled to the zero load node by a certain percentage.
Preferably, the data of engine node scheduling to the node most lightly loaded of pack heaviest are long in the step S5
Degree is pl
nodenonBe not empty detecting and alarm interstitial content for queue to be checked, i' be queue to be checked be not wherein,
Empty node lower label, niFor i-th of detecting and alarm data queue size to be checked, unit is byte, nodenumFor
Detecting and alarm number of nodes, umatchFor the detection rates of detecting and alarm node, unit Mbps, △ t is load monitoring adjustment
Gap periods, unit s.
Preferably, if there is not unloaded node, the scheduling ratio of data packet described in step S5 are as follows:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte,
M is the data total flow to be checked that different engine nodes are assigned in the unit time, and unit Mbps, η are by variable null_numi
And nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is for i-th of detecting and alarm node data queue to be checked
Empty number, nullnumIt is empty engine node number for data queue to be checked.
It is another object of the present invention to solve the defect of the prior art, a kind of scheduling of resource distribution system of matching primitives is provided
System, the technical solution adopted is as follows:
A kind of scheduling of resource distribution system of matching primitives, including load sensor, load analyzer, traffic scheduler and
Multiple detecting and alarms, the load sensor are connect with detecting and alarm and load analyzer respectively, the load analyzer also with
Traffic scheduler connection, the load sensor are used to bear at interval of the work of each engine node of t time △, dynamic detection time
Situation is carried, the load analyzer is to the loading condition of current time each detecting and alarm node according to by carrying out to light sequence again
Sequence, the traffic scheduler is used for when there is overload node or unloaded node, as unit of session that present load is most heavy
The data to be tested packet of engine node be scheduled to node most lightly loaded by a certain percentage, and will be in next △ t period
Session be successively distributed in proportion sequence after each engine node.
Preferably, the load sensor is detected so that the degree of resource consumption of detecting and alarm node and bottom pour ladle rate are foundation
The loading condition of engine node illustrates corresponding detecting and alarm node overload if bottom pour ladle rate is not zero.
Preferably, the load analyzer is to the loading condition of current time each detecting and alarm node according to by again to light
Sequence be ranked up and specifically include:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiIt is corresponding
Engine node ID insert Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix,
According to niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large is filled in
Data.
Preferably, when load analyzer detect unloaded node only one then traffic scheduler is directly single with session
The data to be tested packet of the most heavy engine node of present load is scheduled to the zero load node by position by a certain percentage, if detect
The quantity of unloaded node is greater than 1, then traffic scheduler is randomly selected one of unloaded node and will currently be born as unit of session
The data to be tested packet for carrying most heavy engine node is scheduled to the zero load node by a certain percentage.
Preferably, traffic scheduler gives the session in next △ t period to each detecting and alarm by following pro rate
Node:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte,
M is the data total flow to be checked that different engine nodes are assigned in the unit time, and unit Mbps, η are by variable null_numi
And nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is for i-th of detecting and alarm node data queue to be checked
Empty number, nullnumIt is empty engine node number for data queue to be checked.
Compared with prior art, beneficial effects of the present invention:
1, the present invention maintains the integrity of detection data using session as thread, solves with data message as scheduling
The problem of integrality of breaking unit intrusion detection, reduces and fails to report rate of false alarm raising.
2, the present invention to be to occur overloading or unloaded node is opportunity, avoid using △ t as scheduling occasion encounter with what
The problem of for according to △ t value is determined.
3, the present invention is not to occur overloading or unloaded node is scheduling purpose, and realizes load with computational load outside jot
Equilibrium is engineering objective, the load of each engine node is balanced to a certain extent, than the method more section of mean allocation flow
It learns, and additional operation expense is lower.
Detailed description of the invention
Fig. 1 is flow chart when there is unloaded node in embodiment 1；
Fig. 2 is the entire flow figure of the distribution method of embodiment 1；
Fig. 3 is the structural schematic diagram of the distribution system of embodiment 2.
Specific embodiment
The invention will be described in further detail with reference to the accompanying drawings and examples.
Embodiment 1:
As shown in Figure 1 to Figure 2, the scheduling of resource distribution method of a kind of matching primitives, comprising:
At interval of the computational load of each engine node of △ t time real-time monitoring, and to the load weight of each engine node into
Row sequence；
When there is unloaded or overload node, by the data to be tested packet of the most heavy node of present load as unit of session
It is scheduled to node most lightly loaded according to a certain percentage, and traverse node carries out the adjusting of load balancing.
The present embodiment specifically comprises the following steps:
S1: initialization △ t；
S2: crawl data packet and by allocation of packets to each detecting and alarm node；
The workload situation of each detecting and alarm node is detected after S3: △ t time, and to current time each detecting and alarm section
The loading condition of point to light sequence according to by being ranked up again；
S4: detect whether unloaded or overload node occur；
S5: if then as unit of session by the data to be tested packet of the most heavy engine node of present load by a certain percentage
It is scheduled to node most lightly loaded, and returns to S2, if it is not, data packet dispatching ratio is then calculated, by the meeting in next △ t period
Words are successively distributed to each engine node after sequence in proportion, then return to S3.
It is that main monitoring carrys out detecting and alarm node with the degree of resource consumption of detecting and alarm node and bottom pour ladle rate in the present invention
Loading condition illustrates corresponding detecting and alarm node overload if bottom pour ladle rate is not zero.
Further, the step S3 is specifically included:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiIt is corresponding
Engine node ID insert Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix,
According to niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large is filled in
Data.
In the present embodiment, in the step S5, when there is unloaded node, if unloaded node directly with meeting if only one
The data to be tested packet of the most heavy engine node of present load is scheduled to the zero load node for unit by words by a certain percentage, if empty
The quantity for carrying node is greater than 1, then randomly selects one of unloaded node as unit of session for engine that present load is most heavy
The data to be tested packet of node is scheduled to the zero load node by a certain percentage.
In the present embodiment, data of the engine node scheduling of pack heaviest to node most lightly loaded in the step S5
Length is pl
Wherein, wherein nodenonIt is not empty detecting and alarm interstitial content for queue to be checked, i' is that queue to be checked is not empty
Node lower label, niFor i-th of detecting and alarm data queue size to be checked, unit is byte, nodenumFor detecting and alarm section
Point quantity, umatchFor the detection rates of detecting and alarm node, unit isMbps,△ t is the gap periods of load monitoring adjustment, single
Position is s.
In the present embodiment, if there is not unloaded node, the scheduling ratio of data packet described in step S5 are as follows:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte,
M is the data total flow to be checked that different engine nodes are assigned in the unit time, and unit Mbps, η are by variable null_numi
And nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is for i-th of detecting and alarm node data queue to be checked
Empty number, nullnumIt is empty engine node number for data queue to be checked.
Embodiment 2:
As shown in figure 3, a kind of scheduling of resource distribution system of matching primitives, including load sensor, load analyzer, stream
Amount scheduler and multiple detecting and alarms, the load sensor are connect with detecting and alarm and load analyzer respectively, the load
Analyzer is also connect with traffic scheduler, and the load sensor is used at interval of each engine section of t time △, dynamic detection time
The workload situation of point, the load analyzer is to the loading condition of current time each detecting and alarm node according to by again to light
Sequence be ranked up, the traffic scheduler is used to work as unit of session when there is overload node or unloaded node
The data to be tested packet of the engine node of preceding pack heaviest is scheduled to node most lightly loaded by a certain percentage, and by next △
Session in the t period is successively distributed to each engine node after sequence in proportion.
In the present embodiment, the load sensor is examined so that the degree of resource consumption of detecting and alarm node and bottom pour ladle rate are foundation
The loading condition for surveying engine node, illustrates corresponding detecting and alarm node overload if bottom pour ladle rate is not zero.
In the present embodiment, the load analyzer to the loading condition of current time each detecting and alarm node according to by again to
Light sequence, which is ranked up, to be specifically included:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiIt is corresponding
Engine node ID insert Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix,
According to niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large is filled in
Data.
In the present embodiment, when load analyzer detect unloaded node only one then traffic scheduler is directly with session
The data to be tested packet of the most heavy engine node of present load is scheduled to the zero load node by unit by a certain percentage, if detecting
The quantity of unloaded node be greater than 1, then traffic scheduler randomly select one of unloaded node will be current as unit of session
The data to be tested packet of the engine node of pack heaviest is scheduled to the zero load node by a certain percentage.
In the present embodiment, when there is unloaded or overload node, the engine node scheduling of pack heaviest is to most lightly loaded
The data length of node is pl
Wherein, nodenonIt is not empty detecting and alarm interstitial content for queue to be checked, i' is that queue to be checked is not empty section
Point lower label, niFor i-th of detecting and alarm data queue size to be checked, unit is byte, nodenumFor detecting and alarm number of nodes
Amount, umatchFor the detection rates of detecting and alarm node, unit Mbps, △ t is the gap periods of load monitoring adjustment, and unit is
s。
In the present embodiment, traffic scheduler is drawn the session in next △ t period by following pro rate to each detection
Hold up node:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte,
M is the data total flow to be checked that different engine nodes are assigned in the unit time, and unit Mbps, η are by variable null_numi
And nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is for i-th of detecting and alarm node data queue to be checked
Empty number, nullnumIt is empty engine node number for data queue to be checked.
Claims (5)
1. a kind of scheduling of resource distribution method of matching primitives characterized by comprising
It is arranged at interval of the computational load of each engine node of △ t time real-time monitoring, and to the load weight of each engine node
Sequence；
When there is unloaded or overload node, as unit of session by the data to be tested packet of the most heavy node of present load according to
Certain ratio is scheduled to node most lightly loaded, and traverse node carries out the adjusting of load balancing；
Specifically comprise the following steps:
S1: initialization △ t；
S2: crawl data packet and by allocation of packets to each detecting and alarm node；
The workload situation of each detecting and alarm node is detected after S3: △ t time, and to current time each detecting and alarm node
Loading condition to light sequence according to by being ranked up again；
S4: detect whether unloaded or overload node occur；
S5: if then being dispatched the data to be tested packet of the most heavy engine node of present load by a certain percentage as unit of session
To node most lightly loaded, and S2 is returned to, if it is not, then calculating data packet dispatching ratio, the session in next △ t period is pressed
Ratio is successively distributed to each engine node after sequence, then returns to S3；In step s 5, the engine node scheduling of pack heaviest
Data length to node most lightly loaded is pl
Wherein, nodenonIt is not empty detecting and alarm interstitial content for queue to be checked, i' is that queue to be checked is not under empty node
Label, niFor i-th of detecting and alarm data queue size to be checked, unit is byte, nodenumFor detecting and alarm number of nodes,
umatchFor the detection rates of detecting and alarm node, unit Mbps, △ t is the gap periods of load monitoring adjustment, unit s.
2. a kind of scheduling of resource distribution method of matching primitives according to claim 1, which is characterized in that the step S3
It specifically includes:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiCorresponding draws
Hold up node ID filling Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix, according to
niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large fills in data.
3. a kind of scheduling of resource distribution method of matching primitives according to claim 1, which is characterized in that in the step
S5, if there is not unloaded node, the scheduling ratio of data packet described in step S5 are as follows:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte, and M is
The data total flow to be checked of different engine nodes is assigned in unit time, unit Mbps, η are by variable null_numiWith
nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is empty for i-th of detecting and alarm node data queue to be checked
Number, nullnumIt is empty engine node number for data queue to be checked.
4. a kind of scheduling of resource distribution system of matching primitives, which is characterized in that including load sensor, load analyzer, stream
Amount scheduler and multiple detecting and alarms, the load sensor are connect with detecting and alarm and load analyzer respectively, the load
Analyzer is also connect with traffic scheduler, and the load sensor is used at interval of each engine section of t time △, dynamic detection time
The workload situation of point, the load analyzer is to the loading condition of current time each detecting and alarm node according to by again to light
Sequence be ranked up, the traffic scheduler is used to work as unit of session when there is overload node or unloaded node
The data to be tested packet of the engine node of preceding pack heaviest is scheduled to node most lightly loaded by a certain percentage, and by next △
Session in the t period is successively distributed to each engine node after sequence in proportion；
Wherein the load analyzer to the loading condition of current time each detecting and alarm node according to by again to light sequence into
Row sequence specifically includes:
S31: by each engine node length of data queue n to be checkediIt is ranked up from big to small, successively fills Table2[i]；
S32: by each engine node length of data queue n to be checkediIt is ranked up from big to small, by the n after sequenceiCorresponding draws
Hold up node ID filling Table1[i]；
S33: by each engine node length of data queue n to be checkediIt is ranked up from small to large, successively fills Table3[i]；
Wherein, Table1For 1 × nodenumMatrix, record engine nodal scheme, Table2For 1 × nodenumMatrix, according to
niSequence from big to small fills in data, Table3For 1 × nodenumMatrix, according to niSequence from small to large fills in data；
When occurring overloading or when unloaded node, the data length of engine node scheduling to the node most lightly loaded of pack heaviest is
pl
Wherein, nodenonIt is not empty detecting and alarm interstitial content for queue to be checked, i' is that queue to be checked is not under empty node
Label, niFor i-th of detecting and alarm data queue size to be checked, unit is byte, nodenumFor detecting and alarm number of nodes,
umatchFor the detection rates of detecting and alarm node, unit Mbps, △ t is the gap periods of load monitoring adjustment, unit s.
5. a kind of scheduling of resource distribution system of matching primitives according to claim 4, which is characterized in that traffic scheduler
The session in next △ t period is given to each detecting and alarm node by following pro rate:
Wherein, pliFor in next △ t period to Table1The length of data queue of [i] node scheduling, unit are byte, and M is
The data total flow to be checked of different engine nodes is assigned in unit time, unit Mbps, η are by variable null_numiWith
nullnumDetermine the ratio of the data packet to be checked of adjustment, null_numiIt is empty for i-th of detecting and alarm node data queue to be checked
Number, nullnumIt is empty engine node number for data queue to be checked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610710107.4A CN106385381B (en) | 2016-08-23 | 2016-08-23 | A kind of the scheduling of resource distribution method and its system of matching primitives |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610710107.4A CN106385381B (en) | 2016-08-23 | 2016-08-23 | A kind of the scheduling of resource distribution method and its system of matching primitives |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106385381A CN106385381A (en) | 2017-02-08 |
| CN106385381B true CN106385381B (en) | 2019-05-10 |
Family
ID=57916890
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610710107.4A Active CN106385381B (en) | 2016-08-23 | 2016-08-23 | A kind of the scheduling of resource distribution method and its system of matching primitives |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106385381B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108910508B (en) * | 2018-05-08 | 2020-12-04 | 清华大学 | Multi-machine cooperation rule and device in high-speed sorting task |
| CN109947837A (en) * | 2019-03-21 | 2019-06-28 | 上海富数科技有限公司 | Realize the system and method for intelligent scheduling data grabber function |
| CN111400008B (en) * | 2020-03-13 | 2023-06-02 | 北京旷视科技有限公司 | Computing resource scheduling method and device and electronic equipment |
| CN111767189A (en) * | 2020-06-08 | 2020-10-13 | 北京思特奇信息技术股份有限公司 | Cluster state monitoring method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102281329A (en) * | 2011-08-02 | 2011-12-14 | 北京邮电大学 | Resource scheduling method and system for platform as a service (Paas) cloud platform |
| CN104657221A (en) * | 2015-03-12 | 2015-05-27 | 广东石油化工学院 | Multi-queue peak-alternation scheduling model and multi-queue peak-alteration scheduling method based on task classification in cloud computing |
| CN104901881A (en) * | 2015-04-30 | 2015-09-09 | 华为技术有限公司 | Route device, load balancing device, and message distribution method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100157800A1 (en) * | 2008-12-19 | 2010-06-24 | Inventec Corporation | Method for processing network traffic loading balance |
-
2016
- 2016-08-23 CN CN201610710107.4A patent/CN106385381B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102281329A (en) * | 2011-08-02 | 2011-12-14 | 北京邮电大学 | Resource scheduling method and system for platform as a service (Paas) cloud platform |
| CN104657221A (en) * | 2015-03-12 | 2015-05-27 | 广东石油化工学院 | Multi-queue peak-alternation scheduling model and multi-queue peak-alteration scheduling method based on task classification in cloud computing |
| CN104901881A (en) * | 2015-04-30 | 2015-09-09 | 华为技术有限公司 | Route device, load balancing device, and message distribution method and system |
Non-Patent Citations (1)
| Title |
|---|
| 支持动态调度和负载均衡的云资源管理平台D-Cloud的设计与实;刘坚;《中国优秀硕士学位论文全文数据库》;20160229;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106385381A (en) | 2017-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106385381B (en) | A kind of the scheduling of resource distribution method and its system of matching primitives | |
| US10523692B2 (en) | Load balancing method and apparatus in intrusion detection system | |
| CN105227481B (en) | SDN congestion control method for routing based on path cost and stream scheduling cost minimization | |
| CN103139093B (en) | Based on the express network data stream load equalization scheduling method of FPGA | |
| US10079740B2 (en) | Packet capture engine for commodity network interface cards in high-speed networks | |
| CN101729573B (en) | Dynamic load balancing method of network intrusion detection | |
| CN108712464A (en) | A kind of implementation method towards cluster micro services High Availabitity | |
| EP2730067B1 (en) | Method and apparatus for load balancing | |
| CN102904729B (en) | The intelligent acceleration network card of more applications is supported according to agreement, port shunt | |
| CN107077340A (en) | Load balancing | |
| CN102314336B (en) | A kind of data processing method and system | |
| CN104579996A (en) | Cluster load balancing method and system | |
| CN105917632A (en) | A method for scalable distributed network traffic analytics in telco | |
| CN103428008A (en) | Big data distribution strategy oriented to multiple user groups | |
| CN110493043B (en) | Distributed situation awareness calling method and device | |
| CN108200156A (en) | The dynamic load balancing method of distributed file system under a kind of cloud environment | |
| CN106657191A (en) | Load balancing method and associated apparatus and system | |
| CN113489674A (en) | Malicious traffic intelligent detection method and application for Internet of things system | |
| CN103873523A (en) | Client cluster access method and device | |
| Wu et al. | A Distributed Intrusion Detection Model via Nondestructive Partitioning and Balanced Allocation for Big Data. | |
| CN106131227A (en) | Balancing method of loads, meta data server system and load balance system | |
| CN107102897A (en) | A kind of database active defense method of many GPU parallel processings | |
| CN115277574A (en) | Data center network load balancing method under SDN architecture | |
| CN102404211A (en) | Method and device for realizing load balancing of processors under AMP framework | |
| CN107453948A (en) | The storage method and system of a kind of network measurement data |
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 |