CN101883113B - Method and physical nodes for realizing overlay network load balance - Google Patents

Method and physical nodes for realizing overlay network load balance Download PDF

Info

Publication number
CN101883113B
CN101883113B CN201010219261.4A CN201010219261A CN101883113B CN 101883113 B CN101883113 B CN 101883113B CN 201010219261 A CN201010219261 A CN 201010219261A CN 101883113 B CN101883113 B CN 101883113B
Authority
CN
China
Prior art keywords
node
load
capacity
physical
module
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.)
Expired - Fee Related
Application number
CN201010219261.4A
Other languages
Chinese (zh)
Other versions
CN101883113A (en
Inventor
王治平
聂晓文
周旭
张棪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ZTE Corp
Institute of Acoustics CAS
Original Assignee
ZTE Corp
Institute of Acoustics CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by ZTE Corp, Institute of Acoustics CAS filed Critical ZTE Corp
Priority to CN201010219261.4A priority Critical patent/CN101883113B/en
Publication of CN101883113A publication Critical patent/CN101883113A/en
Application granted granted Critical
Publication of CN101883113B publication Critical patent/CN101883113B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Data Exchanges In Wide-Area Networks (AREA)
  • Computer And Data Communications (AREA)

Abstract

The invention discloses a method for realizing overlay network load balance, which comprises that: each physical node detects self node, determines itself as a low-load physical node (L node) or a high-load physical node (H node), and acquires corresponding residual load or overload; the L node informs the neighboring physical node of L node per se and residual load; and when searching that the neighboring physical node has the L node and the residual load of the L node is no less than the overload of the H node, the H node migrates all or partial load of a self virtual VS with highest load factor to the VS of the L node. The invention also discloses the physical nodes for realizing the overlay network load balance at the same time. The scheme of the invention can realize overlay network load balance, has the flexibility of the method adopting the VS, can adapt to the real application scene, and reduces the overhead of maintaining the VS at the same time.

Description

A kind of method and physical node of realizing overlay network load balance
Technical field
The present invention relates to the overlapping network constructing technology of internet arena, relate in particular to a kind of method and physical node of realizing overlay network load balance.
Background technology
Along with the development of Internet technology, new demand and problem also continue to bring out.Under the constant prerequisite in the Internet that keeps bottom, by overlapping network, providing required performance is a kind of method of network evolution.Overlapping network framework is layer on the internet, mode with seamless evolve builds, form the overlay network infrastructure of a multilayer, not only can to the distributed control of network and management, support for bottom-layer network operator provides, for government provides the support to network behavior supervision and improvement, can also, by network service interface and programming development interface that to top layer, application provides, support to comprise that equity (P2P, Peer to Peer) network is in interior various whole new set of applications service.
Support that various P2P application are most important purposes of overlapping network, it provides network address translation (NAT by node, Network Address Translation) penetrate auxiliary and overlay network distributed hashtable (DHT, Distributed Hash Table) networking and distributed resource index function thereof, support organized P2P overlay network, thereby exploitation and the operation of for P2P, applying provide a set of public infrastructure.That is to say, node can be formed overlay network by DHT agreement, typical in Chord agreement, forms a structurized P2P network, thereby provides one towards overall resource management and the service providing platform of whole the Internet.
In structurized P2P network, as the P2P network based on Chord ring, conventionally adopt consistent hash function that resource identifier (ID) and node ID are mapped to a unified space, then resource is stored in the immediate subsequent node ID of resource ID under server node on.According to the test of the P2P network to based on Chord ring, need the server node of at least thousands of orders of magnitude, the basis equalization of guarantee load.Therefore the system far below the above-mentioned order of magnitude for physical node number, need to adopt special load balancing.In real P2P structured network, physical node is because the situations such as computing capability, the network bandwidth, disk size cause load capacity to be not quite similar, namely the isomerism problem of physical node.The responsible ID territory of distribution physical node how can be fair, simultaneously, better utilize the resource of each physical node of system, give full play to the ability of each physical node, thereby the load of each physical node is reached unanimity, finally make whole system reach load balancing, this is a good problem to study in overlapping network P2P application.
In existing technical scheme, load balancing is broadly divided into two kinds of methods, and a kind of is the scheme that adopts virtual server (VS, Virtual server), and this is a kind of address space equalization scheme of taking into account physical node isomerism.But this scheme is the resource object size in hypothesis arrival system be equate and resource be under this prerequisite being evenly distributed on the address space of system, depart from real resource object distribution scene, and cannot avoid VS to keep the overhead of routing iinformation.Another kind of scheme is not use VS, adopts the scheme of resource migration, and this mode can reflect real resources distribution scene preferably, but when resource is moved frequently, will bring larger bandwidth consumption and index maintenance expense.
Summary of the invention
In view of this, main purpose of the present invention is to provide a kind of method and physical node of realizing overlay network load balance, realizes overlay network load balance, and, both there is the flexibility that adopts VS method, and can adapt to real application scenarios again, also reduced the expense of safeguarding VS simultaneously.
For achieving the above object, technical scheme of the present invention is achieved in that
A kind of method that realizes overlay network load balance provided by the invention, the method comprises:
Each physical node detects self load, determines certainly as underload physical node (L node) or high load capacity physical node (H node), and obtains corresponding residual negative carrying capacity or overload capacity;
L node will be from notifying neighbours physical node as L node and residual negative carrying capacity;
H node is when the residual negative carrying capacity that inquires neighbours' physical node and have L node and L node is not less than the overload capacity of H node, by all or part of VS that moves to described L node of the load of the highest virtual server (VS) of self load factor.
In such scheme, before each physical node detection self load, the method further comprises: each physical node is invented to n logical node VS according to load capacity size, and each VS is responsible for the different address spaces on overlapping network.
In such scheme, describedly determine from as L node or H node and obtain corresponding residual negative carrying capacity or overload capacity, be specially: in advance each physical node is set to underload thresholding and high load capacity thresholding, each physical node detects self load, when the load capacity of self is less than underload thresholding, mark is from as L node, and obtains residual negative carrying capacity according to high load capacity thresholding; When the load capacity of self is greater than high load capacity thresholding, mark, from as H node, obtains overload capacity according to high load capacity thresholding.
In such scheme, the method further comprises: when the overload capacity of H node is greater than the residual negative carrying capacity of L node, described H node is selected a VS in virtual all VS, when inquiring the current place of described VS physical node and be L node, according to the residual negative carrying capacity of described L node, judge whether to be not less than the overload capacity of self, if the determination result is YES, by all or part of VS that moves to L node of the load of the highest VS of self load factor; If the determination result is NO, continue to select next VS to judge.
In such scheme, the described load by the highest VS of self load factor is all or part of to move on the VS of L node, be specially: the highest load capacity of VS and the residual negative carrying capacity of described L node of load factor that self is fictionalized compares, if the load capacity of the VS that load factor is the highest is not more than the residual negative carrying capacity of described L node, the node ID of the highest VS of load factor is assigned to the new VS that described L node starts; If the load capacity of the VS that load factor is the highest is greater than the residual negative carrying capacity of described L node, the highest ID of load factor in the highest sector address space that VS was responsible for of load factor is assigned to the new VS that described L node starts.
In such scheme, the method further comprises: when physical node adds overlapping network, each VS that described physical node fictionalizes self joins the region that present load rate is higher and removes load sharing;
And/or, when physical node exits overlapping network, described physical node is notified self each VS load that corresponding descendant node will proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the VS of suitable L node startup; If can not surpass, the load of the VS that will exit is proceeded to descendant node;
And/or, while there is L node to the physical node of the adjacent VS of all VS on the physical equipment of place in L querying node, using the physical node of adjacent VS, be that the VS of L node is as exiting node, using described adjacent VS as its descendant node, when the load capacity that exits node is not more than the residual negative carrying capacity of L node at the current place of descendant node, the load of exiting node is proceeded to descendant node.
A kind of physical node of realizing overlay network load balance provided by the invention, this physical node comprises: determination of power load module, search module, comparison module, load migration module, state notifying module, load receiver module; Wherein,
Determination of power load module, for detection of self load, is defined as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity, when as H node, and notice neighbor state module; When as L node, notify status notification module;
Neighbor state module, for when inquiring neighbours' physical node and have L node, notice comparison module;
Comparison module, while being not more than the residual negative carrying capacity of L node for the overload capacity comparing self, notification payload transferring module;
Load migration module, for all or part of VS that moves to described L node of load of the highest VS of load factor that self is fictionalized;
State notifying module, for will be from notifying neighbours physical node as L node and residual negative carrying capacity;
Load receiver module, receives for starting VS the load that neighbours' physical node is the migration of H node.
In such scheme, described comparison module, be further used for when the overload capacity that compares self is greater than the residual negative carrying capacity of L node, in self virtual all VS, select in order a VS, when inquiring the current place of described VS physical node and be L node, according to the residual negative carrying capacity of described L node, judge whether to be not less than the overload capacity of self, if the determination result is YES, notification payload transferring module; If the determination result is NO, continue to select next VS to judge.
In such scheme, described physical node, further comprises: add module and/or exit module and/or polymerization module, wherein,
Add module, for when adding overlapping network, each VS that self is fictionalized joins the region that present load rate is higher and removes load sharing;
Exit module, for when self exits overlapping network, the load of notifying descendant node that each VS is corresponding to proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the VS of suitable L node startup; If can not surpass, the load of the VS that will exit is proceeded to descendant node;
Polymerization module, for determine certainly as L node in determination of power load module after, if inquire the physical node of the adjacent VS of all VS on the physical equipment of place, there is L node, using the physical node of adjacent VS, be that the VS of L node is as exiting node, using described adjacent VS as its descendant node, when the load capacity that exits node is not more than the residual negative carrying capacity of L node at the current place of descendant node, the load of exiting node is proceeded to descendant node.
The invention provides a kind of method and physical node of realizing overlay network load balance, each physical node detects self load, determines certainly as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity; L node will be from notifying neighbours physical node as L node and residual negative carrying capacity; H node is when the residual negative carrying capacity that inquires neighbours' physical node and have L node and L node is not less than the overload capacity of H node, and the load of the VS that load factor that self is fictionalized is the highest is all or part of to move on the VS of described L node; So, can realize overlay network load balance, and, both there is the flexibility that adopts VS method, can adapt to real application scenarios again, also reduced the expense of safeguarding VS simultaneously.
Accompanying drawing explanation
Fig. 1 is the schematic diagram that the present invention realizes a kind of method flow of overlay network load balance;
Fig. 2 is that the present invention realizes the schematic diagram that physical node adds the method flow of overlapping network;
Fig. 3 be the present invention realize to adjacent and respectively two VS on two L nodes carry out the schematic diagram of the method flow of polymerization;
Fig. 4 is the schematic diagram of the structure of the present invention's physical node of realizing a kind of overlay network load balance.
Embodiment
Basic thought of the present invention is: each physical node detects self load, determines certainly as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity; L node will be from notifying neighbours physical node as L node and residual negative carrying capacity; H node is when the residual negative carrying capacity that inquires neighbours' physical node and have L node and L node is not less than the overload capacity of H node, and the load of the VS that load factor that self is fictionalized is the highest is all or part of to move on the VS of described L node.
Below by drawings and the specific embodiments, the present invention is described in further detail.
The present invention realizes a kind of method of overlay network load balance, and as shown in Figure 1, the method comprises following step:
Step 101: each physical node detects self load, according to underload thresholding and the high load capacity thresholding set, determines certainly as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity;
Concrete, each physical node is invented to n VS according to load capacity size, each VS is responsible for the different address spaces on overlapping network, described n gets natural number, in order to increase the probability of Load Balanced, this n address space can be scattered in the discontinuity zone on overlapping network as far as possible, in advance each physical node is set to underload thresholding and high load capacity thresholding, each physical node constantly detects self load, when the load capacity of self is less than underload thresholding, mark is certainly as L node, and obtain residual negative carrying capacity according to high load capacity thresholding, by residual negative carrying capacity with from notify neighbours' physical node as L node, when the load capacity of self is greater than high load capacity thresholding, mark, from as H node, obtains overload capacity according to high load capacity thresholding, and to the residual negative carrying capacity of neighbours' L node notice with from carrying out record as L node,
For example, presetting underload thresholding is with high load capacity thresholding , the load capacity that physical node i detects self is W iif, , physical node i mark is certainly as L node, and residual negative carrying capacity (RL) is: , oneself is notified to neighbours' physical node as L node and RL; If , physical node i mark is certainly as H node, and overload capacity (OL) is: and to the RL of neighbours' L node notice with from carrying out record as L node; Described can be 20% of the rated load of physical node i, it can be 80% of the rated load of physical node i;
Further, when obtaining overload capacity, in order not allow physical node again enter very soon higher load condition, need to set load buffer value r, an overload capacity .
Step 102:H node obtains the highest VS(VSj of load factor self fictionalizing);
Concrete, H node calculates the load factor of each VS, and relatively obtains the VS of a load factor maximum, the load factor P of described i VS vsireferring to formula (1):
P vsi = W vsi d i - - - ( 1 )
Wherein, W vsibe the load capacity of i VS, d ibe i the address space length that VS is responsible; After the load factor of each VS is compared, the VS that obtains load factor maximum is VSj;
Whether the neighbor node in step 103:H querying node record has L node, if had, performs step 104; Otherwise, execution step 105;
Step 104:H node judges whether the residual negative carrying capacity of each L node has the overload capacity that is not less than H node self, whether has RL>=OL, if had, performs step 106; If no, perform step 105;
Step 105:H node is selected in order a VS in virtual all VS, when the current place of this VS physical node is L node, the RL according to this L node, has judged whether RL>=OL, if RL>=OL, performs step 106; If RL<OL, re-executes this step;
The P2P network of take based on Chord ring is example, H node is selected in order a VS in the Finger of virtual all VS table, to described VS, send FindLoad message, whether inquire about the current place of described VS physical node is L node, when the current place of this VS physical node is L node, calculate whether there is RL>=OL, if RL>=OL performs step 106; If RL<OL, re-executes this step; When the current place of this VS physical node is not L node, also re-execute this step;
Wherein, if all physical nodes all cannot be shared the OL of H node, transition process failure.
Step 106: by the load capacity (W of VSj vsj) compare with RL, if RL>=W vsj, perform step 107; If RL<W vsj, perform step 108;
It is VSn that step 107:H node notifies this L node to start a new VS, and the load of VSj is all moved to VSn, namely the node ID of VSj is assigned to VSn, process ends;
It is VSn that step 108:H node notifies this L node to start a new VS, the node ID using the highest ID of the load factor in the sector address space that VSj was responsible for as VSn, and the load migration of this part address space of VSj is to VSn;
Concrete, it is VSn that H node notifies this L node to start a new VS, obtains the node ID of VSn according to the VS resource hash table of H node, calculates load factor corresponding to each ID in the hash table of VS resource that is:, the node ID using the highest ID of corresponding load factor as VSn; The hash table of described VS resource is arranged in each VS, for depositing the resource size that each ID is corresponding, as shown in table 1.
ID value Resource size
vp
id=33 20M
v
Table 1
Wherein, v represents the node ID of certain VS, and the ID of forerunner's node (predecessor) of this VS is vp.
Further, when having physical node to add overlapping network, said method also comprises: each VS of the physical node adding is joined to the region that present load rate is higher and remove load sharing, wherein, in order to reduce the routing cost that increasing of VS is brought as far as possible, the principle that keeps a VS coarsegrain, the present invention first only allows 1~2 VS of a physical node add, and choosing by physical node self-ability of described 1 and 2 determines.Concrete, when adding a VS, as shown in Figure 2, comprise following step:
Step 201: the VS newly adding, in the address space of overlapping network, determines the node ID of one or more VS at random;
In this step, if determine at random the node ID of a VS,, directly as the node ID of the VS newly adding, do not perform step 202;
Below to determine that at random the node ID of three VS describes.
Step 202: find this three corresponding descendant nodes of node ID, be made as respectively N1, N2, N3, resource hash table by N1, N2, N3 obtains three load factors that node ID is corresponding, be respectively PID1, PID2, PID3, the node ID using the node ID of corresponding maximum load rate as the VS newly adding;
Wherein, at hypothesis PID1, be maximum, the node ID that PID1 is corresponding is ID1, ID1 is exactly the point adding, the node ID using ID1 as the VS that will add.
The described node ID using ID1 as the VS that will add, be specially: in the P2P network based on Chord ring, ID1 is assigned to the VS newly adding, as its node ID, carry out join (ID1) function of Chord agreement, the VS newly adding is joined in Chord ring.
Further, when having physical node to exit overlapping network, said method also comprises: each VS of the physical node exiting notifies its load that corresponding descendant node will proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the new VS of suitable L node startup, VS and the descendant node that exit all exit; If can not surpass, the load of the VS that will exit is proceeded to descendant node; Wherein, the described load by the VS that will exit and descendant node proceeds in the new VS of suitable L node startup, be specially: the physical node exiting is using the summation of the load capacity of the load capacity of the VS that will exit and descendant node as overload capacity, search the L node that residual negative carrying capacity is not less than described overload capacity, the load of the load of the VS that exit and its descendant node is proceeded in the new VS of this L node startup;
Along with increasing of VS in system, can increase the expense of physical node maintaining routing information.For keeping as far as possible the coarsegrain of VS, reduce under-utilized VS quantity, said method further comprises: to adjacent and respectively two VS on two L nodes carry out polymerization, as shown in Figure 3, the method comprises following step:
On the physical equipment of step 301:L querying node place, whether the physical node of the adjacent VS of all VS there is L node, if existed, performs step 302, if there is no, and execution step 304;
Step 302:L node is that the VS of L node is as exiting node using the physical node of adjacent VS, using described adjacent VS as its descendant node, whether the load capacity that node is exited in described L node judgement is not more than the residual negative carrying capacity of the L node at the current place of descendant node, if the judgment is Yes, perform step 303, otherwise, execution step 304;
Step 303: the load of exiting node is proceeded to descendant node, exit node and exit;
Described the load of exiting node is proceeded to descendant node, generally, the load information that exits node is proceeded to descendant node by message; The described node that exits exits, and generally, exits node and rolls off the production line.
Step 304: process ends.
In order to realize said method, the present invention also provides a kind of physical node of realizing overlay network load balance, as shown in Figure 4, this physical node comprises: determination of power load module 41, neighbor state module 42, comparison module 43, load migration module 44, state notifying module 45, load receiver module 46; Wherein,
Determination of power load module 41, for detection of self load, according to underload thresholding and the high load capacity thresholding set, determines certainly as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity, when as H node, and notice neighbor state module 42; When as L node, notify status notification module 45;
Neighbor state module 42, for when inquiring neighbours' physical node and have L node, notice comparison module 43;
Comparison module 43, while being not more than the residual negative carrying capacity of L node for the overload capacity comparing self, notification payload transferring module 44;
Load migration module 44, for all or part of VS that moves to described L node of load of the highest VS of load factor that self is fictionalized;
Concrete, the highest load capacity of VS and the residual negative carrying capacity of described L node of load factor that described load migration module 44 fictionalizes self compares, if the load capacity of the VS that load factor is the highest is not more than the residual negative carrying capacity of described L node, the node ID of the highest VS of load factor is assigned to the VS that the load receiver module of described L node starts; If the load capacity of the VS that load factor is the highest is greater than the residual negative carrying capacity of described L node, the highest ID of load factor in the highest sector address space that VS was responsible for of load factor is assigned to the VS that the load receiver module 46 of described L node starts, detailed process is referring to step 108;
State notifying module 45, for will be from notifying neighbours physical node as L node and residual negative carrying capacity;
Load receiver module 46, for starting a new VS, receives the load that neighbours' physical node is the migration of H node;
Described comparison module 43, be further used for when the overload capacity that compares self is greater than the residual negative carrying capacity of L node, in self virtual all VS, select in order a VS, when inquiring the current place of this VS physical node and be L node, according to the residual negative carrying capacity of this L node, judge whether to be not less than the overload capacity of self, if the determination result is YES, notification payload transferring module 44; If the determination result is NO, continue to select next VS to judge; Wherein, selecting all VS still there is no marquis when qualified, end operation;
Further, described physical node, further comprises: add module 47 and/or exit module 48 and/or polymerization module 49, wherein,
Add module 47, for when adding overlapping network, each VS that self is fictionalized joins the region that present load rate is higher and removes load sharing;
Take that to add a VS be example, described in add module 47 in the address space of overlapping network, determine at random the node ID of one or more VS; Wherein, if determine at random the node ID of a VS, the node ID of the VS that directly conduct newly adds;
Below to determine that at random the node ID of three VS describes.
The described module 47 that adds finds this three corresponding descendant nodes of node ID, be made as respectively N1, N2, N3, resource hash table by N1, N2, N3 obtains three load factors that node ID is corresponding, be respectively PID1, PID2, PID3, the node ID using the node ID of corresponding maximum load rate as the VS newly adding;
Exit module 48, for when self exits overlapping network, the load of notifying descendant node that each VS is corresponding to proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the new VS of suitable L node startup; If can not surpass, the load of the VS that will exit is proceeded to descendant node;
Polymerization module 49, for determine certainly as L node in determination of power load module after, if inquire the physical node of the adjacent VS of all VS on the physical equipment of place, there is L node, using the physical node of adjacent VS, be that the VS of L node is as exiting node, using described adjacent VS as its descendant node, whether the load capacity that node is exited in judgement is not more than the residual negative carrying capacity of the L node at the current place of descendant node, if the judgment is Yes, the load of exiting node is proceeded to descendant node; If the judgment is No, end operation.
The above, be only preferred embodiment of the present invention, is not intended to limit protection scope of the present invention, all any modifications of doing within the spirit and principles in the present invention, is equal to and replaces and improvement etc., within all should being included in protection scope of the present invention.

Claims (8)

1. a method that realizes overlay network load balance, is characterized in that, the method comprises:
Each physical node detects self load, determines certainly as underload physical node (L node) or high load capacity physical node (H node), and obtains corresponding residual negative carrying capacity or overload capacity;
L node will be from notifying neighbours physical node as L node and residual negative carrying capacity;
H node is when the residual negative carrying capacity that inquires neighbours' physical node and have L node and L node is not less than the overload capacity of H node, the highest load capacity of VS and the residual negative carrying capacity of described L node of load factor that self is fictionalized compares, if the load capacity of the VS that load factor is the highest is not more than the residual negative carrying capacity of described L node, the node ID of the highest VS of load factor is assigned to the new VS that described L node starts; If the load capacity of the VS that load factor is the highest is greater than the residual negative carrying capacity of described L node, the highest ID of load factor in the highest sector address space that VS was responsible for of load factor is assigned to the new VS that described L node starts.
2. method according to claim 1, it is characterized in that, before each physical node detection self load, the method further comprises: each physical node is invented to n logical node VS according to load capacity size, and each VS is responsible for the different address spaces on overlapping network.
3. method according to claim 1, it is characterized in that, describedly determine from as L node or H node and obtain corresponding residual negative carrying capacity or overload capacity, be specially: in advance each physical node is set to underload thresholding and high load capacity thresholding, each physical node detects self load, when the load capacity of self is less than underload thresholding, mark is from as L node, and obtains residual negative carrying capacity according to high load capacity thresholding; When the load capacity of self is greater than high load capacity thresholding, mark, from as H node, obtains overload capacity according to high load capacity thresholding.
4. method according to claim 3, it is characterized in that, the method further comprises: when the overload capacity of H node is greater than the residual negative carrying capacity of L node, described H node is selected a VS in virtual all VS, when inquiring the current place of described VS physical node and be L node, according to the residual negative carrying capacity of described L node, judge whether to be not less than the overload capacity of self, if the determination result is YES, by all or part of VS that moves to L node of the load of the highest VS of self load factor; If the determination result is NO, continue to select next VS to judge.
5. method according to claim 1, is characterized in that, the method further comprises: when physical node adds overlapping network, each VS that described physical node fictionalizes self joins the region that present load rate is higher and removes load sharing;
And/or, when physical node exits overlapping network, described physical node is notified self each VS load that corresponding descendant node will proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the VS of suitable L node startup; If can not surpass, the load of the VS that will exit is proceeded to descendant node;
And/or, while there is L node to the physical node of the adjacent VS of all VS on the physical equipment of place in L querying node, using the physical node of adjacent VS, be that the VS of L node is as exiting node, using described adjacent VS as its descendant node, when the load capacity that exits node is not more than the residual negative carrying capacity of L node at the current place of descendant node, the load of exiting node is proceeded to descendant node.
6. a physical node of realizing overlay network load balance, is characterized in that, this physical node comprises: determination of power load module, neighbor state module, comparison module, load migration module, state notifying module, load receiver module; Wherein,
Determination of power load module, for detection of self load, is defined as L node or H node, and obtains corresponding residual negative carrying capacity or overload capacity, when as H node, and notice neighbor state module; When as L node, notify status notification module;
Neighbor state module, for when inquiring neighbours' physical node and have L node, notice comparison module;
Comparison module, while being not more than the residual negative carrying capacity of L node for the overload capacity comparing self, notification payload transferring module;
Load migration module, for the highest load capacity of VS and the residual negative carrying capacity of described L node of load factor that self is fictionalized, compare, if the load capacity of the VS that load factor is the highest is not more than the residual negative carrying capacity of described L node, the node ID of the highest VS of load factor is assigned to the new VS that described L node starts; If the load capacity of the VS that load factor is the highest is greater than the residual negative carrying capacity of described L node, the highest ID of load factor in the highest sector address space that VS was responsible for of load factor is assigned to the new VS that described L node starts;
State notifying module, for will be from notifying neighbours physical node as L node and residual negative carrying capacity;
Load receiver module, receives for starting VS the load that neighbours' physical node is the migration of H node.
7. physical node according to claim 6, it is characterized in that, described comparison module, be further used for when the overload capacity that compares self is greater than the residual negative carrying capacity of L node, in virtual all VS, select in order a VS, when inquiring the current place of described VS physical node and be L node, according to the residual negative carrying capacity of described L node, judge whether to be not less than the overload capacity of self, if the determination result is YES, notification payload transferring module; If the determination result is NO, continue to select next VS to judge.
8. physical node according to claim 6, is characterized in that, described physical node, further comprises: add module and/or exit module and/or polymerization module, wherein,
Add module, for when adding overlapping network, each VS that self is fictionalized joins the region that present load rate is higher and removes load sharing;
Exit module, for when self exits overlapping network, the load of notifying descendant node that each VS is corresponding to proceed to, judge whether the load capacity that each descendant node receives after described load can surpass high load capacity thresholding, if surpassed, the load of the VS that will exit and descendant node is proceeded in the VS of suitable L node startup; If can not surpass, the load of the VS that will exit is proceeded to descendant node;
Polymerization module, for determine certainly as L node in determination of power load module after, if inquire the physical node of the adjacent VS of all VS on the physical equipment of place, there is L node, using the physical node of adjacent VS, be that the VS of L node is as exiting node, using described adjacent VS as its descendant node, when the load capacity that exits node is not more than the residual negative carrying capacity of L node at the current place of descendant node, the load of exiting node is proceeded to descendant node.
CN201010219261.4A 2010-06-25 2010-06-25 Method and physical nodes for realizing overlay network load balance Expired - Fee Related CN101883113B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201010219261.4A CN101883113B (en) 2010-06-25 2010-06-25 Method and physical nodes for realizing overlay network load balance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201010219261.4A CN101883113B (en) 2010-06-25 2010-06-25 Method and physical nodes for realizing overlay network load balance

Publications (2)

Publication Number Publication Date
CN101883113A CN101883113A (en) 2010-11-10
CN101883113B true CN101883113B (en) 2014-08-20

Family

ID=43054999

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201010219261.4A Expired - Fee Related CN101883113B (en) 2010-06-25 2010-06-25 Method and physical nodes for realizing overlay network load balance

Country Status (1)

Country Link
CN (1) CN101883113B (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102223254A (en) * 2011-06-17 2011-10-19 南京中兴新软件有限责任公司 Monitoring system and method
CN102523166B (en) * 2011-12-23 2014-10-01 中山大学 Structured network system applicable to future internet
CN102609303B (en) * 2012-01-18 2014-12-17 华为技术有限公司 Slow-task dispatching method and slow-task dispatching device of Map Reduce system
WO2014047902A1 (en) * 2012-09-28 2014-04-03 华为技术有限公司 Load balancing method, device, system and computer readable medium
CN104683254A (en) * 2013-11-29 2015-06-03 英业达科技有限公司 Route control method and route control device
US9632840B2 (en) 2014-04-22 2017-04-25 International Business Machines Corporation Load balancing with granularly redistributable workloads
CN105635199B (en) * 2014-10-28 2019-03-15 睿哲科技股份有限公司 A kind of self-organizing cluster server of holding load equilibrium
CN105592126A (en) * 2014-11-14 2016-05-18 株式会社日立制作所 Agent-free automatic server system
CN106230997B (en) * 2016-09-30 2020-06-05 腾讯科技(北京)有限公司 Resource scheduling method and device
CN106341844A (en) * 2016-11-30 2017-01-18 北京邮电大学 SDN based wireless local area network load balancing method and device the same
CN109617989B (en) * 2018-12-28 2021-11-26 浙江省公众信息产业有限公司 Method, apparatus, system, and computer readable medium for load distribution

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1777120A (en) * 2005-12-07 2006-05-24 中国科学院计算技术研究所 Distributed load equalizing method for structured P2P system
EP2148493A1 (en) * 2008-07-24 2010-01-27 Nokia Siemens Networks OY P2P overlay network for administrative services in a digital network

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1777120A (en) * 2005-12-07 2006-05-24 中国科学院计算技术研究所 Distributed load equalizing method for structured P2P system
EP2148493A1 (en) * 2008-07-24 2010-01-27 Nokia Siemens Networks OY P2P overlay network for administrative services in a digital network

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
A.Rao ,et al..Load Balancing in Structured P2P Systems.《IPTPS》.2003,第2735卷第68-79页.
A.Rao,et al..Load Balancing in Structured P2P Systems.《IPTPS》.2003,第2735卷第68-79页. *

Also Published As

Publication number Publication date
CN101883113A (en) 2010-11-10

Similar Documents

Publication Publication Date Title
CN101883113B (en) Method and physical nodes for realizing overlay network load balance
Rahmati Near optimum random routing of uniformly load balanced nodes in wireless sensor networks using connectivity matrix
Shen et al. Achieving dynamic load balancing through mobile agents in small world P2P networks
CN110336875A (en) A method of improving calculating and the storage speed of Internet of Things application
CN102082832A (en) Data storage method and device in distributed file system
Duan et al. A novel load balancing scheme for mobile edge computing
Messina et al. Decentralised resource finding in cloud/grid computing environments: A performance evaluation
CN101465753B (en) Method, device and system for managing P2P system set
CN101645831A (en) Node organization method in P2P system
CN109947764B (en) Query enhancement system and method for constructing elastic site based on time delay
CN101902388A (en) Expandable fast discovery technology for multi-stage sequencing resources
EP1926276A1 (en) Load balancing in a peer-to-peer system
Han et al. A hybrid P2P overlay network for high efficient search
Msadek et al. Trustworthy self-optimization in organic computing environments
CN1972206A (en) A method for building combination-based P2P system
Li et al. A role-based distributed publish/subscribe system in IoT
Hacini et al. A new overlay P2P network for efficient routing in group communication with regular topologies
Liu The research of resource searching algorithm in mobile P2P network
Li et al. Ontology-based clustering and routing in peer-to-peer networks
Makikawa et al. Balance and proximity-aware Skip Graph construction
Haryanti et al. Improving resource allocation performance in mobile ad hoc grid with mobility prediction
CN101650714B (en) Accurate index-based peer-to-peer searching method
JP6048888B2 (en) Structured P2P network construction method, node device and program thereof
Zeng et al. A novel lookup and routing protocol based on CAN for structured P2P network
Armetta et al. Self-organized routing for unstructured peer-to-peer networks

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20140820

Termination date: 20170625

CF01 Termination of patent right due to non-payment of annual fee