CN101840356A - Multi-core CPU load balancing method based on ring and system thereof - Google Patents
Multi-core CPU load balancing method based on ring and system thereof Download PDFInfo
- Publication number
- CN101840356A CN101840356A CN200910243573A CN200910243573A CN101840356A CN 101840356 A CN101840356 A CN 101840356A CN 200910243573 A CN200910243573 A CN 200910243573A CN 200910243573 A CN200910243573 A CN 200910243573A CN 101840356 A CN101840356 A CN 101840356A
- Authority
- CN
- China
- Prior art keywords
- cpu
- ring
- core cpu
- load
- core
- 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.)
- Granted
Links
Images
Landscapes
- Multi Processors (AREA)
Abstract
The invention relates to a multi-core CPU load balancing method based on ring and a system thereof. The method comprises the following steps: detecting the load balancing conditions of all CPUs in the multi-core CPU in real time and judging whether the multi-core CPU is balanced; evaluating the loading conditions of all rings under the condition that of the multi-core CPU is not balanced; and adjusting binding relationship among the ring and the CPUs by using the ring as a unit. The method uses the ring as a least unit to adjust the CPU load degree, the division of the flow rate of a network card is finer, and further the multi-core CPU load is more balanced. The method and the system thereof of the invention can be widely applied in the multi-core CPU.
Description
Technical field
The present invention relates to multi-core CPU (Central Processing Unit, CPU (central processing unit)), relate in particular to the Network Load Balance of multi-core CPU.
Background technology
Current network is very universal, and the internet high speed development makes that access to netwoks is more and more frequent, therefore causes that the user is also more and more higher to the requirement of the network bandwidth, access to netwoks speed.Network interface card in the network equipment is along with the user develops into PCI-Express to the requirement of access to netwoks speed from ten Broadcoms, hundred Broadcoms.
Along with the raising of each core portfolio of existing network and the quick growth of visit capacity and data traffic, ability of data processing and calculating strength require also to strengthen greatly, make single server equipment to bear.In the case, do a large amount of HardwareUpgrings if throw away existing equipment, this will cause the existing resource huge waste, if face again when portfolio promotes next time, the great number cost input that will cause again HardwareUpgring again, even the remarkable again equipment of performance can not satisfy the demand that the current business amount increases.
A kind of cheapness of deriving out at this situation effectively transparent technology is load balancing, and load balancing is a kind of technology of expanding conventional network equipment and server bandwidth, increase handling capacity, strengthening network data-handling capacity, raising network more flexible and availability.
A plurality of network interface cards and a plurality of CPU are arranged in the multi-core CPU system, generally, the only corresponding CPU of network interface card, this moment, the minimum load balancing unit was a network interface card, CPU corresponding with it when the network interface card flow is very big will be very busy, therefore cause Network Packet Loss probably, and then influence accesses network speed greatly.
Summary of the invention
The invention provides a kind of method and system that can improve the cpu load equilibrium.
In first aspect, the invention provides a kind of multi-core CPU load-balancing method based on ring, this method detects in this multi-core CPU each cpu load equilibrium situation at first in real time and judges whether this multi-core CPU is balanced; Under the unbalanced situation of this multi-core CPU, assess each ring loading condition then; Last is that unit adjusts the binding relationship between ring and the CPU with ring.
In second aspect, the invention provides a kind of multi-core CPU SiteServer LBS based on ring, this system comprises balance module, is used for detecting in real time each cpu load equilibrium situation of this multi-core CPU and judges whether this multi-core CPU is balanced; Under the unbalanced situation of this multi-core CPU, assess each ring loading condition; And be that unit adjusts the binding relationship between ring and the CPU with ring.
In one embodiment of the invention, detect the quantity that each cpu load equilibrium situation in the multi-core CPU comprises the actual occupancy of this multi-core CPU of real-time detection, CPU operation number of processes, CPU handle packet quantity, CPU handle packet size, cpu load network interface card ring in real time.
In another embodiment of the present invention, assess each ring loading condition and comprise that each ring in the real-time detection multi-core CPU is by ring data packet number, linking number by ring data packet length, ring.
The present invention is by judging whether cpu load is balanced, detects each ring loading condition, and adjusts ring and CPU binding relationship, solved the multi-core CPU problem of load balancing.The inventive method is that minimum balancing unit is adjusted the cpu load degree with ring, and the network interface card flow is divided thinner, and the multi-core CPU load is balanced more.
Description of drawings
Below with reference to accompanying drawings specific embodiments of the present invention is described in detail, in the accompanying drawings:
Fig. 1 is the multi-core CPU SiteServer LBS block diagram based on ring of one embodiment of the invention;
Fig. 2 is the multi-core CPU load balancing process flow diagram based on ring of one embodiment of the invention.
Embodiment
Fig. 1 is the multi-core CPU SiteServer LBS block diagram based on ring of one embodiment of the invention.This system comprises a CPU111, the 2nd CPU112, the 3rd CPU113, balance module 120 and first network interface card 130, second network interface card 140.Wherein, first network interface card 130 comprises a ring131, the 2nd ring132, the 3rd ring133; Second network interface card 140 comprises the 4th ring141, the 5th ring142, the 6th ring143, the 7th ring144.Among Fig. 1, a CPU111 and the 2nd ring132, the 4th ring141 binding, the 2nd CPU112 and a ring131, the 6th ring143 binding, the 3rd CPU113 and the 3rd ring133, the 5th ring142, the 7th ring144 binding.
For convenience of description, it is three that Fig. 1 only schematically describes out CPU quantity, and network interface card quantity is two, ring quantity is seven, and binding relationship between CPU and the ring, in fact, the quantity of CPU, network interface card, ring is not all limit, and binding relationship is unfixing yet between CPU and the ring.
The one CPU111, the 2nd CPU112, the 3rd CPU113 are used for handling from the data among first network interface card 130 or second network interface card 140 certain or some ring.
First network interface card 130, second network interface card 140 comprise a plurality of ring respectively, and each network interface card ring quantity all can be configured according to system requirements, this first network interface card 130, second network interface card 140 are used for its place system is linked into network, and each ring in this first network interface card 130, second network interface card 140 is used for reception and sends packet from output packet and to network.
Balance module 120 detects the balanced situation of a CPU111, the 2nd CPU112, the 3rd CPU113 in real time, weigh each ring loading condition in first network interface card 130, second network interface card 140 simultaneously, according to the balanced situation of described each CPU and described each ring loading condition, adjust the binding relationship between CPU and the ring, and then make the multi-core CPU load balancing.Elaborate balance module 120 below and how to judge the CPU equilibrium situation, how to weigh the ring loading condition, and how to adjust the binding relationship between CPU and the ring.
Suppose that balance module 120 detects the 2nd ring132 and the binding of the 4th ring141 in second network interface card 140 in current time the one CPU111 and first network interface card 130; Ring131 in the 2nd CPU112 and first network interface card 130, the 6th ring143 in second network interface card 140 binding; The 5th ring142 in the 3rd ring133 in the 3rd CPU123 and first network interface card, second network interface card 140, the 7th ring144 in second network interface card 140 binding, as shown in Figure 1.
The technical indicators such as quantity of the actual occupancy of each CPU of balance module 120 real-time inspections (CPU111, the 2nd CPU112, the 3rd CPU113), operation number of processes, handle packet quantity, handle packet size, load network interface card ring, calculate CPU (CPU111, the 2nd CPU112, the 3rd CPU113) the shared number percent of each technical indicator, and give each technical indicator certain weight respectively, and then obtain each CPU (CPU111, CPU112, CPU113) equilibrium value.
Particularly, suppose that it is 10% that balance module 120 detects the actual occupancy of current time the one CPU111 in real time, the one CPU111 operation number of processes accounts for multi-core CPU and always moves number of processes 20%, the one CPU111 handle packet quantity accounts for the total handle packet quantity 40% of multi-core CPU, the one CPU111 handle packet size accounts for the total handle packet size 20% of multi-core CPU, the one CPU111 load network interface card ring quantity accounts for multi-core CPU total load network interface card ring quantity 60%, and the shared weight of loading level is a, the shared weight of length of run number of processes is b, the shared weight of handle packet quantity is c, the shared weight of handle packet size is d, the shared weight of load network interface card ring quantity is e, then a CPU111 equilibrium value x1 is
x1＝a*10％+b*20％+c*40％+d*20％+e*60％ (1)
Suppose that the actual occupancy of the 2nd CPU112 this moment is 40%, the 2nd CPU112 operation number of processes accounts for multi-core CPU and always moves number of processes 30%, the 2nd CPU112 handle packet quantity accounts for the total handle packet quantity 30% of multi-core CPU, the 2nd CPU112 handle packet size accounts for the total handle packet size 60% of multi-core CPU, the 2nd CPU112 load network interface card ring quantity accounts for multi-core CPU total load network interface card ring quantity 40%, then the 2nd CPU112 equilibrium value x2 is
x2＝a*40％+b*30％+c*30％+d*60％+e*40％ (2)
Then this moment, the 3rd CPU113 equilibrium value x3 was,
X3＝a*50％+b*50％+c*30％+d*20％+e*0％ (3)
Get a=100, b=10, c=10, d=10, e=20, x1=30 then, x2=60, x3=60.
Balance module 120 is according to a CPU111 equilibrium value x1, the 2nd CPU112 equilibrium value x2, the 3rd CPU113 equilibrium value x3, obtain the balanced mean value of multi-core CPU, and obtain difference y1, y2, y3 between this CPU111, CPU112, CPU113 equilibrium value and this mean value respectively according to this mean value, and this difference y1, y2, y3 account for its corresponding equilibrium value x1, x2, x3 number percent u1%, u2%, u3% respectively.Particularly, the y1 value is,
y1＝x1-(x1+x2+x3)/3 (4)
The y2 value is,
y2＝x2-(x1+x2+x3)/3 (5)
The y3 value is,
y3＝x3-(x1+x2+x3)/3 (6)
The u1 value is,
u1＝(y1/x1)100％ (7)
The u2 value is,
u2＝(y2/x2)100％ (8)
The u3 value is,
u3＝(y3/x3)100％ (9)
With formula (1), (2), (3) substitution formula (4), (5), (6), obtain y1=-20, y2=10, y3=10; With y1, y2, y3 substitution formula (7), (8), (9), obtain u1=-66.6%, u2=16.7%, u3=16.7% again.
Suppose that the balanced threshold value n of the CPU that is set by balance module 120 is 50, by balanced threshold value m is 40, because a CPU112 equilibrium value x1 (x1=30) is less than this equilibrium threshold value n (n=50), then a CPU112 does not need to do equilibrium, and the 2nd CPU112, the 3rd CPU113 equilibrium value x2, x3 (x2=60, x3=60) greater than balanced threshold value n (n=50), then the 2nd CPU112, the 3rd CPU113 need do equilibrium.
Simultaneously since a CPU112 equilibrium value x1 (x1=30) less than by balanced threshold value m (m=40), then a CPU112 can be by equilibrium, and the 2nd CPU112, the 3rd CPU113 equilibrium value x2, x3 (x2=60, x3=60) greater than by balanced threshold value m (m=40), then the 2nd CPU112, the 3rd CPU113 cannot be by equilibriums.Hence one can see that, the one CPU111 can be by equilibrium, the 2nd CPU112, the 3rd CPU113 need equilibrium, therefore the 2nd CPU112, the 3rd CPU113 are balanced among the CPU111, concrete equilibrium amount need be finished by assessment ring loading condition, below will elaborate and how assess the ring loading condition.
Balance module 120 detects each ring in real time by this ring data packet number, data packet length, linking number technical indicators such as (UDP, TCP linking numbers), calculate each technical indicator of each ring with its binding CPU in shared number percent, and give each technical indicator certain weight respectively, thereby obtain the load of each ring.
Particularly, suppose that balance module 120 real-time monitors current time the one ring131 and accounts for and its binding CPU by its data packet number, (being CPU112) is by packet sum 20%, the one ring131 accounts for and its binding CPU by packet bag length, (being CPU112) is by packet overall budget long 30%, the one ring131 linking number accounts for and its binding CPU, (being CPU112) total linking number 70%, and described data packet number weight is f, the long weight of packet bag is g, the linking number weight is h, then a ring131 load z1 is
z1＝f*20％+g*30％+h*70％ (10)
In like manner can get the 2nd ring132 load z2, the 3rd ring133 load z3, the 4th ring141 load z4, the 5th ring142 load z5, the 6th ring143 load z6, the 7th ring144 load z7.
According to the ring load can this ring load on the total ring load of its binding CPU in number percent v.Among Fig. 1, the one ring131, the 6th ring143 and the 2nd CPU112 binding, suppose that a ring131 load z1 value is 10, the 6th ring143 load z6 value is 40, then a ring131 load on the total ring load of its binding CPU112 in number percent v1 equal 20% (obtaining) by 10/ (10+40), then the 6th ring143 load on the total ring load of its binding CPU112 in number percent v6 equal 80%.In like manner can get, other ring load on the total ring load of its binding CPU in number percent.
The number percent v2, the v4 that load in the total ring load of a CPU111 at the 2nd ring132, the 4th ring141 of this hypothesis and CPU111 binding are respectively 40%, 60%, and number percent v3, v5, v7 that hypothesis and the 3rd CPU113 the 3rd ring133, the 5th ring142, the 7th ring144 that bind load in the total ring load of the 3rd CPU113 be respectively 10%, 20%, 70%, such as among Fig. 1 mark.
Balance module 120 respectively with each ring load on the total ring load of its binding CPU in number percent v, deduct that difference accounts for this CPU equilibrium value number percent u between the balanced mean value of this CPU equilibrium value and multi-core CPU, try to achieve v again and subtract u minimal difference absolute value, at last binding relationship between corresponding ring of this minimal difference absolute value and the CPU is adjusted into by among the balanced CPU.
Particularly, the number percent v1 that the one ring131 loads in the total ring load of CPU112 equals 20%, difference accounts for this CPU112 equilibrium value number percent u2 and equals 16.7% between the balanced mean value of CPU112 equilibrium value and multi-core CPU, and the absolute value that deducts behind this u2 with this v1 equals 3.3%; And absolute difference equals 63.3% between the u2 (16.7%) of the v6 of the 6th ring143 (80%) and the 2nd CPU112; Because 3.3% less than 63.3%, therefore the binding relationship between a ring131 and the 2nd CPU112 is adjusted into a ring131 is tied on the CPU111.In like manner binding relationship between the 3rd CPU113 and the 5th ring142 is adjusted into the 5th ring143 is tied on the CPU111, as shown in Figure 1, wherein dotted line is adjusted binding relationship.
Fig. 2 is the multi-core CPU load balancing process flow diagram based on ring of one embodiment of the invention.
In step 210, detect in the multi-core CPU the balanced situation of each CPU in real time and judge whether this multi-core CPU is balanced.
In one embodiment of the invention, balance module detects the technical indicators such as quantity of the actual occupancy of each CPU, operation number of processes, handle packet quantity, handle packet size, load network interface card ring in real time, calculate the shared number percent of each technical indicator of CPU, and give each technical indicator certain weight respectively, thereby obtain each CPU equilibrium value; Judge the balanced threshold value of each CPU equilibrium value and default then and by magnitude relationship between the balanced threshold value, when existing certain or some CPU equilibrium values greater than balanced threshold value and exist certain or some CPU equilibrium values, illustrate that this multi-core CPU is unbalanced less than under by balanced threshold value situation.Judging under the unbalanced situation of this multi-core CPU, trying to achieve the mean value of all CPU equilibrium values in this multi-core CPU, and equilibrium value greater than the CPU of balanced mean value for needing balanced CPU, equilibrium value less than the CPU of balanced mean value for needing by balanced CPU; Try to achieve the difference between the balanced mean value of each CPU equilibrium value and this multi-core CPU again, calculate the number percent of this difference and the corresponding CPU equilibrium value of this difference at last.
In step 220, under the unbalanced situation of multi-core CPU, assess each ring loading condition.
In one embodiment of the invention, balance module detects each ring in real time by technical indicators such as this ring data packet number, data packet length, linking numbers, calculate the every technical indicator of ring with its binding CPU in shared number percent, and give each technical indicator certain weight respectively, thereby obtain the load of each ring, and then ask this ring load on the total ring load of its binding CPU in number percent, thereby obtain each ring loading condition in this multi-core CPU.
In step 230, be that unit adjusts binding relationship between ring and the CPU with ring, so that this multi-core CPU load balancing.
In all CPU of needs equilibrium, respectively with each ring load on the total ring load of its binding CPU in number percent, deduct that difference accounts for this CPU equilibrium value number percent between this CPU equilibrium value and the balanced mean value, try to achieve this difference least absolute value again, at last the binding relationship between corresponding ring of this least absolute value and the CPU is adjusted to by among the balanced CPU.
The above algorithm that detects CPU equilibrium situation, assessment ring loading condition has multiple; need to prove, so long as judge that the unbalanced back of CPU method that to be least unit with ring adjust binding relationship between CPU and the ring and relevant device are just all within protection domain of the present invention.
Obviously, under the prerequisite that does not depart from true spirit of the present invention and scope, the present invention described here can have many variations.Therefore, the change that all it will be apparent to those skilled in the art that all should be included within the scope that these claims contain.The present invention's scope required for protection is only limited by described claims.
Claims (11)
1. multi-core CPU load-balancing method based on ring comprises:
Step a detects in this multi-core CPU each cpu load equilibrium situation in real time and judges whether this multi-core CPU is balanced;
Step b under the unbalanced situation of this multi-core CPU, assesses each ring loading condition;
Step c is that unit adjusts the binding relationship between ring and the CPU with ring.
2. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 1 is characterized in that, detects in real time among the described step a that each cpu load equilibrium situation comprises steps d in this multi-core CPU;
Steps d detects each CPU technical indicator in this multi-core CPU in real time.
3. a kind of multi-core CPU load-balancing method as claimed in claim 2 based on ring, it is characterized in that described CPU technical indicator comprises one or more in the quantity of the actual occupancy of CPU, CPU operation number of processes, CPU handle packet quantity, CPU handle packet size, cpu load network interface card ring.
4. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 2 is characterized in that, comprises step e after steps d;
Step e according to each CPU technical indicator in the described real-time detected multi-core CPU, calculates the shared number percent of each technical indicator of CPU, and gives each technical indicator certain weight respectively, thereby obtain each CPU equilibrium value.
5. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 4 is characterized in that, judges among the described step a whether equilibrium comprises step f to this multi-core CPU;
Step f judges the balanced threshold value of each CPU equilibrium value and default and by magnitude relationship between the balanced threshold value, when existing the CPU equilibrium value greater than balanced threshold value and exist the CPU equilibrium value less than by balanced threshold value the time, judges that this multi-core CPU is unbalanced.
6. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 4 is characterized in that, comprises step g after step e;
Step g is asked the mean value of all CPU equilibrium values in this multi-core CPU, and then asks the difference between each CPU equilibrium value and the balanced mean value of multi-core CPU, calculates the number percent of this difference in corresponding CPU equilibrium value with this difference again.
7. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 1 is characterized in that, each ring loading condition of assessment comprises step h among the step b;
Step h detects each ring technical indicator in this multi-core CPU in real time.
8. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 7 is characterized in that, described ring technical indicator comprises by ring data packet number, linking number by ring data packet length, ring.
9. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 8 is characterized in that described step;
Step I, calculate the every technical indicator of ring with its binding CPU in shared number percent, and give each technical indicator certain weight respectively, thereby obtain each ring load, and then ask this ring load on the total ring load of its binding CPU in number percent.
10. a kind of multi-core CPU load-balancing method based on ring as claimed in claim 1 is characterized in that described step c is according to described cpu load equilibrium situation and ring loading condition, is that unit adjusts the binding relationship between ring and the CPU with ring.
11. the multi-core CPU SiteServer LBS based on ring comprises balance module;
Described balance module is used for detecting in real time each cpu load equilibrium situation of this multi-core CPU and judges whether this multi-core CPU is balanced; Under the unbalanced situation of this multi-core CPU, assess each ring loading condition; And be that unit adjusts the binding relationship between ring and the CPU with ring.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102435736A CN101840356B (en) | 2009-12-25 | 2009-12-25 | Multi-core CPU load balancing method based on ring and system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102435736A CN101840356B (en) | 2009-12-25 | 2009-12-25 | Multi-core CPU load balancing method based on ring and system thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101840356A true CN101840356A (en) | 2010-09-22 |
| CN101840356B CN101840356B (en) | 2012-11-21 |
Family
ID=42743740
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009102435736A Active CN101840356B (en) | 2009-12-25 | 2009-12-25 | Multi-core CPU load balancing method based on ring and system thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101840356B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986608A (en) * | 2010-12-13 | 2011-03-16 | 武汉大学 | Method for evaluating heterogeneous overlay network load balance degree |
| CN102707996A (en) * | 2012-05-15 | 2012-10-03 | 江苏中科梦兰电子科技有限公司 | Task scheduling method for heterogeneous multi-core processor |
| CN103713957A (en) * | 2014-01-08 | 2014-04-09 | 厦门雅迅网络股份有限公司 | Method for achieving CPU burden dynamic balance of vehicle-mounted terminal through serial port |
| CN103823712A (en) * | 2014-03-17 | 2014-05-28 | 杭州华三通信技术有限公司 | Data flow processing method and device for multi-CPU virtual machine system |
| CN104901898A (en) * | 2015-06-08 | 2015-09-09 | 东软集团股份有限公司 | Load balancing method and device |
| CN105468522A (en) * | 2015-11-20 | 2016-04-06 | 浪潮电子信息产业股份有限公司 | Software pressure testing method based on Linux system CPU load control |
| CN106559499A (en) * | 2016-12-07 | 2017-04-05 | 郑州云海信息技术有限公司 | The distribution method and system of a kind of storage system and its computing resource |
| CN107562531A (en) * | 2016-06-30 | 2018-01-09 | 华为技术有限公司 | A kind of data balancing method and device |
| CN109865292A (en) * | 2019-01-10 | 2019-06-11 | 珠海金山网络游戏科技有限公司 | A kind of game resource construction method and device based on game engine |
| CN110321266A (en) * | 2019-06-05 | 2019-10-11 | 上海易点时空网络有限公司 | CPU multi-core utilization rate optimization processing method and device for single server |
| CN110362402A (en) * | 2019-06-25 | 2019-10-22 | 苏州浪潮智能科技有限公司 | A kind of load-balancing method, device, equipment and readable storage medium storing program for executing |
| CN110377418A (en) * | 2019-06-05 | 2019-10-25 | 上海易点时空网络有限公司 | CPU multicore utilization rate optimized treatment method and device for single server |
| CN111831409A (en) * | 2020-07-01 | 2020-10-27 | Oppo广东移动通信有限公司 | Thread scheduling method and device, storage medium and electronic equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3947017B2 (en) * | 2002-02-18 | 2007-07-18 | 株式会社沖データ | Dial pulse generator |
| CN1258716C (en) * | 2003-11-26 | 2006-06-07 | 中国人民解放军国防科学技术大学 | Double ring method for monitoring partial cache consistency of on-chip multiprocessors |
-
2009
- 2009-12-25 CN CN2009102435736A patent/CN101840356B/en active Active
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986608B (en) * | 2010-12-13 | 2012-08-15 | 武汉大学 | Method for evaluating heterogeneous overlay network load balance degree |
| CN101986608A (en) * | 2010-12-13 | 2011-03-16 | 武汉大学 | Method for evaluating heterogeneous overlay network load balance degree |
| CN102707996A (en) * | 2012-05-15 | 2012-10-03 | 江苏中科梦兰电子科技有限公司 | Task scheduling method for heterogeneous multi-core processor |
| CN103713957A (en) * | 2014-01-08 | 2014-04-09 | 厦门雅迅网络股份有限公司 | Method for achieving CPU burden dynamic balance of vehicle-mounted terminal through serial port |
| CN103713957B (en) * | 2014-01-08 | 2018-12-14 | 厦门雅迅网络股份有限公司 | A kind of method that car-mounted terminal realizes CPU burden dynamic equilibrium using serial ports |
| CN103823712A (en) * | 2014-03-17 | 2014-05-28 | 杭州华三通信技术有限公司 | Data flow processing method and device for multi-CPU virtual machine system |
| CN103823712B (en) * | 2014-03-17 | 2017-05-10 | 杭州华三通信技术有限公司 | Data flow processing method and device for multi-CPU virtual machine system |
| CN104901898B (en) * | 2015-06-08 | 2018-08-03 | 东软集团股份有限公司 | A kind of load-balancing method and device |
| CN104901898A (en) * | 2015-06-08 | 2015-09-09 | 东软集团股份有限公司 | Load balancing method and device |
| CN105468522A (en) * | 2015-11-20 | 2016-04-06 | 浪潮电子信息产业股份有限公司 | Software pressure testing method based on Linux system CPU load control |
| CN107562531A (en) * | 2016-06-30 | 2018-01-09 | 华为技术有限公司 | A kind of data balancing method and device |
| CN107562531B (en) * | 2016-06-30 | 2020-10-09 | 华为技术有限公司 | Data equalization method and device |
| CN106559499A (en) * | 2016-12-07 | 2017-04-05 | 郑州云海信息技术有限公司 | The distribution method and system of a kind of storage system and its computing resource |
| CN106559499B (en) * | 2016-12-07 | 2020-02-07 | 苏州浪潮智能科技有限公司 | Storage system and method and system for distributing computing resources thereof |
| CN109865292A (en) * | 2019-01-10 | 2019-06-11 | 珠海金山网络游戏科技有限公司 | A kind of game resource construction method and device based on game engine |
| CN109865292B (en) * | 2019-01-10 | 2023-03-14 | 珠海金山数字网络科技有限公司 | Game resource construction method and device based on game engine |
| CN110377418A (en) * | 2019-06-05 | 2019-10-25 | 上海易点时空网络有限公司 | CPU multicore utilization rate optimized treatment method and device for single server |
| CN110321266A (en) * | 2019-06-05 | 2019-10-11 | 上海易点时空网络有限公司 | CPU multi-core utilization rate optimization processing method and device for single server |
| CN110377418B (en) * | 2019-06-05 | 2021-09-03 | 上海易点时空网络有限公司 | CPU multi-core utilization rate optimization processing method and device for single server |
| CN110321266B (en) * | 2019-06-05 | 2021-11-09 | 上海易点时空网络有限公司 | CPU multi-core utilization rate optimization processing method and device for single server |
| CN110362402A (en) * | 2019-06-25 | 2019-10-22 | 苏州浪潮智能科技有限公司 | A kind of load-balancing method, device, equipment and readable storage medium storing program for executing |
| CN110362402B (en) * | 2019-06-25 | 2021-08-10 | 苏州浪潮智能科技有限公司 | Load balancing method, device, equipment and readable storage medium |
| CN111831409A (en) * | 2020-07-01 | 2020-10-27 | Oppo广东移动通信有限公司 | Thread scheduling method and device, storage medium and electronic equipment |
| CN111831409B (en) * | 2020-07-01 | 2022-07-15 | Oppo广东移动通信有限公司 | Thread scheduling method and device, storage medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101840356B (en) | 2012-11-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101840356B (en) | Multi-core CPU load balancing method based on ring and system thereof | |
| CN104901898B (en) | A kind of load-balancing method and device | |
| WO2012057956A3 (en) | Reactive load balancing for distributed systems | |
| CN101783771A (en) | Method and equipment for realizing load balance continuity | |
| CN105260913A (en) | CTR estimation method and system, and DSP server used for Internet advertisement putting | |
| CN109167812A (en) | Evaluation services quality, the method for determining adjustable strategies, server and storage medium | |
| CN108112038A (en) | A kind of method and device for controlling flowing of access | |
| CN105871588A (en) | Load balance configuration method, device and system | |
| CN109324897A (en) | Data uploading method and system, terminal and computer readable storage medium | |
| CN109191135A (en) | Transaction based on block chain retries method, apparatus, equipment and readable storage medium storing program for executing | |
| US11341842B2 (en) | Metering data management system and computer readable recording medium | |
| CN110083775B (en) | Configuration method and configuration device for recommended resources | |
| CN109063024B (en) | Social platform user influence calculation method and device | |
| CN111385328B (en) | Service request processing method, system and electronic equipment | |
| CN108985556B (en) | Method, apparatus, equipment and the computer storage medium of flow scheduling | |
| US9501321B1 (en) | Weighted service requests throttling | |
| CN108111567A (en) | Realize the uniform method and system of server load | |
| CN112468573A (en) | Data pushing method, device, equipment and storage medium | |
| CN109995818A (en) | A kind of method and device of server load balancing | |
| CN110297948B (en) | Relational network construction method and device | |
| CN103078806B (en) | A kind of load balance scheduling algorithm based on Q value method | |
| CN106612296A (en) | A method and apparatus for assigning user equipment connection requests | |
| CN113657817B (en) | Transaction processing method and device, electronic equipment and readable storage medium | |
| CN105721342B (en) | The method for connecting network and system of multi-process equipment | |
| CN102567416A (en) | Load-balanced distribution type data statistical method and 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |