CN109218442A - A kind of load-balancing method and device of distributed file system - Google Patents
A kind of load-balancing method and device of distributed file system Download PDFInfo
- Publication number
- CN109218442A CN109218442A CN201811215901.7A CN201811215901A CN109218442A CN 109218442 A CN109218442 A CN 109218442A CN 201811215901 A CN201811215901 A CN 201811215901A CN 109218442 A CN109218442 A CN 109218442A
- Authority
- CN
- China
- Prior art keywords
- node
- destination
- address
- use information
- destination node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
-
- 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
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The embodiment of the invention discloses a kind of load-balancing method of distributed file system and devices, this method comprises: receiving the current use information of each node, the use information of each node includes the address of node and memory usage;Destination node is determined according to the memory usage of each node, and obtains the destination address of destination node;The destination address of destination node is sent to the client for initiating request, establishes connection with destination node so that client is based on destination address.In this way, it effectively avoids the problem that the request for being based only on client determines and loads unreasonable distribution caused by destination node, by the current service condition for considering each node, suitable destination node is distributed based on this for the client, and the address of the destination node is fed back into client, it efficiently and is reliably connected to the network between client and the distribution node with realizing, to strengthen network data processing ability, improves the flexibility and availability of load balancing.
Description
Technical field
The present invention relates to load-balancing technique fields, more particularly to a kind of load-balancing method of distributed file system
And device.
Background technique
Currently, for distributed file system, the load balancing of use includes polling method, weighted polling method, most
Small connection number method, randomized, source address Hash method etc..Specific load balancing process, which may is that work as, receives client transmission
Request after, according to the request, using poll, Smallest connection number, random or source address Hash scheduling algorithm, taken from multiple rear ends
Determination will be the back-end server of client distribution in business device (also referred to as, node), to achieve the purpose that load balancing.
But so carry out load balancing, just for receive client transmission request, without consider it is each after
The service condition for holding server is easy to cause the back-end server for client distribution to make since the free memory of itself is insufficient
At processing efficiently and accurately can not be carried out to the request of client.It would therefore be highly desirable to provide a kind of more efficient load balancing
Strategy, to improve the treatment effeciency for the request that back-end server issues client.
Summary of the invention
Above-mentioned in order to solve, the present invention provides a kind of load-balancing method of distributed file system and devices, so that
Before handling the request that client is initiated, avoid be based only on client request determine caused by destination node bear
The problem of carrying unreasonable distribution, it can be considered that the current service condition of each node, the service condition based on each node are
The client distributes suitable destination node, and the address of the destination node is fed back to client, to realize client and be somebody's turn to do
It efficiently and is reliably connected to the network between distribution node, to strengthen network data processing ability, improves load balancing plan
Flexibility and availability slightly.
In a first aspect, providing a kind of load-balancing method of distributed file system, comprising:
Receive the current use information of each node, the use information of each node includes the address of node and interior
Deposit utilization rate;
Destination node is determined according to the memory usage of each node, and obtains the destination address of the destination node;
The destination address of the destination node is sent to the client for initiating request, so that the client is based on
The destination address and the destination node establish connection.
Optionally, the address of node for including in the use information is the physical address of the node.
Optionally, the memory usage according to each node determines destination node and obtains the destination node
Destination address, comprising:
The smallest node of memory usage is searched, destination node is denoted as;
Obtain the corresponding use information of the destination node from the use information of each node, and by the destination node pair
The physical address that the use information answered includes is determined as the destination address of the destination node.
Optionally, the address of node for including in the use information is the physical address and the node of the node
At least one virtual address.
Optionally, the memory usage according to each node determines destination node and obtains the destination node
Destination address, comprising:
The smallest node of memory usage is searched, destination node is denoted as;
Obtain the corresponding use information of the destination node from the use information of each node, and from the destination node pair
At least one corresponding virtual address of the destination node is obtained in the use information answered;
Determine a virtual address as the target section from least one corresponding virtual address of the destination node
The destination address of point.
Second aspect additionally provides a kind of load balancing apparatus of distributed file system, comprising:
Receiving unit, for receiving the current use information of each node, the use information of each node includes should
Address of node and memory usage;
Acquiring unit for determining destination node according to the memory usage of each node, and obtains the target section
The destination address of point;
Transmission unit, for the destination address of the destination node to be sent to the client for initiating request, so as to
The client is based on the destination address and the destination node establishes connection.
Optionally, the address of node for including in the use information is the physical address of the node.
Optionally, the acquiring unit, comprising:
First lookup subelement is denoted as destination node for searching the smallest node of memory usage;
First obtains subelement, corresponding using letter for obtaining the destination node from the use information of each node
It ceases, and the physical address that the corresponding use information of the destination node includes is determined as to the destination address of the destination node.
Optionally, the address of node for including in the use information is the physical address and the node of the node
At least one virtual address.
Optionally, the acquiring unit, comprising:
Second lookup subelement is denoted as destination node for searching the smallest node of memory usage;
Second obtains subelement, corresponding using letter for obtaining the destination node from the use information of each node
Breath, and at least one corresponding virtual address of the destination node is obtained from the corresponding use information of the destination node;
Subelement is determined, for determining a virtual address from least one corresponding virtual address of the destination node
Destination address as the destination node.
In embodiments of the present invention, a kind of load-balancing method of distributed file system is provided, this method comprises: first
First, the current use information of each node is received, the use information of each node includes that the address of node and memory make
With rate;Then, destination node is determined according to the memory usage of each node, and with obtaining the target of the destination node
Location;Then, the destination address of the destination node is sent to the client for initiating request, so that the client is based on
The destination address and the destination node establish connection.In this way, using method provided in an embodiment of the present invention, it can be effective
It avoids the problem that the request for being based only on client determines and loads unreasonable distribution caused by destination node, by considering each node
Current service condition, the service condition based on each node is that the client distributes suitable destination node, and by the mesh
Mark address of node feeds back to client, efficiently and is reliably connected to the network between client and the distribution node with realizing, from
And network data processing ability is strengthened, improve the flexibility and availability of load balancing.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations recorded in the present invention
Example, for those of ordinary skill in the art, is also possible to obtain other drawings based on these drawings.
Fig. 1 is system framework schematic diagram involved in an application scenarios in the embodiment of the present invention;
Fig. 2 is a kind of flow diagram of the load-balancing method of distributed file system provided in an embodiment of the present invention;
Fig. 3 is a kind of structural schematic diagram of the load balancing apparatus of distributed file system provided in an embodiment of the present invention.
Specific embodiment
Inventor has found that the load balancing of use is based only on for current distributed file system
Request transmitted by the client received determines the node of the request to be processed using various Load Sharing Algorithms, still, this
Kind is held just for the load balancing mode of the request of client due to absolutely not considering the current service condition of each node
The node for client distribution is easily led to since itself current memory utilization rate is higher (that is, free memory is insufficient), causing can not
Processing efficiently and accurately is carried out to the request of client.
Based on this, in embodiments of the present invention, the currently used situation based on each node determines that memory usage is suitable
Node be destination node, by the destination node distribute to initiate request client, specifically can be by the mesh of the destination node
Mark address is sent to client, connection is established according to the destination address and destination node by client, to realize by the memory
The reasonable destination node of utilization rate handles the request of client initiation, it is ensured that carries out processing efficiently and accurately to the request.
In this way, the load-balancing method provided effectively avoids and is based only on the request of client and determines load point caused by destination node
With unreasonable problem, destination node that is efficient and being reliably connected to the network can be established for client distribution therewith by realizing,
To strengthen network data processing ability, the flexibility and availability of load balancing are improved.
For example, one of the scene of the embodiment of the present invention, can be applied in scene as shown in Figure 1.The scene
In include: node 1, node 2 ..., node N (N is integer, N >=2), wherein in a kind of situation, any one node
For periodically collecting the use information of each node, and to carry out load balancing as the executing subject in the embodiment of the present invention;Separately
In a kind of situation, and by other load-balancing device or with the use information for periodically collecting each node can also be born
The other equipment for carrying equalization function execute the embodiment of the present invention, are not especially limited in the present invention.It should be noted that this reality
Applying can be used as executing subject in example with node 1 and is illustratively illustrated.
As an example, node 1, node 2 ..., node N calculates the memory usage of itself, and memory made
With rate U1、U2、……、UNWith the address ip of itself1、IP2、……、IPNIt carries in use information, and each node can be with base
It will be respective in Simple Network Management Protocol (English: simple network management protocol, abbreviation: snmp)
Use information is sent to node 1;After node 1 receives the current use information of N number of node, it can be made according to the memory of each node
Suitable memory usage U is determined with rate U5Corresponding node 5, is denoted as destination node, and obtains the target of the destination node
Address ip5;In this way, node 1 can be by destination address IP when receiving the transmission request of client 1005It is sent to the client
End 100, so that the client 100 is based on the destination address IP5Connection is established with the destination node (that is, node 5).Such as
This, the memory usage of the destination node of distribution is suitable, realizes reasonable, efficient and reliable load balancing effect.
It is understood that above-mentioned scene is only a Sample Scenario provided in an embodiment of the present invention, the embodiment of the present invention
It is not limited to this scene.
With reference to the accompanying drawing, by embodiment come the negative of distributed file system a kind of in the present invention will be described in detail embodiment
Carry the specific implementation of equalization methods and device.
Fig. 2 is a kind of flow diagram of the load-balancing method of distributed file system provided in an embodiment of the present invention.
Referring to fig. 2, this method can specifically include:
Step 201, the current use information of each node is received, the use information of each node includes the node
Address and memory usage.
It is understood that each node can periodically (such as every 10 seconds) detect certainly in distributed file system
The relevant parameter index of body, such as: total memory headroom, free memory space and disk buffering space, and joined based on the correlation
Number index calculates the current memory usage of the node, for characterizing the memory service condition of the node.
It can be in the meminfo in each periodic duty/proc file system for each node as an example
File obtains the relevant parameter index of the service condition of current memory, the relevant parameter index specifically got, need using
Index have: total memory size MemTotal, free memory size MemFree, block disk buffering size Buffers and file
Disk buffering size Cached;Using above-mentioned 4 relevant parameter indexs, it is current that the node can be calculated according to the following formula
Memory usage U:
U=100%* (MemTotal-MemFree-Buffers-Cached)/MemTotal
Wherein, MemTotal-MemFree-Buffers-Cached specifically: reject current idle with total memory size
Space (that is, current space available), obtain currently used space size;Currently used space size and total memory size
Ratio, can withdraw deposit the current memory usage of the node.
It is understood that each node can periodically calculate the memory usage of itself, and latest computed is obtained
Memory usage and Self address be carried in current use information to be sent to collect and use information and carry out load balancing
Node;The node updates existing use information to the newest use information of each node received, as current use
Information.
It should be noted that carrying out load balancing to the request that client is initiated in the embodiment of the present invention, it is all based on most
What new use information (that is, current use information) was realized.Such as: assuming that be A in the use information that 13:00:00 is received,
It is B in the use information that 13:00:10 is received, if initiating to request in 13:00:12 client, it is based on use information B
Corresponding node is distributed for the client.
For the address of node in current use information, it can specifically refer to the Internet protocol address of the node
(English: Internet Protocol Address, referred to as: IP address).In a kind of situation, finger joint point can be physically
Location;In another case, being also possible to the physical address of node and at least one virtual address of the node.
Wherein, physical address refers to the actual address or absolute address of node, only one general physics of a node
Address;In contrast, virtual address is not then the actual address of necessary being, and a node can configure at least one virtually
Location, in this way, when different users wants access to identical node the node can be accessed based on different virtual addresses.
It should be noted that carrying the address of node in the current use information of node, being for subsequent load balancing
When facilitate the address for obtaining determining destination node, so that the address of destination node is directly fed back to client, thus quickly,
It is effective to establish network connection.In other implementations, the current memory usage of each node can also be directly received, by
The node for receiving memory usage obtains the corresponding address of node of each current memory usage, and the memory that node is current
The corresponding preservation of utilization rate and address.
It should be noted that respectively current use information can be sent to this reality by agreements such as snmp by each node
Apply the executing subject of example.
Step 202, destination node is determined according to the memory usage of each node, and obtains the mesh of the destination node
Mark address.
When specific implementation, in step 202 according to the memory usage of each node determines destination node and obtains
The destination address of destination node, can specifically include: searching the suitably described node of memory usage, is denoted as destination node;From
The corresponding use information of the destination node is obtained in the use information of each node, and corresponding use of the destination node is believed
The address that breath includes is determined as the destination address of the destination node.
As an example, if the address for including in the current use information of each node received in step 201 is
The physical address of each node, then, destination node is determined according to the memory usage of each node in step 202 and obtains institute
The destination address for stating destination node, can specifically include: searching the smallest node of memory usage, is denoted as destination node;
Obtain the corresponding use information of the destination node from the use information of each node, and by the corresponding use of the destination node
The physical address that information includes is determined as the destination address of the destination node.
It is understood that the memory usage of node can indicate the node currently to the processing capacity of information, specifically
For, the memory usage of node is smaller, indicates that the node is currently used or the space occupied is less, and current there are also a large amount of
Available memory headroom, that is, indicate that the current processing capacity of the node is stronger;, whereas if the memory usage of node is larger,
Indicate that the node is currently used or the space occupied is more, currently available memory headroom is smaller, that is, indicate that the node is worked as
Preceding processing capacity is weaker.
Wherein it is possible to determine that memory uses by for statistical analysis to the current use information of each node received
The smallest node of rate.The smallest node of the memory usage, can indicate in all nodes the node be currently used or
The ratio that the space occupied accounts for total memory headroom is minimum, and current processing capacity is most strong, then, using the node as destination node
It goes to handle the request that following client is sent, it can be ensured that the efficiency and reliability of processing.
As another example, if the address for including in the current use information of each node received in step 201
For each node physical address and at least one virtual address, then, according to the memory usage of each node in step 202
It determines destination node and obtains the destination address of the destination node, can specifically include: searching the smallest institute of memory usage
Node is stated, destination node is denoted as;Obtain the corresponding use information of the destination node from the use information of each node, and from institute
It states and obtains at least one corresponding virtual address of the destination node in the corresponding use information of destination node;Then, from described
Destination address of the virtual address as the destination node is determined at least one corresponding virtual address of destination node.
Wherein, it in order to which each node can be accessed by multiple client, can be configured for each node corresponding multiple
Virtual address, in this way, section at the same time, multiple client can be by different virtual addresses simultaneously and orderly access
Identical node.
It is somebody's turn to do it is understood that determining that a virtual address is used as from least one corresponding virtual address of destination node
The destination address of destination node can specifically determine destination address from virtual address by the way of poll, can also use
The mode randomly selected determines destination address from virtual address, does not limit it in the present embodiment and determines unique virtual address
Implementation.
Such as: assuming that determining that the smallest node of memory usage is node C, receiving the current use information of node C
In, the physical IP including node C: xxx.xxx.x.x and virtual IP address 1:yyy.yyy.y.y and virtual IP address 2:zzz.zzz.z.z,
At this point it is possible to virtual IP address 1 and virtual IP address 2 are determined from the use information, and to the virtual IP address 1 and virtual IP address 2, it is possible to
It is determined as destination address by randomly selecting one of virtual IP address, such as chooses virtual IP address 2 and be used as destination address.
It should be noted that the destination address of the destination node no matter got be physical address or virtual address,
The connection with destination node can be realized by the destination address, do not influence to carry out load balancing to distributed file system.
It is understood that getting the destination address of destination node, load is made in the request initiated for subsequent client
Balanced and processing request provides data basis.
Step 203, the destination address of the destination node is sent to the client for initiating request, so as to the visitor
Family end group establishes connection in the destination address and the destination node.
When specific implementation, after getting the destination address of destination node, that is, it may wait for client and initiate request.When there is visitor
When request is initiated at family end, this is used to periodically collect the use information of each node and carries out the node of load balancing can be by step
203 destination addresses got are sent to client;It, can be corresponding to the destination address after client receives destination address
Node sends the request or access request for establishing connection, in this way, the destination node can establish connection and be handled with client
A series of requests of the client.
In some implementations, if receiving asking for multiple client in the period for receiving use information at one
It asks, if destination address is virtual address, can determine different virtual addresses as the corresponding target of each client
Location is embodied as different clients and distributes the same section so that each destination address is sent respectively to corresponding client
Point (that is, the smallest node of current memory usage).
Such as: assuming that client first initiates request 1 in 13:00:02, client second initiates request 2 in 13:00:05, and saves
Point X the smallest node of memory usage in the use information that 13:00:00 is received is that node C (receives the period of use information
It is 20 seconds), and including 4 virtual IP addresses of C, respectively 1~virtual IP address of virtual IP address 4 in the use information of node C, it is possible to
It is that first determines that destination address is virtual IP address 2 by way of poll, is that second determines that destination address is virtual IP address 4, in this way, the node
Virtual IP address 2 can be sent to first by X, and virtual IP address 4 is sent to second, in this way, client first and second can pass through virtual IP address 2 respectively
It is connected on node C with virtual IP address 4, provides service by the request that node C initiates for client first and second.
In other implementations, if receiving asking for multiple client in the period for receiving use information at one
It asks, first determining destination node can be distributed to and initiate to request earliest client;Then to other than destination node
Other nodes are for statistical analysis, search the smallest node of memory usage in these remaining nodes, are determined as initiating request the
The destination node of two early clients;And so on, until distributing corresponding target section to the client of all initiation requests
Point.
Such as: assuming that client first initiates request 1 in 13:00:02, client second initiates request 2 in 13:00:05, and saves
Point X the smallest node of memory usage in the use information that 13:00:00 is received is node C, the smallest section of memory usage
Point is node D, (period of reception use information is 20 seconds), it is possible to be to initiate request earliest by way of poll
Client first determines that destination address is virtual IP address 2;Then, determine that memory usage is minimum in the node other than node C
Node, that is, node D, as the corresponding destination node of client second, it is possible to be that client second determines that destination address is
Physical IP is sent to second in this way, virtual IP address 2 can be sent to first by the nodes X by physical IP, in this way, client first and second can
It is the request that client first is initiated by node C to be connected respectively on node C and node D by virtual IP address 2 and physical IP respectively
1 provides service, is that the request 2 that client second is initiated provides service by node D.
As it can be seen that a kind of load-balancing method of the distributed file system provided through the embodiment of the present invention, this method packet
Include: firstly, receive the current use information of each node, the use information of each node includes the address of node and interior
Deposit utilization rate;Then, destination node is determined according to the memory usage of each node, and obtains the target of the destination node
Address;Then, the destination address of the destination node is sent to the client for initiating request, so as to client's end group
Connection is established in the destination address and the destination node.In this way, the request for being based only on client is effectively avoided to determine mesh
The problem of marking load unreasonable distribution caused by node is based on each section by considering the current service condition of each node
The service condition of point is that the client distributes suitable destination node, and the address of the destination node is fed back to client, with
It realizes and efficiently and is reliably connected to the network between client and the distribution node, to strengthen network data processing ability, mentioned
The high flexibility and availability of load balancing.
Correspondingly, the embodiment of the invention also provides a kind of load balancing apparatus of distributed file system, such as Fig. 3 institute
Show, which can specifically include:
Receiving unit 301, for receiving the current use information of each node, the use information of each node includes
The address of node and memory usage;
Acquiring unit 302 for determining destination node according to the memory usage of each node, and obtains the target
The destination address of node;
Transmission unit 303, for the destination address of the destination node to be sent to the client for initiating request, with
Toilet states client and establishes connection based on the destination address and the destination node.
Optionally, the address of node for including in the use information is the physical address of the node.
Optionally, the acquiring unit 302, comprising:
First lookup subelement is denoted as destination node for searching the smallest node of memory usage;
First obtains subelement, corresponding using letter for obtaining the destination node from the use information of each node
It ceases, and the physical address that the corresponding use information of the destination node includes is determined as to the destination address of the destination node.
Optionally, the address of node for including in the use information is the physical address and the node of the node
At least one virtual address.
Optionally, the acquiring unit 302, comprising:
Second lookup subelement is denoted as destination node for searching the smallest node of memory usage;
Second obtains subelement, corresponding using letter for obtaining the destination node from the use information of each node
Breath, and at least one corresponding virtual address of the destination node is obtained from the corresponding use information of the destination node;
Subelement is determined, for determining a virtual address from least one corresponding virtual address of the destination node
Destination address as the destination node.
Foregoing description is a kind of associated description of the load balancing apparatus of distributed file system, wherein specific implementation side
Formula and the effect reached may refer to a kind of retouching for the load-balancing method embodiment of distributed file system shown in Fig. 2
It states, which is not described herein again.
As seen through the above description of the embodiments, those skilled in the art can be understood that above-mentioned implementation
All or part of the steps in example method can add the mode of general hardware platform to realize by software.Based on this understanding,
Technical solution of the present invention can be embodied in the form of software products, which can store is situated between in storage
In matter, such as read-only memory (English: read-only memory, ROM)/RAM, magnetic disk, CD etc., including some instructions to
So that a computer equipment (can be the network communication equipments such as personal computer, server, or router) executes
Method described in certain parts of each embodiment of the present invention or embodiment.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for device reality
For applying example, since it is substantially similar to the method embodiment, so describing fairly simple, related place is referring to embodiment of the method
Part explanation.The apparatus embodiments described above are merely exemplary, wherein mould as illustrated by the separation member
Block may or may not be physically separated, and the component shown as module may or may not be physics
Module, it can it is in one place, or may be distributed over multiple network units.It can select according to the actual needs
Some or all of the modules therein achieves the purpose of the solution of this embodiment.Those of ordinary skill in the art are not paying creation
Property labour in the case where, it can understand and implement.
The above is only a preferred embodiment of the present invention, it is not intended to limit the scope of the present invention.It should refer to
Out, for those skilled in the art, under the premise of not departing from the present invention, can also make several improvements
And retouching, these modifications and embellishments should also be considered as the scope of protection of the present invention.
Claims (10)
1. a kind of load-balancing method of distributed file system characterized by comprising
The current use information of each node is received, the use information of each node includes that the address of node and memory make
With rate;
Destination node is determined according to the memory usage of each node, and obtains the destination address of the destination node;
The destination address of the destination node is sent to the client for initiating request, so that the client is based on described
Destination address and the destination node establish connection.
2. the method according to claim 1, wherein the address of node for including in the use information is institute
State the physical address of node.
3. according to the method described in claim 2, it is characterized in that, the memory usage according to each node determines mesh
Mark node and the destination address for obtaining the destination node, comprising:
The smallest node of memory usage is searched, destination node is denoted as;
The corresponding use information of the destination node is obtained from the use information of each node, and the destination node is corresponding
The physical address that use information includes is determined as the destination address of the destination node.
4. the method according to claim 1, wherein the address of node for including in the use information is institute
State the physical address of node and at least one virtual address of the node.
5. according to the method described in claim 4, it is characterized in that, the memory usage according to each node determines mesh
Mark node and the destination address for obtaining the destination node, comprising:
The smallest node of memory usage is searched, destination node is denoted as;
The corresponding use information of the destination node is obtained from the use information of each node, and corresponding from the destination node
At least one corresponding virtual address of the destination node is obtained in use information;
Determine a virtual address as the destination node from least one corresponding virtual address of the destination node
Destination address.
6. a kind of load balancing apparatus of distributed file system characterized by comprising
Receiving unit, for receiving the current use information of each node, the use information of each node includes the node
Address and memory usage;
Acquiring unit for determining destination node according to the memory usage of each node, and obtains the destination node
Destination address;
Transmission unit, for the destination address of the destination node to be sent to the client for initiating request, so as to described
Client is based on the destination address and the destination node establishes connection.
7. device according to claim 6, which is characterized in that the address of node for including in the use information is institute
State the physical address of node.
8. device according to claim 7, which is characterized in that the acquiring unit, comprising:
First lookup subelement is denoted as destination node for searching the smallest node of memory usage;
First obtains subelement, for obtaining the corresponding use information of the destination node from the use information of each node, and
The physical address that the corresponding use information of the destination node includes is determined as to the destination address of the destination node.
9. device according to claim 6, which is characterized in that the address of node for including in the use information is institute
State the physical address of node and at least one virtual address of the node.
10. device according to claim 9, which is characterized in that the acquiring unit, comprising:
Second lookup subelement is denoted as destination node for searching the smallest node of memory usage;
Second obtains subelement, for obtaining the corresponding use information of the destination node from the use information of each node, and
At least one corresponding virtual address of the destination node is obtained from the corresponding use information of the destination node;
Subelement is determined, for determining a virtual address conduct from least one corresponding virtual address of the destination node
The destination address of the destination node.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811215901.7A CN109218442A (en) | 2018-10-18 | 2018-10-18 | A kind of load-balancing method and device of distributed file system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811215901.7A CN109218442A (en) | 2018-10-18 | 2018-10-18 | A kind of load-balancing method and device of distributed file system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109218442A true CN109218442A (en) | 2019-01-15 |
Family
ID=64981025
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811215901.7A Pending CN109218442A (en) | 2018-10-18 | 2018-10-18 | A kind of load-balancing method and device of distributed file system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109218442A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101815033A (en) * | 2010-03-12 | 2010-08-25 | 成都市华为赛门铁克科技有限公司 | Method, device and system for load balancing |
US8874775B2 (en) * | 2008-10-15 | 2014-10-28 | Aster Risk Management Llc | Balancing a distributed system by replacing overloaded servers |
CN106210147A (en) * | 2016-09-13 | 2016-12-07 | 郑州云海信息技术有限公司 | A kind of load-balancing method based on poll and device |
CN107087014A (en) * | 2017-01-24 | 2017-08-22 | 无锡英威腾电梯控制技术有限公司 | A kind of load-balancing method and its controller |
CN107493351A (en) * | 2017-10-09 | 2017-12-19 | 郑州云海信息技术有限公司 | A kind of client accesses the method and device of the load balancing of storage system |
-
2018
- 2018-10-18 CN CN201811215901.7A patent/CN109218442A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8874775B2 (en) * | 2008-10-15 | 2014-10-28 | Aster Risk Management Llc | Balancing a distributed system by replacing overloaded servers |
CN101815033A (en) * | 2010-03-12 | 2010-08-25 | 成都市华为赛门铁克科技有限公司 | Method, device and system for load balancing |
CN106210147A (en) * | 2016-09-13 | 2016-12-07 | 郑州云海信息技术有限公司 | A kind of load-balancing method based on poll and device |
CN107087014A (en) * | 2017-01-24 | 2017-08-22 | 无锡英威腾电梯控制技术有限公司 | A kind of load-balancing method and its controller |
CN107493351A (en) * | 2017-10-09 | 2017-12-19 | 郑州云海信息技术有限公司 | A kind of client accesses the method and device of the load balancing of storage system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11418620B2 (en) | Service request management | |
US10715479B2 (en) | Connection redistribution in load-balanced systems | |
Abujoda et al. | DistNSE: Distributed network service embedding across multiple providers | |
CN108512885A (en) | Network packet to being identified as message queue telemetering transmission packet executes specific action | |
CN110430274A (en) | A kind of document down loading method and system based on cloud storage | |
WO2013104172A1 (en) | Method for realizing load balancing, dns server, and client | |
WO2016180188A1 (en) | Distributed link establishment method, apparatus and system | |
CN110086886B (en) | Dynamic session holding method and device | |
CN112751897B (en) | Load balancing method, device, medium and equipment | |
CN107666474B (en) | Network message processing method and device and network server | |
KR20110036573A (en) | Providing access over an ip network to a server application program | |
CN106506701A (en) | A kind of server load balancing method and load equalizer | |
US20130297703A1 (en) | Peer node and method for improved peer node selection | |
CN103024085A (en) | System and method for processing P2P (peer-to-peer) node request | |
JP2006174374A (en) | Relay system | |
CN105656964B (en) | The implementation method and device of data-pushing | |
CN106027356B (en) | A kind of conversion method and device of Tunnel Identifier | |
CN103401799A (en) | Method and device for realizing load balance | |
JP5154313B2 (en) | SIP message distribution method and SIP message distribution apparatus | |
JP2007219637A (en) | Load balancing system and program therefor | |
JP5917678B1 (en) | Information processing apparatus, method, and program | |
CN109218442A (en) | A kind of load-balancing method and device of distributed file system | |
CN111614792A (en) | Transparent transmission method, system, server, electronic device and storage medium | |
CN107888643A (en) | A kind of UDP load-balancing methods, device and system | |
US9912757B2 (en) | Correlation identity generation method for cloud environment |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190115 |