CN115002105A - Balanced distribution method and device based on disk rate and network rate - Google Patents
Balanced distribution method and device based on disk rate and network rate Download PDFInfo
- Publication number
- CN115002105A CN115002105A CN202210414212.9A CN202210414212A CN115002105A CN 115002105 A CN115002105 A CN 115002105A CN 202210414212 A CN202210414212 A CN 202210414212A CN 115002105 A CN115002105 A CN 115002105A
- Authority
- CN
- China
- Prior art keywords
- network
- disk
- time
- rate
- real
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000011156 evaluation Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 4
- 230000000737 periodic effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000002904 solvent Substances 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/101—Server selection for load balancing based on network conditions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1023—Server selection for load balancing based on a hash applied to IP addresses or costs
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Debugging And Monitoring (AREA)
Abstract
The embodiment of the invention provides a balanced distribution method based on disk rate and network rate, which comprises the following steps: the method comprises the steps of obtaining the real-time speed of a disk and a network, counting the average speed of the disk and the network based on the real-time speed, evaluating the lowest expected speed of the disk and the network, determining a load balancing parameter based on the average speed of the disk and the network and the lowest expected speed of the disk and the network, and determining an allocation strategy based on the load balancing parameter. The invention can ensure that the distribution is strictly carried out according to the capacity proportion of the nodes, the nodes with larger capacity are distributed with more tasks at the same time, and when the number of the tasks is enough, all the nodes can be finally distributed to the full at the same time point. In addition, the embodiment of the invention provides a balanced distribution device based on the disk rate and the network rate.
Description
Technical Field
The embodiment of the invention relates to the technical field of communication, in particular to a method and a device for balanced distribution based on disk rate and network rate.
Background
This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
The traditional load balancing method is a load balancing scheme based on the CPU utilization rate, and specifically, a node with the lowest CPU utilization rate is selected from a cluster, and a load balancing result is returned to a client according to the node. However, the CPU utilization in this scheme is calculated by continuously acquiring two pieces of CPU performance data, and since the time interval is short, the calculated CPU utilization is an instantaneous value, and the reliability is poor. For example, for an unbalanced scenario, the CPU utilization rate at a certain time may be high, and the CPU utilization rate at another time may be low, in which case the above scheme cannot really know the true condition of the node CPU utilization rate, and the load balancing performance is affected. It can be seen that, in the prior art, almost all the tasks are based on the balancing of CPU loads, and the task of balanced distribution by IO loads of nodes is not considered. However, pure CPU load balancing may cause a situation that a part of nodes are fully allocated and another part of nodes are still idle, which may result in unbalanced task allocation.
Therefore, how to avoid the phenomenon of unbalanced distribution simply based on the balance of CPU loads is a problem to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a method and a device for balanced distribution based on disk rate and network rate, which aim to realize load balance of network and disk intensive tasks in a cluster. The specific scheme is as follows:
in this context, embodiments of the present invention are intended to provide a method and an apparatus for balanced allocation based on disk rate and network rate.
In a first aspect of the embodiments of the present invention, a method for balanced allocation based on a disk rate and a network rate is provided, where the method includes: acquiring the real-time speed of a disk and a network; counting the average speed of the disk and the network based on the real-time speed; evaluating the lowest expected rates of the disk and the network; determining load balancing parameters based on the average rates of the disk and the network and the lowest expected rates of the disk and the network; determining an allocation policy based on the load balancing parameters.
In an embodiment of the present invention, the real-time rates of the disk and the network include: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations.
In another embodiment of the present invention, the acquiring the real-time rates of the disk and the network includes: acquiring the data volume of network reading operation, the starting time of the network reading operation and the ending time of the network reading operation; determining a real-time rate of the network read operation based on the data volume of the network read operation, the start time of the network read operation, and the end time of the network read operation.
In another embodiment of the present invention, the acquiring the real-time rates of the disk and the network further includes: acquiring the data volume of the network write operation, the starting time of the network write operation and the ending time of the network write operation; determining a real-time rate of the network write operation based on the data volume of the network write operation, the start time of the network write operation, and the end time of the network write operation.
In another embodiment of the present invention, the acquiring the real-time rates of the disk and the network further includes: acquiring the data volume of disk reading operation, the starting time of the disk reading operation and the ending time of the disk reading operation; and determining the real-time rate of the disk reading operation based on the data volume of the disk reading operation, the starting time of the disk reading operation and the ending time of the disk reading operation.
In another embodiment of the present invention, the acquiring the real-time rates of the disk and the network further includes: acquiring the data volume of the disk write operation, the start time of the disk write operation and the end time of the disk write operation; and determining the real-time rate of the disk writing operation based on the data volume of the disk writing operation, the starting time of the disk writing operation and the ending time of the disk writing operation.
In another embodiment of the present invention, said counting the average speed of the disk and the network based on the real-time speed comprises: and respectively counting the average rate of the network writing operation, the average rate of the network reading operation, the average rate of the disk writing operation and the average rate of the disk reading operation based on the real-time rate of the network writing operation, the real-time rate of the network reading operation, the real-time rate of the disk reading operation and the real-time rate of the disk writing operation.
In yet another embodiment of the present invention, said evaluating the lowest expected rates of disk and network comprises: evaluating a lowest expected network read rate, evaluating a lowest expected network write rate, evaluating a lowest expected disk read rate, and evaluating a lowest expected disk write rate, either alone or in combination.
In yet another embodiment of the present invention, the determining the load balancing parameter based on the average speed of the disk and the network and the lowest expected speed of the disk and the network comprises: acquiring the number of tasks allowed to be distributed by the nodes when only the average rate of network read operation is considered; acquiring the number of tasks allowed to be allocated by the node when only the average rate of the network write operation is considered; acquiring the number of tasks allowed to be distributed by the nodes when only the average speed of the disk write operation is considered; acquiring the number of tasks allowed to be allocated by the nodes when only the average speed of the disk read operation is considered; and determining a load balancing parameter based on the number of tasks allowed to be allocated by the node when only the average rate of network read operation is considered, the number of tasks allowed to be allocated by the node when only the average rate of network write operation is considered, the number of tasks allowed to be allocated by the node when only the average rate of disk read operation is considered, and the lowest expected rate of the disk and the network.
In a further embodiment of the invention, the load balancing parameter comprises the number of tasks a node is allowed to allocate.
In a further embodiment of the present invention, the obtaining the number of tasks allowed to be allocated by the node only considering the average rate of network read operations includes: and determining the number of tasks allowed to be allocated by the node when only the average rate of the network read operation is considered based on the average rate of the network read operation and the lowest expected network read rate.
In another embodiment of the present invention, the obtaining the number of tasks allowed to be allocated by the node considering only the average rate of network write operations includes: the number of tasks allowed to be allocated by the node considering only the average rate of network write operations is determined based on the average rate of network write operations and the lowest expected network write rate.
In a further embodiment of the present invention, the obtaining the number of tasks allowed to be allocated by the node while considering only the average rate of the disk read operations includes: and determining the number of tasks allowed to be allocated by the node when only the average speed of the disk read operation is considered based on the average speed of the disk read operation and the lowest expected disk read speed.
In yet another embodiment of the present invention, the obtaining the number of tasks that the node is allowed to allocate when only the average rate of disk write operations is considered includes: the number of tasks that a node is allowed to allocate when considering only the average rate of disk write operations is determined based on the average rate of disk write operations and the lowest expected disk write rate.
In yet another embodiment of the present invention, the determining an allocation policy based on the load balancing parameter comprises: and determining the number of tasks distributed to different nodes at the same time based on the load balancing parameters.
In a second aspect of the embodiments of the present invention, there is provided an apparatus for balanced allocation based on disk rate and network rate, the apparatus comprising: the acquisition module is used for acquiring the real-time speed of the disk and the network; the statistical module is used for counting the average speed of the disk and the network based on the real-time speed; the evaluation module is used for evaluating the lowest expected speed of the disk and the network; a load balancing parameter determining module, configured to determine a load balancing parameter based on the average rate of the disk and the network and a lowest expected rate of the disk and the network; and the distribution strategy determining module is used for determining a distribution strategy based on the load balancing parameters.
In an embodiment of the present invention, the real-time rates of the disk and the network include: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations.
In another embodiment of the present invention, the obtaining module includes: the module is used for acquiring the data volume of the network reading operation, the starting time of the network reading operation and the ending time of the network reading operation; means for determining a real-time rate of the network read operation based on the amount of data for the network read operation, a start time of the network read operation, and an end time of the network read operation.
In yet another embodiment of the present invention, the obtaining module further includes: the module is used for acquiring the data volume of the network write operation, the starting time of the network write operation and the ending time of the network write operation; means for determining a real-time rate of the network write operation based on the amount of data for the network write operation, a start time of the network write operation, and an end time of the network write operation.
In yet another embodiment of the present invention, the obtaining module further includes: the module is used for acquiring the data volume of the disk reading operation, the starting time of the disk reading operation and the ending time of the disk reading operation; and determining the real-time rate of the disk read operation based on the data volume of the disk read operation, the start time of the disk read operation and the end time of the disk read operation.
In yet another embodiment of the present invention, the obtaining module further includes: the module is used for acquiring the data volume of the disk write operation, the start time of the disk write operation and the end time of the disk write operation; means for determining a real-time rate of the disk write operation based on the amount of data for the disk write operation, a start time of the disk write operation, and an end time of the disk write operation.
In still another embodiment of the present invention, the statistical module includes: and the module is used for respectively counting the average rate of the network writing operation, the average rate of the network reading operation, the average rate of the disk writing operation and the average rate of the disk reading operation based on the real-time rate of the network writing operation, the real-time rate of the network reading operation, the real-time rate of the disk reading operation and the real-time rate of the disk writing operation.
In yet another embodiment of the present invention, the evaluation module comprises: evaluating a lowest expected network read rate, evaluating a lowest expected network write rate, evaluating a lowest expected disk read rate, and evaluating a lowest expected disk write rate, either alone or in combination.
In yet another embodiment of the present invention, the determining the load balancing parameter module includes: a module for acquiring the number of tasks allowed to be allocated by the node when only the average rate of network read operation is considered; means for obtaining a number of tasks that the node is allowed to allocate when only an average rate of network write operations is considered; a module for acquiring the number of tasks allowed to be allocated by the node when only the average speed of the disk write operation is considered; a module for obtaining the number of tasks allowed to be allocated by the node when only the average speed of the disk read operation is considered; and the module is used for determining the load balancing parameters based on the number of tasks allowed to be allocated by the nodes when only the average rate of network read operation is considered, the number of tasks allowed to be allocated by the nodes when only the average rate of network write operation is considered, the number of tasks allowed to be allocated by the nodes when only the average rate of disk read operation is considered, and the lowest expected rate of the disk and the network.
In yet another embodiment of the invention, the load balancing parameter comprises the number of tasks that the node is allowed to allocate.
In a further embodiment of the present invention, the module for obtaining the number of tasks allowed to be allocated by the node only when the average rate of network read operations is considered includes: means for determining a number of tasks that the node is allowed to allocate while considering only the average rate of network read operations based on the average rate of network read operations and the lowest desired network read rate.
In a further embodiment of the present invention, the module for obtaining the number of tasks that a node is allowed to allocate when only the average rate of network write operations is considered comprises: means for determining a number of tasks that the node is allowed to allocate when considering only the average rate of network write operations based on the average rate of network write operations and the lowest expected network write rate.
In a further embodiment of the present invention, the module for obtaining the number of tasks allowed to be allocated by a node only when the average rate of disk read operations is considered includes: means for determining a number of tasks that the node is allowed to allocate when considering only the average rate of disk read operations based on the average rate of disk read operations and the lowest expected disk read rate.
In yet another embodiment of the present invention, the module for obtaining the number of tasks that a node is allowed to allocate when only the average rate of disk write operations is considered comprises: means for determining a number of tasks that the node is allowed to allocate when considering only the average rate of disk write operations based on the average rate of disk write operations and the lowest expected disk write rate.
In yet another embodiment of the present invention, the determining an allocation policy module comprises: and determining the number of tasks distributed to different nodes at the same time based on the load balancing parameters.
According to the method and the device for the balanced distribution based on the disk rate and the network rate, data transmission is balanced and distributed based on the disk rate and the network rate, so that tasks can be balanced and distributed according to real-time IO (input/output) loads (IO rates), meanwhile, CPU (central processing unit) loads of nodes are considered, the balance of task distribution can be greatly guaranteed, the nodes are ensured not to be overloaded, and the method and the device are suitable for balanced distribution of any disk/network IO intensive tasks.
Drawings
The above and other objects, features and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description, which proceeds with reference to the accompanying drawings. Several embodiments of the invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which:
FIG. 1 is a flow diagram that schematically illustrates a method for implementing disk rate and network rate based balanced allocation, in accordance with an embodiment of the present invention;
FIG. 2 is a schematic flow chart diagram for implementing disk rate and network rate based balanced allocation, according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an apparatus for implementing balanced allocation based on disk rate and network rate according to an embodiment of the present invention.
In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts.
Detailed Description
The principles and spirit of the present invention will be described with reference to several exemplary embodiments. It is understood that these embodiments are given solely for the purpose of enabling those skilled in the art to better understand and to practice the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
As known to those skilled in the art, the embodiment of the invention can be realized as a method and a device for balanced allocation based on disk rate and network rate. Accordingly, the present disclosure may be embodied in the form of: entirely hardware, entirely software (including firmware, resident software, micro-code, etc.), or a combination of hardware and software.
According to the embodiment of the invention, a method and a device for balanced distribution based on disk rate and network rate are provided.
The principles and spirit of the present invention are explained in detail below with reference to several representative embodiments of the invention.
Referring to fig. 1-2, a flowchart of a method for implementing balanced allocation based on disk and network rates and a diagram of a method for implementing balanced allocation based on disk and network rates according to an embodiment of the present invention are schematically shown. The method comprises the following steps:
s100, acquiring the real-time speed of the disk and the network.
By way of example, the real-time rates of the disk and the network include: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations. The acquiring real-time rates of the disk and the network comprises: the method comprises the steps of obtaining the data volume of the network reading operation, the starting time of the network reading operation and the ending time of the network reading operation, and determining the real-time rate of the network reading operation based on the data volume of the network reading operation, the starting time of the network reading operation and the ending time of the network reading operation. Specifically, each time a network read operation occurs, a real-time rate NetReadSpeed of the network read operation is calculated:
NetReadSpeed＝DataBlockSize/(NetReadEndTime-NetReadStartTime)
where DataBlockSize is the data size in bytes Byte. NetReadEndTime is the end time of the read operation, in nanoseconds ns. NetReadStartTime is the start time of the read operation in nanoseconds ns.
Similarly, when the real-time rate of the network write operation, the real-time rate of the disk read operation, and the real-time rate of the disk write operation are obtained, the real-time rates can be obtained according to the above methods, and details are not repeated here.
And S110, counting the average speed of the disk and the network based on the real-time speed.
By way of example, the counting the average rates of the disk and the network based on the real-time rate includes: and respectively counting the average rate of the network writing operation, the average rate of the network reading operation, the average rate of the disk writing operation and the average rate of the disk reading operation based on the real-time rate of the network writing operation, the real-time rate of the network reading operation, the real-time rate of the disk reading operation and the real-time rate of the disk writing operation. Specifically, the average rate of the network read operations is counted as an example, and the detailed description is given by taking the average rate of the network read operations as an example, and the real-time rates netread speed of all the network read operations in the past N (the value of N may be adjusted according to the actual situation, and it is assumed that N times of network reads occur in N seconds) seconds may be weighted and averaged to obtain the average rate avgnnetread speed of the network read operations in this period of time:
AvgNetReadSpeed＝(NetReadSpeed[0]*NetReadWeight[0]+NetReadSpeed[1]*NetReadWeight[1]+...+NetReadSpeed[n-1]*NetReadWeight[n-1])/n
wherein, need to satisfy: NetReadStartTime [ N ] -NetReadStartTime [0] < ═ N × 1000000000; NetReadWeight [ i ]/ElapsedTime [ i ], ElapsedTime [ i ] indicates how long the ith reading operation is away from the current, and the Factor [ i ] can be designed and valued according to the actual situation, which indicates that the longer the reading operation is away from the current time, the smaller the weight value is; wherein NetReadWeight [0] + NetReadWeight [1] +. NetReadWeight [ n-1] ═ 1.
And written AvgSpeed [0] ═ AvgNetReadSpeed.
Similarly, when the average rate of the network write operation, the average rate of the disk read operation, and the average rate of the disk write operation are obtained, they can be obtained according to the above method, and details are not described here.
That is, the average rate of network write operations, AvgNetWriteSpeed, is the same as AvgNetReadSpeed.
Let AvgSpeed [1] ═ avgnetwriteexpected;
the statistical method of the average speed AvgDiskReadSpeed of the disk reading operation is the same as AvgNetReadSpeed.
Remember avgseed [2] ═ AvgNetReadSpeed;
the average speed AvgDiskWriteSpeed of disk write operation is the same as AvgNetReadSpeed.
Let AvgSpeed [3] ═ AvgDiskWriteSpeed.
As an example, the maximum rates of the network and disk may be counted as follows:
(1) maximum network read rate
Maxseed [0] ═ NetReadSpeedMax ═ Max { avgtnetreadspeed [ i ] }; wherein i is 0, 1.
I.e., NetReadSpeedMax is the maximum value of historical avgnettedspeed.
(2) Maximum network write rate
Maxseed [1] ═ netwritespedmax ═ Max { avgtnetwritespespesped [ i ] }; wherein i is 0, 1.
I.e. netWriteSpeedMax is the maximum value of historical AvgNetWriteSpeed.
(3) Maximum disk read rate
Maxseed [2] ═ diskheadspeedmax ═ Max { AvgDiskReadSpeed [ i ] }; wherein i is 0, 1.
I.e., diskheadspeedmax is the maximum value of historical AvgDiskReadSpeed.
(4) Maximum disk write rate
Maxseed [3] ═ distwriteformat Max ═ Max { avgdiskwritecoded [ i ] }; wherein i is 0, 1.
I.e., DiskWritePeedMax is the maximum value of historical AvgDiskWritePeped.
Moreover, in order to prevent the phenomenon that the statistical data cannot be updated due to no data transmission, a periodic detection mechanism is required to be added: if the distance from the last IO exceeds 10s (which can be adjusted according to practical conditions), the AvgSpeed [ i ] value needs to be actively increased, MinDesiredSpeed [ i ] can be used as the increasing step size, but the MaxSpeed [ i ] cannot be exceeded. The reason is that when no IO occurs, the IO available bandwidth can be rapidly increased even the IO is in an idle state, the condition that the IO is idle can be effectively avoided through a periodic detection mechanism, and the condition that the rate is too high in the process of actively increasing the AvgSpeed [ i ] can be avoided by taking MinDesiredSpeed [ i ] as an increased step length and maxSpeed [ i ] as a limit.
And S120, evaluating the lowest expected speed of the disk and the network.
By way of example, the evaluating the lowest expected rates of disk and network includes: evaluating a lowest expected network read rate, evaluating a lowest expected network write rate, evaluating a lowest expected disk read rate, and evaluating a lowest expected disk write rate, either alone or in combination. Specifically, evaluation
(1) Minimum expected network read rate
MinDesiredSpeed[0]＝NetReadMinDesiredSpeed＝10MB/s
The above representation reads a 1GB file from the network, and takes about 1.7 minutes at a read rate of 10MB/s, and if it takes longer, the experience is degraded. It should be noted that the minimum expected network read rate can be adjusted according to actual conditions.
(2) Minimum expected network write rate
MinDesiredSpeed[1]＝NetWriteMinDesiredSpeed＝3MB/s
(3) Minimum expected disk read rate
MinDesiredSpeed[2]＝DiskReadMinDesiredSpeed＝50MB/s
(4) Minimum expected disk write rate
MinDesiredSpeed[3]＝DiskWriteMinDesiredSpeed＝20MB/s
It should be noted that the minimum expected bandwidth requirement is continuously adjusted and determined by the relevant technical personnel according to the actual service requirement and the actual customer experience, so that the determined final value can better fit the actual requirement of the user.
And S130, determining load balancing parameters based on the average speed of the disk and the network and the lowest expected speed of the disk and the network.
As an example, the determining a load balancing parameter based on the average rate of the disk and the network and the lowest expected rate of the disk and the network comprises: the method comprises the steps of obtaining the number of tasks allowed to be distributed by a node when only the average rate of network read operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of network write operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of disk read operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of disk write operation is considered, determining a load balancing parameter based on the number of tasks allowed to be distributed by the node when only the average rate of network read operation is considered, the number of tasks allowed to be distributed by the node when only the average rate of disk write operation is considered, the number of tasks allowed to be distributed by the node when only the average rate of disk read operation is considered, and the lowest expected rate of a disk and a network. In particular, the amount of the solvent to be used,
(1) allowing the number of tasks to be allocated taking into account only the average rate of network read operations
AllowedTaskNum[0]＝AvgSpeed[0]/MinDesiredSpeed[0]；
(2) Allowing the number of tasks to be allocated considering only the average rate of network writes
AllowedTaskNum[1]＝AvgSpeed[1]/MinDesiredSpeed[1]；
(3) Allowing the number of tasks to be allocated taking into account only the average rate of disk read operations
AllowedTaskNum[2]＝AvgSpeed[2]/MinDesiredSpeed[2]；
(4) Allowing the number of tasks to be allocated taking into account only the average rate of disk write operations
AllowedTaskNum[3]＝AvgSpeed[3]/MinDesiredSpeed[3]；
And, as can be seen from the lowest expected rate evaluated above, due to the lowest demand for rate by a single task:
(1) the lowest network read rate MinDesiredSpeed [0 ];
(2) lowest network write rate MinDesiredSpeed [1 ];
(3) the lowest disk read rate MinDesiredSpeed [2 ];
(4) the lowest disk write rate MinDesiredSpeed [3 ];
thus, the minimum requirement to be met if the task is desired to be able to complete the transmission is that the above 4 points need to be met simultaneously. That is, the number of tasks a node is allowed to allocate is
PeportedAllowedTaskNum＝Min{AllowedTaskNum[i]},i＝0,1,2,3
And S140, determining a distribution strategy based on the load balancing parameters.
As an example, the determining an allocation policy based on the load balancing parameter includes: and determining the number of tasks distributed to different nodes at the same time based on the load balancing parameters. Specifically, assume that a cluster has M nodes, and each node currently reports a value of petortedalloweddtask num to the task distributor.
Now, if there is a task to be allocated, the probability of the task being allocated to each node can be calculated as
Probability[i]＝PeportedAllowedTaskNum[i]/(PeportedAllowedTaskNum[0]+...+PeportedAllowedTaskNum[M-1]).
Whether to assign the task to the node i may be determined by a random number function rand.
If it is not
(PeportedAllowedTaskNum[0]+..+PeportedAllowedTaskNum[i-1])<(rand()％(PeportedAllowedTaskNum[0]+...+PeportedAllowedTaskNum[N-1])+1)<＝(PeportedAllowedTaskNum[0]+..+PeportedAllowedTaskNum[i])
It is assigned to the inode, otherwise it is assigned to other nodes.
Therefore, the distribution can be strictly carried out according to the capacity (PeportedAllowedTaskNum value) ratio of the nodes, the nodes with larger capacity are distributed with more tasks at the same moment, and when the number of the tasks is large enough, all the nodes can be finally distributed to be full at the same time point, so that the phenomenon that one part of the nodes are distributed to be full and the other part of the nodes are relatively idle is avoided.
By way of example, the data can be distributed according to the use condition of the CPU. Because even the IO-intensive tasks may use the CPU simultaneously for a large number of tasks, this may cause the CPU to be overloaded. To prevent this, the CPU utilization is checked periodically and if it exceeds 85%, allowedtimeknum will be set to a very small value, e.g., 1, immediately. This may prevent the dispenser from continuing to dispense a large number of tasks. And after the CPU utilization rate is reduced to a certain proportion (for example, 75 percent), restoring AllowedTaskNum.
According to the method and the device, data transmission is distributed in a balanced mode based on the disk speed and the network speed, so that tasks can be distributed in a balanced mode according to real-time IO (input/output) loads, meanwhile, the CPU loads of the nodes are considered, balance of task distribution can be guaranteed greatly, the nodes are prevented from being overloaded, and meanwhile the method and the device are suitable for balanced distribution of any disk/network IO intensive tasks.
Exemplary devices
Having described the method of the exemplary embodiment of the present invention, reference is next made to fig. 3, which is a schematic diagram of an apparatus for implementing balanced allocation based on disk rate and network rate of the exemplary embodiment of the present invention, the apparatus includes the following modules:
the obtaining module 300 is configured to obtain a real-time rate of a disk and a network.
By way of example, the real-time rates of the disk and the network include: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations. The acquiring the real-time rates of the disk and the network comprises the following steps: the method comprises the steps of obtaining the data volume of network reading operation, the starting time of the network reading operation and the ending time of the network reading operation, and determining the real-time rate of the network reading operation based on the data volume of the network reading operation, the starting time of the network reading operation and the ending time of the network reading operation. Specifically, each time a network read operation occurs, the real-time rate NetReadSpeed of the network read operation is calculated as follows:
NetReadSpeed＝DataBlockSize/(NetReadEndTime-NetReadStartTime)
where DataBlockSize is the data size in bytes Byte. NetReadEndTime is the end time of the read operation, and the unit nanosecond is ns. NetReadStartTime is the start time of the read operation in nanoseconds ns.
Similarly, when the real-time rate of the network write operation, the real-time rate of the disk read operation, and the real-time rate of the disk write operation are obtained, the real-time rates can be obtained according to the above methods, and details are not repeated here.
And a statistic module 310, configured to count an average rate of the disk and the network based on the real-time rate.
As an example, the statistics of the average rate of the disk and the network based on the real-time rate includes: and respectively counting the average rate of the network writing operation, the average rate of the network reading operation, the average rate of the disk writing operation and the average rate of the disk reading operation based on the real-time rate of the network writing operation, the real-time rate of the network reading operation, the real-time rate of the disk reading operation and the real-time rate of the disk writing operation. Specifically, the average rate of the network read operations is counted as an example, and the detailed description is given by taking the average rate of the network read operations as an example, and the real-time rates netread speed of all the network read operations in the past N (the value of N may be adjusted according to the actual situation, and it is assumed that N times of network reads occur in N seconds) seconds may be weighted and averaged to obtain the average rate avgnnetread speed of the network read operations in this period of time:
AvgNetReadSpeed＝(NetReadSpeed[0]*NetReadWeight[0]+NetReadSpeed[1]*NetReadWeight[1]+...+NetReadSpeed[n-1]*NetReadWeight[n-1])/n
wherein, need to satisfy: NetReadStartTime [ N ] -NetReadStartTime [0] < ═ N × 1000000000; NetReadWeight [ i ]/ElapsedTime [ i ], ElapsedTime [ i ] indicates how long the ith reading operation is away from the current, and the Factor [ i ] can be designed and valued according to the actual situation, which indicates that the longer the reading operation is away from the current time, the smaller the weight value is; wherein NetReadWeight [0] + NetReadWeight [1] +. NetReadWeight [ n-1] ═ 1.
And remember avgseed [0] ═ AvgNetReadSpeed.
Similarly, when the average rate of the network write operation, the average rate of the disk read operation, and the average rate of the disk write operation are obtained, they can be obtained according to the above method, and details are not described here.
That is, the average rate of network write operations, AvgNetWriteSpeed, is the same as AvgNetReadSpeed.
Let AvgSpeed [1] ═ avgnetwriteexpected;
the statistical method of the average speed AvgDiskReadSpeed of the disk reading operation is the same as AvgNetReadSpeed.
Remember avgseed [2] ═ AvgNetReadSpeed;
the average speed AvgDiskWriteSpeed of disk write operation is the same as AvgNetReadSpeed.
Let AvgSpeed [3] ═ AvgDiskWriteSpeed.
As an example, the maximum rates of the network and disk may be counted as follows:
(1) maximum network read rate
Maxseed [0] ═ NetReadSpeedMax ═ Max { avgtnetreadspeed [ i ] }; wherein i is 0, 1.
I.e., NetReadSpeedMax is the maximum value of historical avgnettedspeed.
(2) Maximum network write rate
Maxseed [1] ═ netwritespedmax ═ Max { avgtnetwritesped [ i ] }; wherein i is 0, 1.
I.e. netWriteSpeedMax is the maximum value of historical AvgNetWriteSpeed.
(3) Maximum disk read rate
Maxseed [2] ═ diskheadspeedmax ═ Max { AvgDiskReadSpeed [ i ] }; wherein i is 0, 1.
I.e., diskheadspeedmax is the maximum value of historical AvgDiskReadSpeed.
(4) Maximum disk write rate
Maxseed [3] ═ distwriteformat Max ═ Max { avgdiskwritecoded [ i ] }; wherein i is 0, 1.
I.e., DiskWritePeedMax is the maximum value of historical AvgDiskWritePeped.
Moreover, in order to prevent the phenomenon that the statistical data cannot be updated due to no data transmission, a periodic detection mechanism is required to be added: if the distance from the last IO exceeds 10s (which can be adjusted according to practical conditions), the AvgSpeed [ i ] value needs to be actively increased, MinDesiredSpeed [ i ] can be used as the increasing step size, but the MaxSpeed [ i ] cannot be exceeded. The reason is that under the condition that no IO occurs, the available bandwidth of the IO can be rapidly increased, even the IO is in an idle state, the condition that the IO is idle can be effectively avoided through a periodic detection mechanism, and the condition that the rate is too high in the process of actively increasing the value of the AvgSpeed [ i ] can also be avoided through taking MinDesiredSpeed [ i ] as an increased step length and taking MaxSpeed [ i ] as a limit.
And an evaluation module 320 for evaluating the lowest expected rates of the disk and the network.
By way of example, the evaluating the lowest expected rates of disk and network includes: evaluating a minimum expected network read rate, evaluating a minimum expected network write rate, evaluating a minimum expected disk read rate, and evaluating a minimum expected disk write rate, either alone or in combination. Specifically, evaluation
(1) Minimum expected network read rate
MinDesiredSpeed[0]＝NetReadMinDesiredSpeed＝10MB/s
The above representation reads a 1GB file from the network, and takes about 1.7 minutes at a read rate of 10MB/s, and if it takes longer, the experience is degraded. It should be noted that the minimum expected network read rate can be adjusted according to actual conditions.
(2) Minimum expected network write rate
MinDesiredSpeed[1]＝NetWriteMinDesiredSpeed＝3MB/s
(3) Minimum expected disk read rate
MinDesiredSpeed[2]＝DiskReadMinDesiredSpeed＝50MB/s
(4) Minimum expected disk write rate
MinDesiredSpeed[3]＝DiskWriteMinDesiredSpeed＝20MB/s
It should be noted that the minimum expected bandwidth needs are continuously adjusted and determined by the related technical personnel according to the actual service requirements and the actual customer experience, so that the determined final value can better fit the actual requirements of the user.
A determine load balancing parameters module 330 for determining load balancing parameters based on the average rate of the disk and the network and the lowest expected rate of the disk and the network.
As an example, the determining a load balancing parameter based on the average rate of the disk and the network and the lowest expected rate of the disk and the network comprises: the method comprises the steps of obtaining the number of tasks allowed to be distributed by a node when only the average rate of network read operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of network write operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of disk read operation is considered, obtaining the number of tasks allowed to be distributed by the node when only the average rate of disk write operation is considered, determining a load balancing parameter based on the number of tasks allowed to be distributed by the node when only the average rate of network read operation is considered, the number of tasks allowed to be distributed by the node when only the average rate of disk write operation is considered, the number of tasks allowed to be distributed by the node when only the average rate of disk read operation is considered, and the lowest expected rate of a disk and a network. In particular, the amount of the solvent to be used,
(1) allowing the number of tasks to be allocated considering only the average rate of network read operations
AllowedTaskNum[0]＝AvgSpeed[0]/MinDesiredSpeed[0]；
(2) Allowing the number of tasks to be allocated taking into account only the average rate of network writes
AllowedTaskNum[1]＝AvgSpeed[1]/MinDesiredSpeed[1]；
(3) Allowing the number of tasks to be allocated taking into account only the average rate of disk read operations
AllowedTaskNum[2]＝AvgSpeed[2]/MinDesiredSpeed[2]；
(4) Allowing the number of tasks to be allocated taking into account only the average rate of disk write operations
AllowedTaskNum[3]＝AvgSpeed[3]/MinDesiredSpeed[3]；
And, from the minimum expected rate evaluated above, due to the minimum rate requirement of a single task:
(1) the lowest network read rate MinDesiredSpeed [0 ];
(2) the lowest network write rate MinDesiredSpeed [1 ];
(3) the lowest disk read rate MinDesiredSpeed [2 ];
(4) the lowest disk write rate MinDesiredSpeed [3 ];
thus, the minimum requirement to be met if the task is desired to be able to complete the transmission is that the above 4 points need to be met simultaneously. That is, the number of tasks a node is allowed to allocate is
PeportedAllowedTaskNum＝Min{AllowedTaskNum[i]},i＝0,1,2,3
A determine allocation policy module 340 for determining an allocation policy based on the load balancing parameters.
As an example, the determining an allocation policy based on the load balancing parameter includes: and determining the number of tasks distributed to different nodes at the same time based on the load balancing parameters. Specifically, assume that a cluster has M nodes, and each node currently reports a value of petortedalloweddtask num to the task distributor.
Now there is a task to be distributed, then the probability of distributing the task to each node can be calculated as
Probability[i]＝PeportedAllowedTaskNum[i]/(PeportedAllowedTaskNum[0]+...+PeportedAllowedTaskNum[M-1]).
Whether to assign a task to node i may be determined by a random number function rand.
If it is not
(PeportedAllowedTaskNum[0]+..+PeportedAllowedTaskNum[i-1])<(rand()％(PeportedAllowedTaskNum[0]+...+PeportedAllowedTaskNum[N-1])+1)<＝(PeportedAllowedTaskNum[0]+..+PeportedAllowedTaskNum[i])
The node is assigned to the i node, otherwise, the node is assigned to other nodes.
Therefore, the distribution can be strictly carried out according to the capacity (PeportedAllowedTaskNum value) ratio of the nodes, the nodes with larger capacity are distributed with more tasks at the same moment, and when the number of the tasks is large enough, all the nodes can be finally distributed to be full at the same time point, so that the phenomenon that one part of the nodes are distributed to be full and the other part of the nodes are relatively idle is avoided.
By way of example, the data can be distributed according to the use condition of the CPU. Because even IO-intensive tasks may have a large number of tasks simultaneously using the CPU, this may result in an excessive CPU load. To prevent this, the CPU utilization is checked periodically and if it exceeds 85%, allowedtimeknum will be set to a very small value, e.g., 1, immediately. This may prevent the dispenser from continuing to dispense a large number of tasks. And after the CPU utilization rate is reduced to a certain proportion (for example, 75 percent), restoring AllowedTaskNum.
It should be noted that although in the above detailed description reference has been made to several units/modules or sub-units/modules of an apparatus for equalized distribution based on disk rate and network rate, such division is merely exemplary and not mandatory. Indeed, the features and functions of two or more of the units/modules described above may be embodied in one unit/module according to embodiments of the invention. Conversely, the features and functions of one unit/module described above may be further divided into embodiments by a plurality of units/modules.
Moreover, while the operations of the method of the invention are depicted in the drawings in a particular order, this does not require or imply that the operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.
While the spirit and principles of the invention have been described with reference to several particular embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, nor is the division of aspects, which is for convenience only as the features in such aspects cannot be combined to advantage. The invention is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.
Claims (10)
1. A balanced distribution method based on disk rate and network rate is characterized by comprising the following steps:
acquiring the real-time speed of a disk and a network;
counting the average speed of the disk and the network based on the real-time speed;
evaluating the lowest expected rates of the disk and the network;
determining a load balancing parameter based on the average speed of the disk and the network and the lowest expected speed of the disk and the network;
determining an allocation policy based on the load balancing parameters.
2. The method of claim 1, wherein the real-time rates of the disk and the network comprise: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations.
3. The method of claim 2, wherein obtaining the real-time rates of the disk and the network comprises:
acquiring the data volume of network reading operation, the starting time of the network reading operation and the ending time of the network reading operation;
determining a real-time rate of the network read operation based on the data volume of the network read operation, the start time of the network read operation, and the end time of the network read operation.
4. The method of claim 2, wherein obtaining the real-time rates of the disk and the network further comprises:
acquiring the data volume of the network write operation, the starting time of the network write operation and the ending time of the network write operation;
determining a real-time rate of the network write operation based on the data volume of the network write operation, the start time of the network write operation, and the end time of the network write operation.
5. The method of claim 2, wherein obtaining the real-time rates of the disk and the network further comprises:
acquiring the data volume of disk reading operation, the starting time of the disk reading operation and the ending time of the disk reading operation;
and determining the real-time rate of the disk read operation based on the data volume of the disk read operation, the start time of the disk read operation and the end time of the disk read operation.
6. An apparatus for balanced allocation based on disk rate and network rate, the apparatus comprising:
the acquisition module is used for acquiring the real-time speed of the disk and the network;
the statistical module is used for counting the average speed of the disk and the network based on the real-time speed;
the evaluation module is used for evaluating the lowest expected speed of the disk and the network;
a load balancing parameter determining module, configured to determine a load balancing parameter based on the average rate of the disk and the network and a lowest expected rate of the disk and the network;
and the distribution strategy determining module is used for determining a distribution strategy based on the load balancing parameters.
7. The apparatus of claim 6, wherein the real-time rates of the disk and network comprise: one or a combination of a real-time rate of network read operations, a real-time rate of network write operations, a real-time rate of disk read operations, and a real-time rate of disk write operations.
8. The apparatus of claim 7, wherein the obtaining module comprises:
the module is used for acquiring the data volume of the network read operation, the starting time of the network read operation and the ending time of the network read operation;
means for determining a real-time rate of the network read operation based on the amount of data for the network read operation, a start time of the network read operation, and an end time of the network read operation.
9. The apparatus of claim 7, wherein the obtaining module further comprises:
the module is used for acquiring the data volume of the network write operation, the starting time of the network write operation and the ending time of the network write operation;
means for determining a real-time rate of the network write operation based on the amount of data for the network write operation, a start time of the network write operation, and an end time of the network write operation.
10. The apparatus of claim 7, wherein the obtaining module further comprises:
the module is used for acquiring the data volume of the disk reading operation, the starting time of the disk reading operation and the ending time of the disk reading operation;
and determining the real-time rate of the disk read operation based on the data volume of the disk read operation, the start time of the disk read operation and the end time of the disk read operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210414212.9A CN115002105B (en) | 2022-04-20 | 2022-04-20 | Balanced distribution method and device based on disk rate and network rate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210414212.9A CN115002105B (en) | 2022-04-20 | 2022-04-20 | Balanced distribution method and device based on disk rate and network rate |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115002105A true CN115002105A (en) | 2022-09-02 |
| CN115002105B CN115002105B (en) | 2024-08-09 |
Family
ID=83026095
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210414212.9A Active CN115002105B (en) | 2022-04-20 | 2022-04-20 | Balanced distribution method and device based on disk rate and network rate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115002105B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103685544A (en) * | 2013-12-24 | 2014-03-26 | 华中科技大学 | Performance pre-evaluation based client cache distributing method and system |
| CN103731498A (en) * | 2013-12-31 | 2014-04-16 | 浙江鸿程计算机系统有限公司 | Big data real-time enquiry system load balancing method based on copy selection |
| CN105068875A (en) * | 2015-08-12 | 2015-11-18 | 浪潮(北京)电子信息产业有限公司 | Intelligence data processing method and apparatus |
| CN105975634A (en) * | 2016-06-22 | 2016-09-28 | 清华大学 | Method for storing multi-dimensional ordered data in distributed data storage system |
| WO2016192456A1 (en) * | 2015-06-04 | 2016-12-08 | 中兴通讯股份有限公司 | Local mounting processing method, apparatus and system for network disk |
| CN107197053A (en) * | 2017-07-31 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of load-balancing method and device |
| CN108121601A (en) * | 2017-11-08 | 2018-06-05 | 上海格蒂电力科技有限公司 | A kind of application resource dispatching device and method based on weight |
| CN111381971A (en) * | 2020-03-17 | 2020-07-07 | 重庆邮电大学 | Nginx-based dynamic weight load balancing method |
-
2022
- 2022-04-20 CN CN202210414212.9A patent/CN115002105B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103685544A (en) * | 2013-12-24 | 2014-03-26 | 华中科技大学 | Performance pre-evaluation based client cache distributing method and system |
| CN103731498A (en) * | 2013-12-31 | 2014-04-16 | 浙江鸿程计算机系统有限公司 | Big data real-time enquiry system load balancing method based on copy selection |
| WO2016192456A1 (en) * | 2015-06-04 | 2016-12-08 | 中兴通讯股份有限公司 | Local mounting processing method, apparatus and system for network disk |
| CN105068875A (en) * | 2015-08-12 | 2015-11-18 | 浪潮(北京)电子信息产业有限公司 | Intelligence data processing method and apparatus |
| CN105975634A (en) * | 2016-06-22 | 2016-09-28 | 清华大学 | Method for storing multi-dimensional ordered data in distributed data storage system |
| CN107197053A (en) * | 2017-07-31 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of load-balancing method and device |
| CN108121601A (en) * | 2017-11-08 | 2018-06-05 | 上海格蒂电力科技有限公司 | A kind of application resource dispatching device and method based on weight |
| CN111381971A (en) * | 2020-03-17 | 2020-07-07 | 重庆邮电大学 | Nginx-based dynamic weight load balancing method |
Non-Patent Citations (3)
| Title |
|---|
| 余彬彬;张伟;王欢;徐威;祝明发;阮利;: "基于最小负载的Linux Web机群负载均衡算法", 华中科技大学学报(自然科学版), no. 1, 15 June 2010 (2010-06-15) * |
| 郭关飞, 张尧学, 韦理: "基于透明计算的龙星网络计算机系统测试报告", 计算机工程与应用, no. 17, 1 May 2006 (2006-05-01) * |
| 高子妍;王勇;: "面向云服务的分布式消息系统负载均衡策略", 计算机科学, no. 1, 15 June 2020 (2020-06-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115002105B (en) | 2024-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11546644B2 (en) | Bandwidth control method and apparatus, and device | |
| CN108683720B (en) | Container cluster service configuration method and device | |
| CN110858843B (en) | Service request processing method and device and computer readable storage medium | |
| CN110519183B (en) | Node speed limiting method and device, electronic equipment and storage medium | |
| CN114666284B (en) | Flow control method and device, electronic equipment and readable storage medium | |
| CN111857992B (en) | Method and device for allocating linear resources in Radosgw module | |
| CN111324886A (en) | Service request processing method and device and server | |
| CN107113323B (en) | Data storage method, device and system | |
| CN111045808A (en) | Distributed network task scheduling method and device | |
| CN113821340A (en) | Dynamic balancing method, system, terminal and storage medium of distributed system | |
| CN111538572A (en) | Task processing method, device, scheduling server and medium | |
| CN113079062B (en) | Resource adjusting method and device, computer equipment and storage medium | |
| CN113835905A (en) | Message queue load balancing method and device, electronic equipment and medium | |
| CN115002105A (en) | Balanced distribution method and device based on disk rate and network rate | |
| CN114827033B (en) | Data flow control method, device, equipment and computer readable storage medium | |
| CN113810461B (en) | Bandwidth control method, device and equipment and readable storage medium | |
| CN111866951B (en) | Service processing method and device | |
| CN111193760B (en) | Information sending method, device and storage medium | |
| CN113608870A (en) | Load balancing method and device of message queue, electronic equipment and storage medium | |
| CN114443274A (en) | Method, device and readable storage medium for adjusting thread number in micro-service thread pool | |
| CN115878568A (en) | Data updating method, device, equipment and computer readable storage medium | |
| CN108243112B (en) | Chat group network flow control method and device, storage medium and computing equipment | |
| CN114500381B (en) | Network bandwidth limiting method, system, electronic equipment and readable storage medium | |
| CN112463754B (en) | Data node switching method and device in HDFS (Hadoop distributed File System) and computer equipment | |
| CN110650358B (en) | Node distribution method, device, node distribution server and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |