CN101477561A - Large-scale space vector data management method based on content access network - Google Patents

Large-scale space vector data management method based on content access network Download PDF

Info

Publication number
CN101477561A
CN101477561A CNA2009100424510A CN200910042451A CN101477561A CN 101477561 A CN101477561 A CN 101477561A CN A2009100424510 A CNA2009100424510 A CN A2009100424510A CN 200910042451 A CN200910042451 A CN 200910042451A CN 101477561 A CN101477561 A CN 101477561A
Authority
CN
China
Prior art keywords
node
load
space vector
spatial object
spatial
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CNA2009100424510A
Other languages
Chinese (zh)
Other versions
CN101477561B (en
Inventor
景宁
左怀玉
陈荦
吴秋云
李军
唐宇
钟志农
熊伟
王钧
薛丹
程果
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.)
National University of Defense Technology
Original Assignee
National University of Defense Technology
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 National University of Defense Technology filed Critical National University of Defense Technology
Priority to CN2009100424510A priority Critical patent/CN101477561B/en
Publication of CN101477561A publication Critical patent/CN101477561A/en
Application granted granted Critical
Publication of CN101477561B publication Critical patent/CN101477561B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention provides a method for managing large-scale space vector data based on a content access network, and high expansibility of a management mode of the space vector data is obtained by adopting a storage method based on CAN. The effect of rapid retrieval of the space vector data is achieved by adopting a retrieval method based on the CAN and the spatial range of a quadtree. The load balancing degree in the entire space vector data management process is further promoted by adopting the load balancing algorithm, and the expansibility of large-scale space vector data management behaviors based on CAN is enhanced.

Description

The large-scale space vector data management method of content-based accesses network
Technical field
The present invention relates to space vector data management method, relate in particular to a kind of content-based accesses network the large-scale space vector data storage and the search method of (Content-Addressable Network is called for short CAN).
Background technology
Spatial data is one of support key element of typical Geographic Information System application such as location information service in the modern Geographic Information System, intelligent transportation, real estate information service, logistics management and urban infrastructure management, and research spatial data management method is control and the basic demand of regulating flow of material in the modern geography information space, energy stream and stream of people's mobile equilibrium and sustainable development.
Spatial Data Model is to describe space entity and mutual notion of getting in touch thereof in the real world, is the tissue of research spatial data and the basic skills of design space database schema.Spatial Data Model in the modern Geographic Information System has three kinds: key element (Feature) model, network (Network) model and (Field) model.Wherein, feature model is emphasized discrete objects, describes discrete phenomena according to the object edges boundary line, and the realization of this data model is referred to as space vector data.
In the realization of Geographic Information System, the space vector data model shows as space vector data object (abbreviating spatial object as).Spatial object has two key characters: the one, and all spatial objects all embed in the background coordinate space; The 2nd, each spatial object all has coordinate information.Background coordinate space commonly used is a theorem in Euclid space, and reference coordinate system is a cartesian coordinate system.With two-dimentional theorem in Euclid space is example, and its fundamental space object has three classes, is respectively point, line and face.Point is the zero dimension object that ad-hoc location is arranged, and uses coordinate (x, y) expression usually; Line is the dimensional object with border, is made of a plurality of points; Face is the two-dimensional object with zone and border, is made up of a series of closing lines.
In management system based on the space vector data of computer system platform, efficient storage and fast retrieval be two important ingredients.At first, space vector data management is a space vector data of storing various user-defined formats with document form, and this method is absorbed in efficient storage and has been ignored quick retrieval.Along with continuous application and the understanding of people, carry out space vector data management in conjunction with spatial index and become a kind of comparatively ripe method gradually space vector data.Spatial index is to improve the effective ways that space vector data obtained efficient, and it is by taking all factors into consideration the specific aim that Computer Architecture and space vector data multi-dimensional nature strengthen whole retrieval computation process.The development of this technology makes various space vector datas storehouse and management system continue to bring out, and has become the prominent example of space vector data management based on the management mode of these systems.
Though the management method of existing space vector data is in efficient storage and fast comparatively ripe aspect the retrieval, aspect reply space vector data management extensive, meet with new challenge.After the space vector data scale increases to a certain degree, the storage platform of computer system need be expanded and upgrade.Usually the strategy that adopts is to carry out the upgrading of computer system software and hardware by the professional and technical personnel.Can solve the extensive problem of space vector data on this mode surface.But it needs the professional to utilize complex software technology and hardware device to carry out smoothly, so extensibility is powerful inadequately.The complicacy that how to reduce the space vector data management mode of extension becomes people need solution badly in the managing large scale space vector data a problem.
Summary of the invention
The technical problem to be solved in the present invention is: solve the extensibility of large-scale space vector data management, a kind of large-scale space vector data efficient storage of content-based accesses network, the management method of retrieval and controllable load equilibrium fast are provided.
The invention provides a kind of large-scale space vector date storage method, it is characterized in that may further comprise the steps based on CAN:
The first step at computer network storage system platform, makes up structurized CAN;
The background coordinate space R of spatial object is mapped as the virtual coordinates space of CAN;
(VID Neighbors), claims that P is reciprocity storage node, is called for short node for IP, Zone to set up the logic data structure P=of each computing machine among the CAN; Wherein, IP represents the network address of node P; Zone represents the CAN zone of P; VID represents the CAN sign of P, uses binary number representation; Neighbors represents neighbours' node of P, represents with the set of pointers of pointing to other node;
If among the CAN without any node, then new node P 0Directly be set to the CAN node, even P 0Zone=R, P 0VID=0, P 0Neighbors=null;
If there is node among the CAN, then calculate new node P by any node 0Mapping point p; Search the node P that comprises p kDecomposed P kSelf zone and receive the P of new node 0Add; Node P kWith new node P 0The data structure records of each self refresh self;
In second step, set up the spatial object index;
Set up the data structure { X, u ° } of spatial object to be stored, wherein X represents the spatial object feature, u °=(x y) is called the reference mark of spatial object, is used for the position of representation space object quaternary tree;
Foundation is based on the quaternary tree index data structure of background coordinate space R;
Utilize the quaternary tree index that each spatial object is positioned in background coordinate space R, form the index information record of each spatial object;
In the 3rd step, spatial object is stored on the node of CAN;
Each spatial object is mapped as the mapping point among the virtual coordinates space R of CAN;
In CAN, adopt CAN routing algorithm recursive lookup can comprise unique node of this mapping point;
Spatial object is stored on this node, upgrade the data structure records of quadtree's node on this node.
The present invention also provides a kind of spatial dimension search method of large-scale space vector data, and space vector data is stored according to the method described above, it is characterized in that may further comprise the steps:
The first step is set up the data structure of spatial dimension search condition, describes the information of spatial dimension search condition;
Second step is by the node P of request inquiry 0Judge whether that according to index needs carry out the retrieval of this spatial dimension;
The 3rd step, carry out if desired, then calculate the mapping point of quaternary tree root node, and call the CAN routing algorithm and the spatial dimension search condition is sent on the node P at this mapping point place;
In the 4th step, node P carries out querying condition and the result is returned to the node P of request inquiry 0
In the 5th step, node P judges whether that according to index needs continue to carry out this spatial dimension retrieval;
The 6th step, carry out if desired, then node P is according to the mapping point of quaternary tree index calculation child node, calls then on the node that the CAN routing algorithm is sent to the spatial dimension search condition at this mapping point place and carries out recursive query.
The present invention also provides a kind of load-balancing method based on aforesaid space vector data storage means, it is characterized in that may further comprise the steps:
The first step, foundation comprises the data structure of each node of load information.
In second step, each node regularly judges whether to exist load imbalance according to load information.
In the 3rd step,, then carry out load balancing if there is the overweight situation of load in node P.
In CAN, start the formula inquiry mechanism that floods by P and seek other suitable node;
After designated time intervals, judge whether to find a suitable node;
If find a suitable node P 0, then P and P 0Between redistribute load, upgrade data structure records separately.
In the 4th step, the node that load is lighter is removed load balancing.
Judge whether node P exists the light situation of load;
If exist, judge then whether P once carried out the overweight load-balancing algorithm of overload;
If carried out, then remove the load balancing restriction of P.
By taking storage means, obtained the high scalability of space vector data management pattern based on CAN.By taking spatial dimension search method, obtained the space vector data effect of retrieval fast based on CAN and quaternary tree.By adopting load-balancing algorithm, further promoted the load balancing degrees in the whole space vector data management process, strengthened extensibility based on large-scale space vector data management behavior among the CAN.
Description of drawings
Fig. 1 is the schematic flow sheet of the large-scale space vector date storage method based on CAN of the present invention;
Fig. 2, Fig. 3 and Fig. 4 are respectively the schematic flow sheets that the step 11 among Fig. 1, step 12 and step 13 is elaborated according to an embodiment of the present invention;
Fig. 5 is the described spatial dimension search method of an an embodiment of the present invention schematic flow sheet;
Fig. 6 is the described load-balancing method schematic flow sheet of an embodiment of the present invention;
Fig. 7 is the schematic flow sheet that calculates mapping point in an embodiment of the present invention;
Fig. 8 is the CAN routing algorithm schematic flow sheet that an embodiment of the present invention is searched the node that comprises mapping point p;
Fig. 9 is the exemplary plot that the step 11 among Fig. 1 makes up CAN, i.e. Fig. 2 example;
Figure 10 is the exemplary plot that the step 12 among Fig. 1 is set up the spatial object index, i.e. Fig. 3 example.
Embodiment
With reference to the accompanying drawings embodiments of the present invention are described more fully.Under situation about being without loss of generality,, suppose in two-dimensional space, to describe spatial object for explaining conveniently.
Fig. 1 is the schematic flow sheet of the large-scale space vector date storage method based on CAN provided by the invention, comprises three steps: step 11 makes up CAN; Step 12 is set up the spatial object index; Step 13 stores spatial object among the CAN into.Below in conjunction with Fig. 2, Fig. 3 and Fig. 4 each step is described in detail.
Fig. 2 is that its process is according to the detailed description schematic flow sheet of an embodiment of the present invention to the structure of the step 11 among Fig. 1 CAN:
Step 21. is mapped as the virtual coordinates space of CAN with the background coordinate space R of spatial object, and promptly setting CAN virtual coordinates space is R.Wherein, R=((x Min, y Min), (x Max, y Max)), represent (x in the two-dimentional cartesian coordinate space Min, y Min) and (x Max, y Max) rectangular area between two points, the size of rectangular area is determined according to the scope of the spatial object set that will store.
Step 22. is considered as node with the storage unit in the computer storage networks, as in the microcomputer second network storage platform, every microcomputer is considered as a node.The data structure that makes up node P among the CAN be (IP, Zone, VID, Neighbors), wherein, IP represents the network address of P; Zone represents the CAN rectangular area of P, with ((x 0, y 0), (x 1, y 1)) expression; VID represents the CAN sign of P, uses binary number representation; Neighbors represents neighbours' node of P, represents with the set of pointers of pointing to other node.
If among step 23. CAN without any node, then new node P 0Directly add among the CAN, even P 0Zone=R, P 0VID=0, P 0Neighbors=null, algorithm finishes.Otherwise change step 24 over to.
Step 24. is calculated new node P by arbitrary node P among the CAN 0Mapping point p.
Step 25. adopts the CAN routing algorithm to search the node that comprises mapping point p, simultaneously also with P for all neighbours of P 0Information send to this node, be denoted by P k
Step 26.P KIt is two parts that Zone divides equally, and note is made Z 1And Z 2The computation process of dividing equally is: establish P KZone={ (x 1, y 1), (x 2, y 2), then work as | (y 2-y 1)/(y Max-y Min) | | (x 2-x 1)/(x Max-x Min) | the time, Z 1={ (x 1, y 1), (x 2, (y 1+ y 2)/2) }, Z 2={ (x 1, (y 1+ y 2)/2), (x 2, y 2); Otherwise, Z 1={ (x 1, y 1), ((x 1+ x 2)/2, y 2), Z 2={ ((x 1+ x 2)/2, y 1), (x 2, y 2).P KZone=Z 1P 0Zone=Z 2P KVID moves to left one; P 0VID=P K VID+1.
Step 27. is for P kAll neighbours N i, if itself and P kDistance is non-vanishing, i.e. distance (N IZone, P KZone)!=0, then from P kDelete it among the neighbours; If itself and P 0Distance is zero, then it is added P 0Among the neighbours; With P kAnd P 0Join respectively among the other side's the neighbours.
Fig. 9 understands that for example the step 11 among Fig. 1 makes up the process of CAN.As shown in the figure, establish and had A1, B2, C3, D4, six nodes of E5, F6 among the CAN, new node G7 begins to add network from B2, and (Fig. 1 a) is then calculated the mapping point of G7 by B2, search the node D4 (Fig. 1 b) that comprises mapping point; The D4 node decomposes (Fig. 1 c) with self zone; D4 and each self refresh of G7 node self neighborhood (Fig. 1 d).
Fig. 7 calculates node P in the step 24 of Fig. 2 0The process of mapping point p, suitable equally in other processes of embodiment of the present invention, comprise the steps:
Step 71. will need the character string of shining upon, i.e. node P 0IP address corresponding characters string, be converted into Hash code k, k is an integer;
Step 72. is called hash function f 1(k)=(((a 1K+b 1) %n 1) %m 1) k is mapped as integer x;
Step 73. is called hash function f 2(k)=(((a 2K+b 2) %n 2) %m 2) k is mapped as integer y;
Step 74. is returned mapping point p, and the coordinate figure of p is (x+x Min, y+y Min).
Wherein, P 0The IP address spaces be that Hash code k can directly obtain in high level language, as the String.GetHashCode () method in the C# language; Each parameter value situation is in the above-mentioned algorithm: m 1=| x Max-x Min|, n 1=m 1+ 8, a 1=2, b 1=n 1-2, m 2=| y Max-y Min|, n 2=m 2+ 8, a 2=4, b 2=n 2-4, (x Min, y Min) and (x Max, y Max) be the coordinate of virtual coordinates space R.
Fig. 8 is the process of searching the node that comprises mapping point p in the step 25 of Fig. 2, and is suitable equally in other processes of embodiment of the present invention, comprises the steps:
Step 81. is by node P 0Each neighbours' node P i, calculate mapping point p to P IZone apart from d i, establish d kBe wherein minimum one, d kCorresponding neighbours' node is P k
If step 82. p is at P kZone in, i.e. d kEqual 0, then return P k, algorithm finishes.Otherwise, change step 83 over to.
Step 83. makes P 0=P k, change step 81 over to and carry out recurrence.
Fig. 3 sets up the detailed process synoptic diagram of spatial object index according to an embodiment of the present invention to the step 12 of Fig. 1, comprises the steps:
The data structure that step 31. is set up spatial object to be stored is { X, u ° }, wherein, X is point, line, surface or other data structure of describing the spatial object feature, u °=(x y) is called the reference mark of spatial object, is used for the position of representation space object quaternary tree.
Step 32. is set up quaternary tree indexed data structure.Making the regional center of each node representative in the quaternary tree is u=(x u, y u), be called the reference mark of node.After spatial object was by the quaternary tree index, the reference mark u of the quadtree's node at its place was exactly the reference mark u of this spatial object 0Data structure D (u)={ d 1, d 2, d 3, d 4, list} represents that the reference mark is the quadtree's node information of u, d 1, d 2, d 3And d 4Be integer, the subdivision degree of depth of 4 quadrants of expression, list represents to be stored in the spatial object set of this node; With R (u) expression u place, reference mark node region; With the degree of depth of L (u) expression reference mark u in whole quaternary tree tree; With C (u, i) i the offspring (i=1,2,3,4) of expression reference mark u.
Step 33. is calculated reference mark u ° of each spatial object according to data structure invocation step 33.1 to the step 33.3 of the data structure of spatial object and quaternary tree.
Step 33.1. makes reference mark u ° of spatial object { X, u ° } to equal quaternary tree tree root node control point u.
If step 33.2. X is included in the region R (u °) at reference mark u ° place or L (u °) degree of depth reaches the maximum limited depth f of whole quaternary tree tree Max, then returning u °, algorithm finishes.Otherwise turn to step 33.3.
Step 33.3. for each offspring C of u ° (u °, i), if (C (u °, i)) intersects, and then makes u °=C (u ° i), turns to step 33.2 recurrence for X and R.
Wherein, the maximum limited depth f of quaternary tree tree Max, determine f in the present embodiment according to index efficient and index precision Max=10.
Figure 10 for example understands the process of setting up the spatial object index.As shown in the figure, the index process of setting up spatial object W10 and Y11 is respectively to obtain the reference mark O00 of W10 and the reference mark C22 of Y11.For the O00 of reference mark, D (O00)={ 1,1,2,1, { W10}}, R (O00) is whole region-wide, L (O00)=1, C (O00,1)=A20, C (O00,2)=B21, C (O00,3)=C22, C (O00,4)=D23.For the C22 of reference mark, D (C22)={ 1,1,1,1, { Y11}}, R (C22) are 1/4th zones in the lower left corner, L (C22)=2, C (C22,1)=CA, C (C22,2)=CB, C (C22,3)=CC, C (C22,4)=CD.
Fig. 4 stores spatial object among the CAN detailed process synoptic diagram according to an embodiment of the present invention to the step 13 among Fig. 1, comprises the steps:
The mapping point p that step 41. computer memory object { X, u ° } is reference mark u °.
Step 42. is searched the node P that comprises p k
Step 43. is stored in P with spatial object { X, u ° } kOn, and upgrade D (u °) list+=X.
If step 44. u ° father node u pBe sky, then algorithm finishes.Otherwise, by P kCalculate u pMapping point p pFor P kAll neighbours, search and comprise p pNode, note is made P pUpgrade P pOn D (u p) .d i+=1.Make P k=P p, u °=u p, turn to the 44th step recurrence.
Wherein, u ° mapping point computation process is: with u ° coordinate transformation is the algorithm of calling graph 7 again after the character string.
Fig. 5 is the schematic flow sheet of the described spatial dimension search method of an embodiment of the present invention, comprises the steps:
Step 51. is set up data structure (Q, the u of spatial dimension search condition *), wherein, the range of search of Q representation space object.In two-dimensional space, identical in the face of resembling in Q and the space vector model is made of the line of a series of sealings; u *Reference mark in the expression quaternary tree index.
Step 52. is by the node P of request inquiry 0Judging whether needs to carry out retrieval, even u *Equal the reference mark u of quaternary tree root node, judge R (u then *) whether ∩ Q=Φ sets up, if establishment, then algorithm finishes.Otherwise change step 53 over to.
Step 53. is calculated u *Mapping point p, utilize the CAN routing algorithm to seek to comprise the node P of p r
Step 54. is by node P rCarry out inquiry, and call the CAN routing algorithm result is sent to P 0
Step 55. is by P rJudging whether needs to continue to carry out, that is, and and for u *Each offspring C (u *, i), make u *=C (u *, i), if R is (u *) ∩ Q ≠ Φ establishment, then change step 53 recurrence over to, otherwise algorithm finishes.
In the above-mentioned algorithm, because the routing mechanism of quaternary tree index structure and CAN all can provide assurable system performance, therefore whole spatial dimension retrieving also can provide assurable system performance.
From Fig. 1, Fig. 2, Fig. 3, Fig. 4 and example shown in Figure 5 based on the space vector data management flow process of CAN as can be seen: space vector data management method set forth in the present invention is suitable for extensive sight.When the space vector data scale of whole C AN system increases to one when delimiting, only need in CAN, dynamically add node and get final product.Rely on the good self-adaptation extended capability that CAN self is had, whole upgrade process can realize adaptivity fully.
In above-mentioned each implementation procedure of the present invention, the algorithm of employing evenly is mapped as point in the virtual coordinates space with CAN node and spatial object etc., so the load balancing of total system generally is good.Load imbalance when controlling the spatial object skewness the invention provides a kind of controlled load-balancing algorithm.
Fig. 6 is the load balancing schematic flow sheet that the present invention is based on storage of aforesaid space vector data and search method, and process is:
The first step, the data structure of the load information of node is described in foundation.Step 61 as shown, node P among the CAN is expressed as (IP, Zone, VID, Neighbors, Load, ScatterZone), wherein IP, Zone, VID, Neighbors implication are constant; Load represents the load information of P, represents with the spatial object number that is stored on the P; ScatterZone represents the fringe area of P, represents with the pointer that points to other node.
In second step, regularly judge whether to exist load imbalance according to load information by each node.Step 62 as shown, whether load is overweight to judge P, if promptly the Load of P is greater than threshold values L Max, turned to for the 3rd step.Otherwise, turned to for the 4th step.
In the 3rd step, carry out load balancing.Step 63 arrives step 66 as shown:
Step 63. is sought the light node P of load kThat is,, then make P if fringe area P.ScatterZone is not empty k=P.ScatterZone.Otherwise, start the formula inquiry that floods by P, seek Load less than L MinP k
If step 64. does not find P in Preset Time k, then algorithm finishes.Otherwise, turn to step 65.
Step 65. is calculated the minimum frame center u that surrounds of X for any spatial object X of the D among the P (u) .list x, and calculate u xMapping point p, if the horizontal ordinate of p is then deleted X less than the horizontal ordinate of P.Zone central point in the data structure records of node P, be stored to P kIn.Turn to step 66.
Step 66. is with P kJoin among the P.ScatterZone, even P.ScatterZone=P k, algorithm finishes.
In the 4th step, remove load balancing.Step 67 arrives step 69 as shown:
Step 67. is judged node P, and whether load is lighter, and whether the Load that promptly judges P is less than threshold values L MinIf, less than, turn to step 68.Otherwise algorithm finishes.
If the ScatterZone of step 68. P is not empty, promptly P once carried out the overweight load balancing of overload, turned to step 69.Otherwise algorithm finishes.
Step 69. is for any spatial object X that stores in the P.ScatterZone node pointed, calculate the mapping point p at u ° at its reference mark, if p ∈ P.Zone, then X is deleted in former node, be stored among the P, and deletion P.ScatterZone node pointed, even P.ScatterZone=null.Algorithm finishes.
The method of calculating the minimum encirclement frame center of spatial object X in the above-mentioned algorithm in the step 65 is: spatial object X is expressed as a set { X i(i=1,2 ..., n), wherein, each X i=(x i, y i).U then x=(0.5[mm (x i)+max (x i)], 0.5.[mm (y i)+max (y i)]).
In the above-mentioned algorithm, judge when whether P load imbalance occurs, need undetermined coefficient L MaxAnd L Min, the element number of establishing the spatial object set is M, CAN node number is N, then establishes L Max=4M/N; L Min=M/8N.
By adopting the method that content of the present invention provided, can on CAN, store comprising point, line, surface and other complex space object.Because CAN more easily realizes extendability, so whole C AN storage system is embodying stronger adaptive faculty in the face of on the data scale problem.Above-mentioned one embodiment of the present invention is a coordinate space with two-dimentional theorem in Euclid space, and the space vector data of management is point, line, surface or its compound object.To those skilled in the art, it is feasible it being extended to the more management of higher dimensional space vector data.
The present invention must determine the background coordinate space scope R of whole space vector data collection when the storage space vector data, can not support in the storing process that this is a qualification of the scope of application of the present invention to the on-the-fly modifying of R.How the present invention is not particularly limited among the CAN each node at the local index space vector data, has therefore kept the capacity of self-government of node among the CAN.In addition, the load-balancing algorithm among the present invention also can omit according to space vector data collection distribution situation when specific implementation.Above-mentioned various modification or variation are in order to design the various instantiations of special-purpose to those skilled in the art.

Claims (5)

1. large-scale space vector date storage method based on CAN is characterized in that may further comprise the steps:
The first step at computer network storage system platform, makes up structurized CAN;
The background coordinate space R of spatial object is mapped as the virtual coordinates space of CAN;
(VID Neighbors), claims that P is reciprocity storage node, is called for short node for IP, Zone to set up the logic data structure P=of each computing machine among the CAN; Wherein, IP represents the network address of node P; Zone represents the CAN zone of P; VID represents the CAN sign of P, uses binary number representation; Neighbors represents neighbours' node of P, represents with the set of pointers of pointing to other node;
If among the CAN without any node, then new node P 0Directly be set to the CAN node, even P 0Zone=R, P 0VID=0, P 0Neighbors=null;
If there is node among the CAN, then calculate new node P by any node 0Mapping point p; Search the node P that comprises p kDecomposed P kSelf zone and receive new node P 0Adding; Node P kWith new node P 0The data structure records of each self refresh self;
In second step, set up the spatial object index;
Set up the data structure { X, u ° } of spatial object to be stored, wherein X represents the spatial object feature, u °=(x y) is called the reference mark of spatial object, is used for the position of representation space object quaternary tree;
Foundation is based on the quaternary tree index data structure of background coordinate space R;
Utilize the quaternary tree index that each spatial object is positioned in background coordinate space R, form the index information record of each spatial object;
In the 3rd step, spatial object is stored on the node of CAN;
Each spatial object is mapped as the mapping point among the virtual coordinates space R of CAN;
In CAN, adopt the CAN routing algorithm to search unique node that can comprise this mapping point;
Spatial object is stored on this node, upgrade the data structure records of quadtree's node on this node.
2. large-scale space vector date storage method according to claim 1 is characterized in that calculating the process of node mapping point, comprises the steps:
With node P 0The IP address spaces be integer k;
According to the dimension N of background coordinate space R, utilize different hash functions that k is mapped as N integer, N integer is mapped in the scope of background coordinate space R, be that the point of coordinate is mapping point with this N round values.
3. the spatial dimension search method of large-scale space vector data, space vector data is stored in accordance with the method for claim 1, it is characterized in that may further comprise the steps:
The first step is set up the data structure of spatial dimension search condition, describes the information of spatial dimension search condition;
Second step is by the node P of request inquiry 0Judge whether that according to index needs carry out the retrieval of this spatial dimension;
The 3rd step, carry out if desired, then calculate the mapping point of quaternary tree root node, and call the CAN routing algorithm and the spatial dimension search condition is sent on the node P at this mapping point place;
In the 4th step, node P carries out retrieval and the result is returned to the node P of request inquiry 0
In the 5th step, node P judges whether that according to index needs continue to carry out this spatial dimension retrieval;
The 6th step, carry out if desired, then node P is according to the mapping point of quaternary tree index calculation child node, calls then on the node that the CAN routing algorithm is sent to the spatial dimension search condition at this mapping point place and carries out recursive query.
A managing large scale space vector data load-balancing method, use the described method storage of claim 1 large-scale space vector data, it is characterized in that may further comprise the steps:
The first step, the data structure of the load information of node is described in foundation;
Node P among the CAN is expressed as (IP, Zone, VID, Neighbors, Load, ScatterZone), wherein IP, Zone, VID, Neighbors implication are identical with claim 1; Load represents the load information of P, represents with the spatial object number that is stored on the P; ScatterZone represents the fringe area of P, represents with the pointer that points to other node;
In second step, regularly judge whether to exist load imbalance according to load information by each node;
Whether load is overweight to judge P, if the Load of P is greater than threshold values L Max, turned to for the 3rd step; Otherwise, turned to for the 4th step;
In the 3rd step, carry out load balancing;
Seek the light node P of load kIf promptly P.ScatterZone is not empty, then makes P k=P.ScatterZone; Otherwise, start the formula inquiry that floods by P, seek Load less than threshold values L MinP k
If in Preset Time, do not find P k, then method finishes;
Otherwise,, calculate the minimum frame center u that surrounds of X for any spatial object X among the P x, and calculate u xMapping point p, if the horizontal ordinate of p is then deleted X less than the horizontal ordinate of P.Zone central point in the data structure records of node P, be stored to P kIn; With P kJoin among the P.ScatterZone, even P.ScatterZone=P k, method finishes;
In the 4th step, remove load balancing;
Whether load is lighter to judge node P, and whether the Load that promptly judges P is less than threshold values L Min:
If more than or equal to, then method finishes;
If less than, and the ScatterZone of P is not empty, for any spatial object X that stores in the P.ScatterZone node pointed, calculate the mapping point p at u ° at its reference mark, if p ∈ P.Zone then deletes X in former node, be stored among the P, and deletion P.ScatterZone node pointed, even P.ScatterZone=null.
A kind of large-scale space vector data management according to claim 4 load-balancing method, it is characterized in that: threshold values L Max=4M/N, L Min=M/8N, wherein, M is the element number of spatial object set, N is a CAN node number.
CN2009100424510A 2009-01-09 2009-01-09 Large-scale space vector data management method based on content access network Expired - Fee Related CN101477561B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2009100424510A CN101477561B (en) 2009-01-09 2009-01-09 Large-scale space vector data management method based on content access network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2009100424510A CN101477561B (en) 2009-01-09 2009-01-09 Large-scale space vector data management method based on content access network

Publications (2)

Publication Number Publication Date
CN101477561A true CN101477561A (en) 2009-07-08
CN101477561B CN101477561B (en) 2010-12-29

Family

ID=40838277

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2009100424510A Expired - Fee Related CN101477561B (en) 2009-01-09 2009-01-09 Large-scale space vector data management method based on content access network

Country Status (1)

Country Link
CN (1) CN101477561B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101833986A (en) * 2010-05-20 2010-09-15 哈尔滨工业大学 Method for creating three-stage audio index and audio retrieval method
CN102063486A (en) * 2010-12-28 2011-05-18 东北大学 Multi-dimensional data management-oriented cloud computing query processing method
CN102708148A (en) * 2012-03-31 2012-10-03 深圳祥云信息科技有限公司 Duplication eliminating method based on multidimensional lattice data spatial model
CN103744999A (en) * 2014-01-23 2014-04-23 中国人民解放军国防科学技术大学 Spatial vector data online interactive mapping method based on hierarchical-divided storage structure

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1731394A (en) * 2005-08-18 2006-02-08 上海交通大学 Method for implementing electronic urban map system having intelligent query function
CN101246597A (en) * 2007-11-30 2008-08-20 北京师范大学 Method for commercial scale vector map data gradual transmission and dynamic reconstruction

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101833986A (en) * 2010-05-20 2010-09-15 哈尔滨工业大学 Method for creating three-stage audio index and audio retrieval method
CN102063486A (en) * 2010-12-28 2011-05-18 东北大学 Multi-dimensional data management-oriented cloud computing query processing method
CN102063486B (en) * 2010-12-28 2013-06-05 东北大学 Multi-dimensional data management-oriented cloud computing query processing method
CN102708148A (en) * 2012-03-31 2012-10-03 深圳祥云信息科技有限公司 Duplication eliminating method based on multidimensional lattice data spatial model
CN102708148B (en) * 2012-03-31 2014-04-30 深圳祥云信息科技有限公司 Duplication eliminating method based on multidimensional lattice data spatial model
CN103744999A (en) * 2014-01-23 2014-04-23 中国人民解放军国防科学技术大学 Spatial vector data online interactive mapping method based on hierarchical-divided storage structure
CN103744999B (en) * 2014-01-23 2017-01-25 中国人民解放军国防科学技术大学 Spatial vector data online interactive mapping method based on hierarchical-divided storage structure

Also Published As

Publication number Publication date
CN101477561B (en) 2010-12-29

Similar Documents

Publication Publication Date Title
Nishimura et al. MD-HBase: A scalable multi-dimensional data infrastructure for location aware services
CN106528773B (en) Map computing system and method based on Spark platform supporting spatial data management
CN103150394B (en) Distributed file system metadata management method facing to high-performance calculation
Greenstein et al. DIFS: A distributed index for features in sensor networks
KR101687124B1 (en) Method for generating a dataset structure for location-based services and method and system for providing location-based services to a mobile device
CN109885576A (en) A kind of Hash table creation method and system calculate equipment and storage medium
CN102289466A (en) K-nearest neighbor searching method based on regional coverage
CN101477561B (en) Large-scale space vector data management method based on content access network
CN105357247B (en) Multidimensional property cloud resource range lookup method based on layering cloud peer-to-peer network
CN105893605B (en) Distributed Computing Platform and querying method towards space-time data k NN Query
CN104252457A (en) Method and device for managing data set
Kumar et al. M-Grid: a distributed framework for multidimensional indexing and querying of location based data
Zheng et al. Repose: distributed top-k trajectory similarity search with local reference point tries
US9292559B2 (en) Data distribution/retrieval using multi-dimensional index
Corral et al. Algorithms for joining R-trees and linear region quadtrees
CN115408382B (en) Vector space information gridding storage management method for large-scale satellite data
Lee et al. An efficient spatio-temporal index for spatio-temporal query in wireless sensor networks
Gong et al. Object-oriented and integrated spatial data model for managing image, DEM, and vector data
CN105354310A (en) MapReduce based map tile storage layout optimization method
Alamri et al. Vertical indexing for moving objects in multifloor environments
CN117671392B (en) Grid division method, device, equipment and medium for homeland space data
CN114064995B (en) Static skyline query method based on position in road network environment
CN108052536A (en) A kind of file system of IoT equipment
Gao et al. Queries based on data attribute spatial index in wireless sensor networks
KR20060071981A (en) A unified index of moving objects in road 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
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20101229

Termination date: 20140109