CN103269387A - Automatic configuration method for data center network addresses - Google Patents
Automatic configuration method for data center network addresses Download PDFInfo
- Publication number
- CN103269387A CN103269387A CN2013101902342A CN201310190234A CN103269387A CN 103269387 A CN103269387 A CN 103269387A CN 2013101902342 A CN2013101902342 A CN 2013101902342A CN 201310190234 A CN201310190234 A CN 201310190234A CN 103269387 A CN103269387 A CN 103269387A
- Authority
- CN
- China
- Prior art keywords
- node
- fulcrum
- matched
- network
- data center
- 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
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides an automatic configuration method for data center network addresses. The automatic configuration method for the data center network addresses comprises (1) carrying out layer dividing on nodes in a network design blueprint and nodes in a physical network topological graph, (2) adopting the nodes with the highest layer in the network design blueprint as a supporting point, forming a set of nodes to be matched through all the nodes with the highest layer in the physical network topological graph, and finding the nodes capable of being matched with the supporting point of the network design blueprint in the set of the nodes to be matched to form an initial supporting point pair, and (3) matching all the nodes in the network design blueprint and the corresponding nodes in the physical network topological graph from the initial supporting point pair to transverse the network design blueprint. The automatic configuration method for the data center network addresses can obviously improve efficiency of an algorithm and is wide in application range.
Description
Technical field
The present invention relates to the automatic configuring technical of computer network field, specifically, the present invention relates to a kind of data center network address automatic distributing method.
Background technology
Cloud computing is as a kind of emerging computation schema, in recent years in field extensive uses such as science calculating, network service, mass data storage.And data center is also built in a large number as the infrastructure of cloud computing.At present popular data center's scale is often very huge, comprises tens thousand of even hundreds of thousands station server usually.A large amount of servers couple together by switch, and communicate by the ICP/IP protocol of standard, have just formed general data center's structure.
Equipment in the data center need carry out necessary configuration could operate as normal.For considerations such as high usage route, scheduling of resource and disaster-tolerant backups, data center network is encoded to information such as network topology, device location in the device address usually, and this device address is called logical number, for example 10.0.0.4.Simultaneously, in data center, each equipment has a uniqueness sign (for example MAC Address) physically, and this uniqueness sign is called physics numbering, for example 00:42:C9:45:D8:EA.At configuration phase, need namely need the physics numbering of equipment and the logical number of the relevant information of having encoded to be mapped the mapping one by one of forming device physics numbering and logical number according to the devices allocation logical number that is planned in the data center.As mentioned before, data center is often in large scale, may comprise tens thousand of even hundreds of thousands station server and connect their switch, the manual configuration device address takes time and effort and occurs configuration error easily, and therefore how fast automatic configuration device logical number is an important research direction in the cloud computing technology with the requirement at reply large-scale data center.
People such as R.N.Mysore have proposed a kind of Portland topological structure, based on the Portland topological structure, can use distributed location discovery agreement (LDP) to determine device location according to the MAC Address of equipment, dispose automatically thereby finish the network address.But LDP only is applicable to multi-level tree network topology structure, can't guarantee the applicability under other network topology structures.
In addition, also exist a class to utilize network topology study to carry out the network address scheme of configuration automatically.In Ethernet, bridge can be by the topological relation in the automatic learning network of mode of broadcasting packet.In order to make this network topology learning method can be applicable to larger ethernet environment, people such as C.Kim have proposed a kind of SEATTLE placement algorithm, it uses single-hop DHT to distribute ARP state in the switch, thereby reaches purposes such as obtaining network topological information, device location relation.Yet the SEATTLE placement algorithm still needs the switch broadcasting packet, and the load meeting of every switch increases along with the quantity of end host, and this mode of learning that needs to send broadcasting packet still is difficult to be applicable to the data center network environment.
People such as Kai Chen have proposed a kind of mapping algorithm based on subgraph isomorphism (GI), be called for short the DAC algorithm, can reference: Kai Chen, Chuanxiong Guo, Haitao Wu, etal, DAC:Generic and Automatic Address Configuration for Data Center Networks[C] //Proc of SIGCOMM2010.NJ:ACM, 2010.The DAC algorithm is based on the isomorphism character of figure, judges node in the network design blueprint and the mapping relations of the node in the actual physics topology, and then finishes the network address and dispose automatically.Particularly, the DAC algorithm needs network design blueprint (being blue print) and physical network topology figure as input information.Wherein, the network design blueprint comprises the neighbouring relations between logical number, device type and the node of each node, and Fig. 1 shows a simple network design blueprint example.Wherein, the logical number of each node is unique, and these logical numbers through special coding have comprised network topology structure, information such as internodal position relation.The physical network topology map logo physical connection relation of data center apparatus, usually number as its physics with the MAC Address of each equipment, to guarantee uniqueness, and by neighbouring relations between Topology Discovery recording equipment type and node, Fig. 2 shows a simple physics network topology illustrated example.In the DAC algorithm, above input information is abstracted into two non-directed graphs, G
b=(V
b, E
b) represent network design blueprint, G
p=(V
p, E
p) the represents physical network topological diagram.
In the DAC algorithm, can form mapping relations in order to judge among two figure between interdependent node, (decompose), decomposition (split), reconstruct (refinement) operation are extracted in definition out.Wherein suppose figure G=(V, E) in, V is the set on all summits among the G, V is by some disjoint nonvoid subset (0＜i＜n) form.
Extract out (decompose): the subclass θ that has a node v and a V among the figure G
i, v ∈ θ
i,
With v from θ
iThe middle extraction out equals element v from θ
iThe middle taking-up, new set of independent formation v}, and will be { among the v} adding V.
Decompose (split): at figure G=(V, have the subclass θ of V in E)
1, θ
2Use θ
1Decompose θ
2Mean all θ
2In element, calculate at θ
1In with it the number k on the summit that is connected is arranged, and with θ
2In element be divided into some subclass according to the size of k.
Reconstruct (refinement): this operation is to extracting out and being used in combination of decomposing, when to figure G=(V, E) after draw out operation is carried out on a certain summit in, use newly-generated subclass to decompose other subclass successively, when producing, new subclass continues it is carried out draw out operation, do not finish when not having the subclass that can be decomposed in V, the state that scheme G this moment is called stable state.
If the vertex set among two figure all is in stable state, and subclass number unanimity in the vertex set, element number all equates in the vertex subset of position correspondence, claims the vertex set homotype (apposition) of two figure; Through extracting out and decomposing, if having only this summit itself in the subclass at place, a summit, then claim this summit to be in separate state (singleton); When summits all in the vertex set all is in separate state, claim that this vertex set is (disperse) that disperses.When in two figure figure, extracting a pair of node out, and two sets of node are carried out reconstructed operations, if two vertex sets homotype all the time in this process illustrates that this a pair of node shines upon successfully; In two figure, seek node repeatedly and carry out to extract reconstruct out, up to the two vertex set all be in discrete state and in this process homotype all the time, then can judge two isomorphism of graphs, and can make up internodal mapping one by one according to corresponding relation.
In the DAC algorithm, matching process is completely random.Particularly, at first in the network design blueprint, selecting a node at random, and then at random in physical network topology figure, select a node, both mate, if unsuccessful, change a node among the physical network topology figure and mate, up to the match is successful.Next, looking for one at random again in blue print does not have the node that the match is successful, and it is similarly mated, and to the last, the node in the network design blueprint full figure is all finished coupling with the respective nodes of physical network topology figure.For ease of understanding, provide the picked at random node example that the match is successful He it fails to match respectively below in conjunction with Fig. 3.Fig. 3 shows a simple network design blueprint and corresponding physical network topology figure, supposes to select at random in the DAC algorithm network design blueprint G
bIn l
5Node and physical network topology figure G
pIn d
2Node mates, then G
b, G
pInitial condition is:
G
b:{l
1,l
2,l
3,l
4,l
5,l
6,l
7,l
8}
G
p:{d
1,d
2,c1
3,d
4,d
5,d
6,d
7,d
8}
Extract l out
5Node and d
2Become behind the node:
G
b:{{l
5},{l
1,l
2,l
3,l
4,l
6,l
7,l
8}}
C
p:{{d
2},{d
1,d
3,d
4,d
5,d
6,d
7,d
8}}
Reconstruct forms stable state:
G
b:{{l
5},{l
1,l
2,l
7,l
8},{l
3,l
4,l
6,}}
C
p:{{d
2},{d
1},{d
3,d
4,d
5,d
6,d
7,d
8}}
Obvious G
b, G
pReconstruction result can not be corresponding one by one, i.e. the two different shaped.So network design blueprint G
bIn l
5Node and physical network topology figure G
pIn d
2The node coupling is unsuccessful.
Suppose to select at random network design blueprint G
bIn l
5Node and physical network topology figure G
pIn d
1Node mates, then G
b, G
pInitial condition is:
G
b:{l
1,l
2，l
3，l
4，l
5，l
6，l
7，l
8}
C
p:{d
1,d
2,d
3,d
4,d
5,d
6,d
7,d
8}
Extract l out
5Node and d
2Become behind the node:
G
b:{{l
5},{l
1,l
2,l
3,l
4,l
6,l
7,l
8}}
C
p:{{d
1},(d
2,d
3,d
4,d
5,d
6,d
7,d
8}}
First step reconstruction result is:
G
b:{{l
5},{l
1,l
2,l
7,l
8},{l
3,l
4,l
6}}
C
p:{{d
1},{d
2,d
3,d
5,d
7},{d
4,d
6,d
8}}
The second step reconstruct forms stable state:
G
b:{{l
5},{l
1,l
2,l
7l
8},{l
6},{l
3,l
4}}
C
p:{{d
1},{d
2,d
3,d
5,d
7},{d
6},{d
4,d
8}}
Obvious G
b, G
pReconstruction result is corresponding one by one, i.e. the two homotype.So network design blueprint G
bIn l
5Node and physical network topology figure G
pIn d
1The match is successful for node.
The DAC algorithm can be applicable to the network environment of various scales, various topological structures, automatically is the equipment configuration address.Yet, because physical network topology figure and network design blueprint are often larger, usually comprise tens thousand of even hundreds of thousands station server and supporting a large amount of switches, the chance of success of single match is little, it is longer that this causes shining upon preprocessing process in the DAC algorithm, recall step (i.e. a node after it fails to match the step of coupling) again and too much, make that the efficient of the automatic allocation plan in this address is on the low side.
In sum, currently press for a kind of automatic allocation plan in data center network address that can be applicable to that various scales, various topological structure and efficient are high.
Summary of the invention
The purpose of this invention is to provide a kind of automatic allocation plan in data center network address that can be applicable to that various scales, various topological structure and efficient are high.
For achieving the above object, the invention provides a kind of data center network address automatic distributing method, comprise the following steps:
1) is node division level among network design blueprint and the physical network topology figure;
2) with the number of plies is the highest in the network design blueprint node as fulcrum, the highest node of all numbers of plies of physical network topology figure is formed set of node to be matched, and seek in set of node to be matched can to form initial fulcrum right with the node of the fulcrum coupling of network design blueprint;
3) from initial fulcrum to, traverses network devises and draw the blueprint, with all nodes in the network design blueprint all with physical network topology figure in respective nodes pairing.
Wherein, in the described step 1), for network design blueprint and physical network topology figure, carry out following substep respectively:
11) the same set of packing into after the number of plies of all nodes among the figure is put 0, be called not layering set, take out in this is searched node that device type is server in the layering set and from then on it gathered, the node that takes out constitutes a new set, is referred to as reference set;
12) all node levels in the current not layering set are added 1, and do not seeking the node that annexation is arranged with the node of current reference set in the layering set, have the node of annexation to take out the node of these and reference set and constitute new set, this new set is exactly new reference set;
13) repeated execution of steps 12) be combined into sky up to stratified set not.
Wherein, described step 2) comprise following substep:
21) according to network design blueprint G
b=(V
b, E
b) level divide the result, select the node P that the number of plies is the highest
bAs fulcrum;
22) according to physical topology figure G
p=(V
p, E
p) level divide the result, the node that the number of plies is the highest is all put into a set, is referred to as set of node V to be matched
m
23) use P
bDecompose V
b, with fulcrum P
bFrom V
bThe middle extraction out and reconstruct V
b, make V
bReach stable state;
24) at set of node V to be matched
mNode P of middle selection
p, decompose V with this node
p, and reconstruct V
p, work as V
pCompare V after reaching stable state
pWith V
bHomotype whether is if homotype is then selected this node P
pWith fulcrum P
bForm initial fulcrum to (P
b, P
p), otherwise at V
mThe new node of middle searching is carried out and is extracted out and reconstructed operation, can make V until finding
pWith V
bThe node P of homotype
pForm initial fulcrum to (P
b, P
p).
Wherein, in the described step 3), for the present node P that traverses
b', by the connected fulcrum P that matched
bIn physical network topology figure, find out corresponding fulcrum P
p, use and fulcrum P
pConnection and the number of plies and present node P are arranged
b' node that equates forms current set of node to be matched, and in set of node to be matched, seek can with present node P
b' pairing node P
p', form new fulcrum to (P
b', P
p'); Go round and begin again in the network design blueprint all nodes all with physical network topology figure in respective nodes finish pairing.
Wherein, in the described step 3), from initial fulcrum to, travel through described network design blueprint in the mode of depth-first.
Wherein, described step 3) comprises following substep:
31) right based on initial fulcrum, with fulcrum P
bBe starting point, depth-first traversal network design blueprint G
b, in ergodic process, present node is designated as P
b';
32) right based on initial fulcrum, at physical topology figure G
pIn find out and fulcrum P
pConnection and the number of plies and present node P are arranged
b' equal node, and form current set of node V to be matched
m';
33) carry out extraction and reconstructed operation, at current set of node V to be matched
m' middle energy and the P of seeking
b' form the node P of mapping
p', constitute matched node to (P
b', P
p'), continue execution in step 34;
34) continue depth-first traversal network design blueprint G
b, repeating step 31) and to 33), finding out with fulcrum has all matched node of annexation to (P
b', P
p');
35) with matched node to (P
b', P
p') right as new fulcrum, repeating step 31) to 34), until network design blueprint G
bWith physical topology figure G
pAll nodes all are in separate state among two figure, at this moment network design blueprint G
bWith physical topology figure G
pNode match one by one.
Wherein, described step 33) in, if the current set of node V to be matched of traversal
m' after can't find can and P
b' form the node of mapping, the then automatic configuration failure in data center network address.
Wherein, described data center network address automatic distributing method also comprises the steps 4):
4) the pairing result according to step 3) is dispensed to the apparatus logic numbering node with corresponding physics numbering.
Compared with prior art, the present invention has following technique effect:
1, can take full advantage of the network configuration characteristics, make efficiency of algorithm significantly promote.
2, can support the address configuration of the data center of various scales, various topological structures.
Description of drawings
Fig. 1 shows the example of a network design blueprint;
Fig. 2 shows the example of a physical network topology figure;
Fig. 3 shows a network design blueprint and corresponding physical network topology figure;
Fig. 4 illustrates the flow chart of one embodiment of the invention;
Fig. 5 shows network design blueprint and the corresponding physical network topology illustrated example in the one embodiment of the invention.
Embodiment
Below in conjunction with the drawings and specific embodiments the present invention is done description further.
According to one embodiment of present invention, provide a kind of based on fulcrum network address method of automatic configuration, it is divided the equipment level based on the result of Topology Discovery, position and type according to device node are determined the equipment place number of plies, rely on the number of plies to seek fulcrum then and mate, and rely on fulcrum again, dwindle node set to be matched according to the annexation of the number of plies, equipment room, thereby significantly improve the success rate of coupling, reduce and recall step, and then improve the network address efficient of configuration automatically.
With reference to figure 4, the comprising the following steps: based on fulcrum network address method of automatic configuration (hereafter is PFAC) of present embodiment
Step 1: be all node division levels among network design blueprint and the physical network topology figure.
The division of the equipment level of PFAC is determined the equipment place number of plies based on the result of Topology Discovery according to position and the type of device node.According to one embodiment of present invention, carry out the following step for the network design blueprint.
Step 11: the same set of packing into after at first the number of plies of all nodes among the figure being put 0, be called not layering set, then in this set, search node that device type is server and from then on it gathered in take out, the node that takes out constitutes a new set, is referred to as reference set; In data center, server node is generally terminal node (for example leaf node in the tree), therefore can think the node of the bottom;
Step 12: then, all node levels during layering is not gathered are added 1, and do not seeking the node that annexation is arranged with the node of current reference set in the layering set, have the node of annexation to take out the node of these and reference set and constitute new set, this new set is exactly new reference set;
Step 13: repeated execution of steps 12 is combined into sky up to stratified set not.At this moment the level of all nodes is all divided and is finished.
In like manner, for physical network topology figure, also carry out above-mentioned steps 11 to 13 similarly, can all divide the level of all nodes among the physical network topology figure out.
For ease of understanding, provide the example that a level is divided below in conjunction with Fig. 5.Fig. 5 shows a network design blueprint G
bWith a physical network topology figure G
pFor network design blueprint G
b, the same set of packing into after the number of plies of all nodes among the figure put 0 obtains:
{l
1(0),l
2(0),l
3(0),l
4(0),l
5(0),l
6(0),l
7(0),l
8(0),l
9(0)}
Take out server node, all the other node levels add 1, obtain:
{l
1(1),l
2(1),l
3(1),l
4(1),l
5(1)}[l
6(0),l
7(0),l
8(0),l
9(0)]
Take out the node that connection is arranged with [l6, l7, l8, l9], all the other numbers of plies add 1, obtain:
{l
1(2),l
2(2),l
3(2)}[l
4(1),l
5(1)]l
6(0),l
7(0),l
8(0),l
9(0)
Take out the node that connection is arranged with [l4, l5] again, all the other numbers of plies add 1, obtain:
{l
1(3)}[l
2(2),l
3(2)]l
4(1),l
5(1),l
6(0),l
7(0),l
8(0),l
9(0)
All node levels of network design blueprint are all divided and are finished like this.According to the method described above, can the physical network topology figure among Fig. 5 be divided equally, repeat no more herein.
Step 2: it is right to seek initial fulcrum.
After finishing node level (being the equipment level) division, in the network design blueprint, select the node that the number of plies is the highest as fulcrum, the highest node of all numbers of plies is formed set of node to be matched among the physical topology figure, and it is right to seek the node composition fulcrum that can mate with the fulcrum of network design blueprint from set of node to be matched.According to one embodiment of present invention, it comprises that step is as follows:
Step 21: according to network design blueprint G
b=(V
b, E
b) level divide the result, select the node P that the number of plies is the highest
bAs fulcrum;
Step 22: according to physical topology figure G
p=(V
p, E
p) level divide the result, the node that the number of plies is the highest is all put into a set, is referred to as set of node V to be matched
m
Step 23: set of node V more to be matched
mThe number of plies and the node P of middle node
bThe number of plies, if unequal, G is described
bWith G
pIsomorphism can't be not the devices allocation logical number according to the network design blueprint of input, and the PFAC algorithm finishes; Otherwise, execution in step 24;
Step 24: use P
bDecompose V
b, with fulcrum P
bFrom V
bThe middle extraction out and reconstruct V
b, make V
bReach stable state;
Step 25: at V
mNode P of middle selection
p, decompose V with this node
p, and reconstruct V
p, work as V
pCompare V after reaching stable state
pWith V
bHomotype whether is if homotype is then selected this node P
pWith fulcrum P
bForm initial fulcrum to (P
b, P
p), otherwise at V
mThe new node of middle searching is carried out and is extracted out and reconstructed operation, can make V until finding
pWith V
bThe node P of homotype
pForm initial fulcrum to (P
b, P
p), P wherein
pBe the fulcrum among the physical topology figure.
For example, with reference to figure 5, divide the result according to network design blueprint level and select the highest node l of the number of plies
1As fulcrum, obtain after the execution in step 24:
({l
1},{l
2,l
3},{l
4,l
5},{l
6,l
7,l
8,l
9})
At V
mThe middle node d that selects
1, use d
1Decompose V
pAfter reaching stable state be:
({d
1},{d
2,d
6},{d
3,d
7},{d
4,d
5,d
8,d
9})
The l of the two homotype, so node as can be seen
1, d
1Coupling, l
1, d
1Constitute described initial fulcrum to (P
b, P
p).
In the step 25, if traversal V
mAfter fail to find can with P
bForming the right node of initial fulcrum, can't be the devices allocation logical number according to the network design blueprint then.
Step 3: coupling residue node.
Through aforementioned process, it is right to have obtained initial fulcrum, next based on this, the residue node is shone upon, and according to one embodiment of present invention, it comprises that step is as follows:
Step 31: right based on initial fulcrum, with fulcrum P
bBe starting point, depth-first traversal network design blueprint G
b, in ergodic process, present node is designated as P
b';
Step 32: right based on initial fulcrum, at physical topology figure G
pIn find out and fulcrum P
pConnection and the number of plies and present node P are arranged
b' equal node, and form current set of node V to be matched
m';
Step 33: carry out and extract out and reconstructed operation, at current set of node V to be matched
m' middle energy and the P of seeking
b' form the node P of mapping
p', constitute matched node to (P
b', P
p'), continue execution in step 34; If the current set of node V to be matched of traversal
m' after can't find can and P
b' forming the node that shines upon, the PFAC algorithm stops;
Step 34: continue depth-first traversal network design blueprint G
b, repeating step 31 to 33, finding out with fulcrum has all matched node of annexation to (P
b', P
p');
Step 35: with matched node to (P
b', P
p') right as new fulcrum, repeating step 31 to 34 is until network design blueprint G
bWith physical topology figure G
pAll nodes all are in separate state among two figure, and this moment, the PFAC algorithm finished, network design blueprint G
bWith physical topology figure G
pNode match one by one.
Still with reference to figure 5, repeating step 31 to 34 can make network design blueprint G
bWith physical topology figure G
pAll nodes all are in separate state among two figure, network design blueprint G
bEnd-state is:
({l
1},{l
2},{l
3},{l
4},{l
5},{l
6,l
7,l
8,l
9})
Physical topology figure G
pEnd-state is:
({d
1},{d
2},{d
6},{d
3},{d
7},{d
4,d
5,d
8,d
9})
Step 4: the pairing result according to step 3 is dispensed to the apparatus logic numbering node with corresponding physics numbering.
In above-described embodiment, can with a station server as the core calculations node deployment in the data center network environment, and set up communication port at this node and the equipment room that needs assignment logic numbering.In order to guarantee the trouble-free operation of PFAC, can define two application layer protocols: topology information acquisition protocols (PCP) and logical number distribution protocol (LDP).The former is responsible for collecting the physics topology information, and this process adopts mode from bottom to top, and each equipment is oneself physical address, device type, and gives the core calculations node on the physical address of neighbor node, obtains the physical topology figure of whole network with this; The latter is responsible for the distribution of logical number, and this process adopts top-down mode, and the core calculations node will shine upon one by one according to the physics numbering-logical number that PFAC calculates, and logical number is transferred to corresponding device, number automatic assigning process with this completion logic.
It should be noted that at last, above embodiment is only in order to describe technical scheme of the present invention rather than the present technique method is limited, the present invention can extend to other modification, variation, application and embodiment on using, and therefore thinks that all such modifications, variation, application, embodiment are in spirit of the present invention and teachings.
Claims (8)
1. a data center network address automatic distributing method comprises the following steps:
1) is node division level among network design blueprint and the physical network topology figure;
2) with the number of plies is the highest in the network design blueprint node as fulcrum, the highest node of all numbers of plies of physical network topology figure is formed set of node to be matched, and seek in set of node to be matched can to form initial fulcrum right with the node of the fulcrum coupling of network design blueprint;
3) from initial fulcrum to, traverses network devises and draw the blueprint, with all nodes in the network design blueprint all with physical network topology figure in respective nodes pairing.
2. data center network address automatic distributing method according to claim 1 is characterized in that, in the described step 1), for network design blueprint and physical network topology figure, carries out following substep respectively:
11) the same set of packing into after the number of plies of all nodes among the figure is put 0, be called not layering set, take out in this is searched node that device type is server in the layering set and from then on it gathered, the node that takes out constitutes a new set, is referred to as reference set;
12) all node levels in the current not layering set are added 1, and do not seeking the node that annexation is arranged with the node of current reference set in the layering set, have the node of annexation to take out the node of these and reference set and constitute new set, this new set is exactly new reference set;
13) repeated execution of steps 12) be combined into sky up to stratified set not.
3. data center network address automatic distributing method according to claim 1 is characterized in that, described step 2) comprise following substep:
21) according to network design blueprint G
b=(V
b, E
b) level divide the result, select the node P that the number of plies is the highest
bAs fulcrum;
22) according to physical topology figure G
p=(V
p, E
p) level divide the result, the node that the number of plies is the highest is all put into a set, is referred to as set of node V to be matched
m
23) use P
bDecompose V
b, with fulcrum P
bFrom V
bThe middle extraction out and reconstruct V
b, make V
bReach stable state;
24) at set of node V to be matched
mNode P of middle selection
p, decompose V with this node
p, and reconstruct V
p, work as V
pCompare V after reaching stable state
pWith V
bHomotype whether is if homotype is then selected this node P
pWith fulcrum P
bForm initial fulcrum to (P
b, P
p), otherwise at V
mThe new node of middle searching is carried out and is extracted out and reconstructed operation, can make V until finding
pWith V
bThe node P of homotype
pForm initial fulcrum to (P
b, P
p).
4. data center network address automatic distributing method according to claim 1 is characterized in that, in the described step 3), for the present node P that traverses
b', by the connected fulcrum P that matched
bIn physical network topology figure, find out corresponding fulcrum P
p, use and fulcrum P
pConnection and the number of plies and present node P are arranged
b' node that equates forms current set of node to be matched, and in set of node to be matched, seek can with present node P
b' pairing node P
p', form new fulcrum to (P
b', P
p'); Go round and begin again in the network design blueprint all nodes all with physical network topology figure in respective nodes finish pairing.
5. data center network address automatic distributing method according to claim 4 is characterized in that, in the described step 3), from initial fulcrum to, travel through described network design blueprint in the mode of depth-first.
6. data center network address automatic distributing method according to claim 1 is characterized in that, described step 3) comprises following substep:
31) right based on initial fulcrum, with fulcrum P
bBe starting point, depth-first traversal network design blueprint G
b, in ergodic process, present node is designated as P
b';
32) right based on initial fulcrum, at physical topology figure G
pIn find out and fulcrum P
pConnection and the number of plies and present node P are arranged
b' equal node, and form current set of node V to be matched
m';
33) carry out extraction and reconstructed operation, at current set of node V to be matched
m' middle energy and the P of seeking
b' form the node P of mapping
p', constitute matched node to (P
b', P
p'), continue execution in step 34;
34) continue depth-first traversal network design blueprint G
b, repeating step 31) and to 33), finding out with fulcrum has all matched node of annexation to (P
b', P
p');
35) with matched node to (P
b', P
p') right as new fulcrum, repeating step 31) to 34), until network design blueprint G
bWith physical topology figure G
pAll nodes all are in separate state among two figure, at this moment network design blueprint G
bWith physical topology figure G
pNode match one by one.
7. data center network address automatic distributing method according to claim 6 is characterized in that, described step 33) in, if the current set of node V to be matched of traversal
m' after can't find can and P
b' form the node of mapping, the then automatic configuration failure in data center network address.
8. data center network address automatic distributing method according to claim 1 is characterized in that, also comprises the steps 4):
4) the pairing result according to step 3) is dispensed to the apparatus logic numbering node with corresponding physics numbering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101902342A CN103269387A (en) | 2013-05-21 | 2013-05-21 | Automatic configuration method for data center network addresses |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101902342A CN103269387A (en) | 2013-05-21 | 2013-05-21 | Automatic configuration method for data center network addresses |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103269387A true CN103269387A (en) | 2013-08-28 |
Family
ID=49012996
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013101902342A Pending CN103269387A (en) | 2013-05-21 | 2013-05-21 | Automatic configuration method for data center network addresses |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103269387A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016090975A1 (en) * | 2014-12-10 | 2016-06-16 | 中兴通讯股份有限公司 | Detection processing method and apparatus |
| CN106998299A (en) * | 2016-01-22 | 2017-08-01 | 华为技术有限公司 | The recognition methods of the network equipment, apparatus and system in data center network |
| CN109688758A (en) * | 2019-02-11 | 2019-04-26 | 南方科技大学 | For determining method, apparatus, electronic equipment and the medium of data center's failure |
| CN109756381A (en) * | 2019-02-11 | 2019-05-14 | 南方科技大学 | Data center's Fault Locating Method, device, electronic equipment and medium |
| TWI832776B (en) * | 2022-07-07 | 2024-02-11 | 大陸商蘇州磐聯集成電路科技股份有限公司 | Method for pairing nodes |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1514605A (en) * | 2003-03-27 | 2004-07-21 | 中国科学院计算机网络信息中心 | Hierarchical exchange network system |
-
2013
- 2013-05-21 CN CN2013101902342A patent/CN103269387A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1514605A (en) * | 2003-03-27 | 2004-07-21 | 中国科学院计算机网络信息中心 | Hierarchical exchange network system |
Non-Patent Citations (1)
| Title |
|---|
| 张赣等: "基于支点的数据中心网络地址快速自动配置方法研究", 《计算机科学》, vol. 40, no. 5, 15 May 2013 (2013-05-15) * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016090975A1 (en) * | 2014-12-10 | 2016-06-16 | 中兴通讯股份有限公司 | Detection processing method and apparatus |
| CN105743727A (en) * | 2014-12-10 | 2016-07-06 | 中兴通讯股份有限公司 | Detection processing method and apparatus |
| CN105743727B (en) * | 2014-12-10 | 2019-11-05 | 南京中兴新软件有限责任公司 | Detection processing method and device |
| CN106998299A (en) * | 2016-01-22 | 2017-08-01 | 华为技术有限公司 | The recognition methods of the network equipment, apparatus and system in data center network |
| CN106998299B (en) * | 2016-01-22 | 2019-10-18 | 华为技术有限公司 | The recognition methods of the network equipment, apparatus and system in data center network |
| CN109688758A (en) * | 2019-02-11 | 2019-04-26 | 南方科技大学 | For determining method, apparatus, electronic equipment and the medium of data center's failure |
| CN109756381A (en) * | 2019-02-11 | 2019-05-14 | 南方科技大学 | Data center's Fault Locating Method, device, electronic equipment and medium |
| CN109756381B (en) * | 2019-02-11 | 2022-02-25 | 南方科技大学 | Data center fault positioning method and device, electronic equipment and medium |
| CN109688758B (en) * | 2019-02-11 | 2022-04-12 | 南方科技大学 | Method, apparatus, electronic device, and medium for determining data center failure |
| TWI832776B (en) * | 2022-07-07 | 2024-02-11 | 大陸商蘇州磐聯集成電路科技股份有限公司 | Method for pairing nodes |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gkantsidis et al. | Random walks in peer-to-peer networks | |
| CN103269387A (en) | Automatic configuration method for data center network addresses | |
| CN104580519B (en) | A kind of method of rapid deployment openstack cloud computing platforms | |
| Chen et al. | Generic and automatic address configuration for data center networks | |
| Di Fatta et al. | Fault tolerant decentralised k-means clustering for asynchronous large-scale networks | |
| CN109189751A (en) | Method of data synchronization and terminal device based on block chain | |
| CN102403718B (en) | Generating method for power grid topological relationship based on Arcgis | |
| CN105677648B (en) | A kind of Combo discovering method and system based on label propagation algorithm | |
| CN102075402A (en) | Virtual network mapping processing method and system | |
| Chen et al. | DAC: Generic and automatic address configuration for data center networks | |
| CN105099951A (en) | Resource scheduling method and management equipment | |
| CN103281211B (en) | Large-scale network node system for managing in groups and management method | |
| CN107209746A (en) | model parameter fusion method and device | |
| CN102917084A (en) | Automatic allocation method of IP address of node inside fat tree structure networking data center | |
| CN109101338A (en) | A kind of block chain processing framework and its method based on the extension of multichannel chip | |
| CN104091034B (en) | Circuit communication chart automatic mapping system and method based on Graphic Pattern Matching | |
| Samani et al. | Multilayer resource-aware partitioning for fog application placement | |
| CN101222391A (en) | Method for constructing topological tree in server/client terminal structure | |
| CN102298716A (en) | Identifier generation method and device and application method of identifier | |
| CN103840871B (en) | Encoding and reconstruction method of three-dimensional model topological data robust transmission in wireless network | |
| CN105426228A (en) | OpenStack virtual machine placement method facing streaming media live broadcasting and video transcoding | |
| CN106681795B (en) | Virtual network mapping method for node local topology and available resource capacity | |
| Li et al. | HFSA: A semi-asynchronous hierarchical federated recommendation system in smart city | |
| CN113488996A (en) | Power distribution network protogram modeling method based on distributed parallel graph computing framework | |
| CN105207825A (en) | Fast synchronizing system and synchronizing method for network elements |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130828 |