KR20060086225A - Distributed virtual environment management systems and methods - Google Patents

Abstract

The present invention relates to a negotiated distributed virtual environment (DVE) management system. The master server receives a redistribution request from the first slave server, a first tree from the first slave server, and a second tree from the second slave server, the first portion of the first zone being the second slave server. Determine that it is controlled by the second slave server and transmit a result indicating that part of the first area is controlled by the second slave server.

Distributed, managed, virtual, environment, server

Description

DISTRIBUTED VIRTUAL ENVIRONMENT MANAGEMENT SYSTEMS AND METHODS}

1 illustrates one embodiment of a distributed virtual environment (DVE) system;

2 illustrates a hardware environment applicable to master, backup, and slave servers;

3 illustrates a software architecture of one embodiment of a DVE management system;

4 illustrates an example Quadtree for a local area;

5A, 5B, 5C, and 5D are flowcharts of one embodiment of a DVE management method;

6A, 6B, 6C illustrate exemplary quadtrees of different aspects for the local area;

7A, 7B illustrate exemplary quadtrees of different aspects for a local area, FIGS.

8 illustrates an example quadtree, and

9 is a diagram illustrating an exemplary quadtree.

* Description of the main parts of the drawings *

11 processing unit 12 memory

13: storage device 14: output device

15 input device 16 communication device

221, 231, 241, 251, 261: local area management module

211: VE management module

The present invention relates to distributed virtual environment management, and more particularly, to a system and method of AOIM (Area Of Interest Management).

With a distributed virtual environment (DVE) system, geographically dispersed users can share a consistent view of the virtual environment (VE) by interacting with each other and exchanging their state information over a network such as the Internet. have. Since DVE systems are growing in terms of VE size and number of concurrent users, it is important that scalability, i.e., how many concurrent users they support interacting in large VEs without sacrificing the functionality of their interactions. It is becoming. One way to improve scalability is the multiple server architecture employed in some DVE systems, as well as most commercial multiplayer network games. Joining a VE to multiple zones, and also distributing the responsibility for managing that zone across multiple servers, can significantly reduce the workload of individual servers, resulting in a system that has more concurrent users and larger VEs. Can support

However, if users are distributed non-uniformly in the DVE, there is an unbalanced workload among servers, i.e., some servers manage more dense areas than others and suffer from heavy workloads. As a result, users managed by servers with heavy loads experience low interaction performance due to the long latency of server status updates. In order to avoid this degradation of interaction performance, a dynamic load balancing method is needed.

A distributed virtual environment (DVE) management system is provided. One embodiment of such a system includes a master server, a first slave server, and a second slave server. The master server receives the redistribution request and the first tree from the first slave server and receives the second tree from the second slave server. The first tree includes a plurality of first nodes. Each first node includes a first concurrent user quantity indicating how many users are present in the first region or a portion of the first region. The second tree includes a plurality of second nodes. Each second node includes a second number of concurrent users indicating how many users are present in the second area or a portion of the second area. In the master server, if the first portion of the first zone is adjacent to the second zone, the sum of the number of second concurrent users in the second zone and the first concurrent user in the first zone of the first zone is the server load limit. If the value is below a predetermined percentage value or server load threshold, it is determined that the first portion of the first region is controlled by the second slave server. In addition, the master server sends a result indicating that a portion of the first zone is controlled by the second slave server, thereby controlling the first portion of the first zone between the first slave server and the second slave server. Enable exchange of

Also provided is a DVE management method performed by a master server. A computer code set is also provided that performs the DVE management method when executed.

Distributed virtual environment (DVE) management systems and methods will become more apparent by reference to the following detailed description of embodiments and the accompanying drawings.

1 is a distributed system comprising a master server 21, a backup server 26, and a slave server 22, 23, 24, 25 operating on a network (preferably the Internet or intranet) using a local connection. 1 is a diagram illustrating an embodiment of a type virtual environment (DVE) system. Those skilled in the art will appreciate that the master server 21, the backup server 26, and the slave servers 22, 23, 24, 25 are connected in different types of networking environments, such as routers, gateways, access points, base station systems, and the like. It will be appreciated that different types of transport means communicate with each other between different types of networking environments. The four areas 31, 32, 33, 34 in the VE are individually managed by the slave servers 22, 23, 24, 25. The size of these regions may not be the same, and their shape is not limited to squares or rectangles and may be various polymorphs.

FIG. 2 shows a master, backup and slave server 21, including a processing device 11, a memory 12, a storage device 13, an input device 14, an output device 15 and a communication device 16. 25 is a diagram showing a hardware environment applicable to 25). The processing device 11 communicates with the memory 12, the storage device 13, the input device 14, the output device 15 and the communication device 16 of the Von Neumann architecture via the bus 17. Connected. There may be one or more processing units 11, as the processor of the computer includes a single central processing unit (CPU), microprocessing unit (MPU), or a plurality of processing units, commonly referred to as parallel processing environments. Preferably, memory 12 is random access memory (RAM), but may also include read-only memory (ROM) or flash ROM. Preferably, memory 12 stores program modules that are executed by processing device 11 to perform DVE management functions. Generally, program modules that perform particular tasks or implement particular abstract data types include routines, programs, objects, components, scripts, web pages, and the like. In addition, those skilled in the art will appreciate that some embodiments may utilize other computer system configurations, including portable devices, multiprocessor-based electronic devices, microprocessor-based electronic devices, or programmable electronic devices, network PCs, minicomputers, mainframe computers, and the like. It will be appreciated that it can be executed. Some embodiments may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices that are based on various remote access structures such as DCOM, CORBA, Web Objects, Web Services, or other similar structures. The storage device 13 may be a hard drive, a magnetic drive, an optical drive, a portable drive or a nonvolatile memory drive. The drive and associated computer readable medium (if necessary) store computer readable instructions, data structures, and program modules non-volatilely. The processing device 11 controlled by the program module received from the memory 12 and also through the input device by the operator supervises the DVE management function.

3 is a diagram illustrating a software architecture for one embodiment of a DVE management system. The slave servers 22 to 25 and the backup server 26 individually include local area management modules 221 to 261, and the master server 21 includes a VE management module 211. Each zone management module uses the quadtree to maintain the number of concurrent users in the local zone allocated by the master server 21, remove child nodes, and create new child nodes for one node. Insert, output a redistribution request with the quadtree to the VE management module 211 according to the number of concurrent users, and receive a redistribution result from the VE management module 211. The VE management module 211 performs area-receding functions so that, depending on the received quadtree, several areas within the VE and, if necessary, the slave servers 22 to 25 and the backup server 26. Determine if additional servers are needed to manage the reduced area managed by.

Each node in the quadtree is divided into two dimensions into four child nodes, corresponding to all / parts of the local area. 4 is a diagram illustrating an exemplary quadtree for a local area that includes four portions R41a through R41d. The node N41 stores the number of concurrent users 31 in the entire area R41, and the nodes N41a, N41b, N41c, and N41d store the number of concurrent users 5, 5 in the areas R41a, R41b, R41d, and R41c. , 5, 16). The four sub-regions of the northwest (NW), northeast (NE), southeast (SE), and southwest (SW) portions may be sequentially arranged by nodes such as N41a, N41b, N41c, and N41d. Local information about all nodes can be stored separately.

5A, 5B, 5C, and 5D show two parts separated by dashed lines for further clarity, i.e., the left part showing the steps performed by the local area management module 221, 231, 241 or 251, and VE management A flowchart for an embodiment of a DVE management method divided into right parts representing the steps performed by module 211 is shown.

In step S411, a movement message indicating that the user has moved from the original area to the destination area is received together with the information about the original area and the destination area. In step S421, it is determined whether a mobile message is being received from another server, and if so, the process proceeds to step S453, otherwise, the process proceeds to step S423. This determination can be made by detecting whether the original area exists in the maintained quadtree. In step S423, a node corresponding to the received original area is located. In step S425, the number of concurrent users in the located node is reduced by one. In step S431, it is determined whether the located node is the root node, and if so, the process proceeds to step S451; otherwise, the process proceeds to step S433. In step S433, a parent node is located. In step S435, the number of concurrent users in the located node is reduced by one. In step S441, according to the updated number of concurrent users, it is determined whether the child node for the located node needs to be merged. For example, if the number of updated concurrent users is lowered to the zone limit, the child nodes for that located node must be merged. The zone limit is used to limit the number of concurrent users in a single leaf node. In step S443, the child node for the located node is removed.

In step S451, it is determined whether a move message is output to another server, and if so, the process ends, otherwise, the process proceeds to step S453. This determination can be made by detecting whether a destination area exists in the maintained quadtree. In step S453, the node corresponding to the destination area is located. In step S455, the number of concurrent users in the located node is increased by one. In step S461, it is determined whether the located node needs to be split, and if so, the process proceeds to step S463; otherwise, the process proceeds to step S471. For example, if the number of updated concurrent users exceeds the area threshold described above, it is necessary to partition the located node. In step S463, four child nodes are created for the located nodes, and the number of simultaneous users is obtained. In step S471, it is determined whether the located node is the root node, and if so, the process proceeds to step S481, otherwise, the process proceeds to step S473. In step S473, the parent node is located. In step S481, it is determined whether the slave server is overloaded, and if so, the process proceeds to step S483, otherwise, the process ends. For example, if the number of concurrent users in the root node exceeds the server load limit, some of the retained area must be moved to another slave server or backup server. In step S483, a redistribution request of the root node is sent to the VE management module 211. In step S485, the redistribution result is received from the VE management module 211, so that control exchange between the slave server and the backup server (if necessary) for a part of the maintained local area is possible. In step S491, a quadtree request is received. In step S493, the retained quadtree is transmitted to the VE management module 211.

In step S511, a redistribution request of the root node is received. In step S513, the area corresponding to the received root node is defined as a dense area. In step S515, a quadtree request is sent to the corresponding slave server in the dense area. In step S517, the quadtree is received from the corresponding slave server. It should be noted that each node of the received quadtree is marked with a corresponding slave server identification that indicates which area is managed by which slave server. In step S521, a slack region is defined that is adjacent to the dense area and has a minimum number of simultaneous users between the adjacent areas. This definition can be made by detecting the number of concurrent users in the root node. In step S523, the number of simultaneous users in the free area and the dense area is obtained as the number of first and second concurrent users, respectively. In step S531, a root node corresponding to the dense area is located. In step S533, the next node corresponding to the current node is located. The next node decision may use known breadth-first traversal or depth-first traversal. In step S535, it is determined whether the located node is a leaf node, and if so, the process proceeds to step S541, otherwise, the process proceeds to step S533. In step S541, it is determined whether the area corresponding to the positioned node is adjacent to the spare area, and if so, the process proceeds to step S543, otherwise, the process proceeds to step S533. Preferably, neighboring nodes are obtained using an efficient quadtree traversal algorithm introduced by Samet. In step S543, the positioned node is marked for merging with the free area. In step S545, the number of concurrent users in the located node is added to the first number of concurrent users and subtracted from the second number of concurrent users. In step S551, it is determined whether the redistribution process is sufficient, and if so, the process proceeds to step S553, otherwise, the process proceeds to step S561. For example, the redistribution process is sufficient if the number of second concurrent users drops to the server load threshold or a predetermined percentage value of the server load threshold described above. In step S553, the final quadtree is transmitted to the corresponding slave server.

In step S561, it is determined whether the spare area is filled, and if so, the process proceeds to step S571, otherwise, the process proceeds to step S533. For example, if the number of first concurrent users exceeds the server load threshold or a predetermined percentage value of the server load threshold, the free area is filled. In step S571, it is determined whether a backup server exists, and if so, the process proceeds to step S573, otherwise, the process ends. In step S573, the next node corresponding to the current node is located. Note that the traversal approach used is the same as in step S533. In step S575, it is determined whether the located node is a leaf node, and if so, the process proceeds to step S577, otherwise, the process proceeds to step S573. In step S577, it is determined whether the area corresponding to the positioned node is adjacent to the spare area, and if so, the process proceeds to step S578, otherwise, the process proceeds to step S573. In step S578, the located node is marked for the backup server. In step S579, the number of concurrent users in the located node is subtracted from the second number of concurrent users. In step S581, it is determined whether the redistribution process is sufficient, and if so, the process proceeds to step S583, otherwise, the process proceeds to step S573. In step S583, the final quadtree is sent to the corresponding slave server and backup server.

A detailed description of the VE management method is given in the following example. 6A, 6B, and 6C are diagrams illustrating exemplary quadtrees of different aspects for the local area R6 maintained by the local area management module 221. In one embodiment, referring to FIG. 6A, the local region R6 is logically divided into four portions R61, R62, R63, and R64, and the subregion R63 is divided into four portions R631, R632, and R633. , R634). The number of simultaneous users in the areas R6, R61, R62, R63, R64, R631, R632, R633, R634 is 30, 5, 5, 10, 10, 5, 2, 1, 2. In step S411, the movement message indicating that the user U1 is moved from the area R634 to the area R64 is received by the local area management module 221. In steps S421, S423, and S425, the number of simultaneous users of the node N634 is reduced by one to be one. In steps S433 and S435, the number of simultaneous users of the node N63 is reduced by one to nine. The area threshold value is set to 10, and in step S441, the nodes N631, N632, N633, and N634 are determined to be merged. In step S443, the nodes N631, N632, N633, and N634 are removed. In steps S431, S433 and S435, the number of simultaneous users of the root node N6 is reduced by one to 29. Transient results are shown in Figure 6b. Next, in steps S431, S451, S453, and S455, the number of simultaneous users of the node R64 is increased by one, to 11. In step S461, the node R64 is determined to be divided. In step S463, four child nodes N641, N642, N644, N643 are created, and their number of concurrent users is 4, 2, 2, 3. In steps S471, S473, and S455, the number of simultaneous users of the root node N6 is increased by one to thirty. Finally, the process ends and the final result is shown in FIG. 6C.

7A and 7B are diagrams illustrating exemplary quadtrees of different aspects of local area R6 maintained by local area management module 221. In another example, referring to FIG. 7A, the local region R6 is logically divided into four portions R61, R62, R63, and R64, and the subregion R64 is divided into four portions R641, R642, R643, Logically divided into R644). The number of concurrent users in the areas R6, R61, R62, R63, R64, R641, R642, R643, R644 is 30, 5, 5, 9, 11, 4, 2, 3, 2. In step S411, the move message indicating that the user U2 is moved from the other slave server to the area N642 is received by the local area management module 221. In steps S421, S453, and S455, the number of simultaneous users of the node R642 is increased by one, to three. In steps S461, S471, S473, and S455, the number of simultaneous users of the node N64 is increased by one to twelve. In steps S461, S471, S473, and S455, the number of simultaneous users of the root node N6 is increased by one to 31. While the server load limit value is set to 30, in step S481, the slave server 221 is determined to be overloaded. In step S483, the redistribution request of the root node N6 is sent to the master server 21. In step S511, the redistribution request of the root node N6 is received by the VE management module 211. In step S513, the area R6 is defined as a dense area. In step S515, the quadtree request is sent to the local area management modules 221, 231, 241, and 251, respectively. 8 is a diagram illustrating an exemplary quadtree held by local area management modules 221, 231, 241, and 251, respectively. In step S517, quadtrees T6 to T9 are received from local area management modules 221, 231, 241, and 251. In step S521, the area R7 is defined as a spare area. In step S523, a first quantity, i.e., 22 and a second quantity, i.e., 31, are obtained. While the depth-first traversal is used, in steps S531 to S541, the nodes N61 and N62 are sequentially positioned. In step S543, node N62 (ie, area R62) is marked for local area management module 231. In step S545, the first and second numbers are updated to 26 and 27, respectively. In step S551, the redistribution process is determined as sufficient. Finally, in step S553, the final quadtree is sent to the local area management modules 221, 231, to exchange control of the area R62.

As another example, in step S511, the redistribution request of the root node N6 is received by the VE management module 211. In step S513, the area R6 is defined as a dense area. In step S515, the quadtree request is sent to the local area management modules 221, 231, 241, and 251, respectively. 9 is a diagram illustrating an exemplary quadtree maintained by local area management modules 221, 231, 241, and 251, respectively. In step S517, quadtrees T6 to T9 are received from local area management modules 221, 231, 241, and 251. In step S521, the area R7 is defined as a spare area. In step S523, a first number, i.e., 22 and a second number, i.e. 38, are obtained. While the depth first traversal is used, in steps S531 to S541, the nodes N61 and N62 are sequentially positioned. In step S543, node N62 (ie, area N62) is marked for local area management module 231. In step S545, the first and second numbers are updated to 27 and 33, respectively. In step S551, the redistribution process is determined to be insufficient. In step S561, since the first number exceeds 80% of the server load threshold, the free area is determined to be full. In steps S571 to S577, the nodes N63, N64, N641, and N642 are sequentially positioned. At step S578, node N642 is marked for local area management module 261. In step S579, the second number is updated to 28. In step S581, the redistribution process is determined to be sufficient. Finally, in step S583, the final quadtree is sent to local area management modules 221, 231, and 261, so that control of areas R62, R642 can be exchanged. The embodiment described above discloses that the received quadtrees are treated individually as four separate trees in the master server 21, but is not limited to this. Those skilled in the art can merge the received quadtree into a single quadtree without departing from the spirit and scope of the present invention and modify it appropriately for a particular process step.

DVE management systems and methods, or certain aspects or portions thereof, may be in the form of program code (ie, instructions) embedded in a type of medium, such as a floppy diskette, CD-ROMS, hard drive, or other machine-readable storage medium. In this case, when the program code is loaded and executed on a machine such as a computer, the machine becomes an apparatus for practicing the present invention. The disclosed methods and systems can also be implemented in the form of program code transmitted via some transmission medium, such as electrical wiring or cable, via an optical fiber, or through other forms of transmission, wherein the program code is received and When loaded and executed on a machine such as a computer, the machine becomes an apparatus for practicing the present invention. When implemented on a general purpose processor, the program code is combined with the processor to provide a dedicated device that operates similarly to certain logic circuits.

Although the present invention has been described in connection with the preferred embodiment, it is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Accordingly, the scope of the invention may be defined and protected by the following claims and their equivalents.

According to the present invention, it is possible to provide a dynamic load balancing method that avoids the degradation of the interaction execution.

Claims (10)

In the distributed virtual environment (DVE) management system, Receiving a redistribution request from the first slave server, Receive a first tree from the first slave server, the first tree comprising a plurality of first nodes individually comprising a first number of concurrent users indicating how many users are in a first zone or a portion of the first zone; , Receive a second tree from the second slave server, the second tree comprising a plurality of second nodes individually including a second number of concurrent users indicating how many users are present in the second realm or part of the second realm; , The first portion of the first region is adjacent to the second region, and the sum of the number of the second concurrent users of the second region and the number of the first concurrent users of the first portion of the first region is a server load. If lower than a threshold or a predetermined percentage of the server load threshold, determine that the first portion of the first region is controlled by the second slave server, Transmit a result indicating that a portion of the first region is controlled by the second slave server, thereby controlling control of the first portion of the first region between the first slave server and the second slave server. Distributed Virtual Environment (DVE) management system that includes a master server that enables exchange. 3. The third server of claim 1, wherein the master server comprises a plurality of third nodes, each comprising a third number of concurrent users indicating how many users are present in a third area or part of the third area. Receiving a tree, and further, if the first portion of the first region is adjacent to the third region and the number of second concurrent users in the second region is lower than the number of third concurrent users in the third region, And determine that a portion of the first zone is controlled by the second slave server. The distributed virtual environment management system of claim 1, wherein the first tree and the second tree are compatible with quadtrees. 2. The master server of claim 1, wherein the second portion of the first region is adjacent to the second region, and wherein the number of the second concurrent users in the second region and the first portion of the first region are determined. The sum of the first number of concurrent users and the number of first concurrent users of the second portion of the first region exceed the server load threshold or a predetermined percentage value of the server load threshold; Determining that a second portion is controlled by a backup server, thereby enabling exchange of control of the second portion of the first region between the first slave server and the backup server. Management system. 2. The distributed virtual environment (DVE) management system of claim 1, wherein the determination of the first portion of the first region is accomplished by traversing the first tree with depth first traversal or width first traversal. 2. The distributed type of claim 1 wherein the first slave server sends the redistribution request when it is detected that the number of the first concurrent users of the root node in the first tree exceeds a server load threshold. Virtual Environment (DVE) Management System. The method of claim 1, wherein the first slave server removes all child nodes for one of the first nodes when it is detected that the first number of concurrent users is lower than an area limit value. Create four child nodes for segmentation when the first number of concurrent users is detected to exceed the region threshold.  8. The method of claim 7, wherein the second slave server removes all child nodes for one of the second nodes when the second concurrent user number is detected to be lower than an area limit, and the second concurrent server. Create four child nodes for one of the second nodes when the number of users is detected to exceed the area threshold. In the distributed virtual environment (DVE) management method that is loaded and executed by the master server, Receiving, by the master server, a redistribution request from a first slave server; The first slab includes a first tree comprising a plurality of first nodes, the master server individually comprising a first number of concurrent users indicating how many users are in a first zone or a portion of the first zone. Receiving from the Eve server; The second slab includes a second tree comprising a plurality of second nodes individually including a second number of concurrent users indicating how many users are present in a second zone or a portion of the second zone. Receiving from the Eve server; Wherein the master server has a first portion of the first region adjacent to the second region, the number of second concurrent users in the second region and the number of first concurrent users in the first portion of the first region; Determining that the first portion of the first region is controlled by the second slave server if the sum of is lower than a server load threshold or a predetermined percentage value of the server load threshold; And Sending, by the master server, a result indicating that a portion of the first region is controlled by the second slave server; And enabling exchange of control of the first portion of the first region between the first slave server and the second slave server. A computer code set for distributed virtual environment (DVE) management that is loaded and executed by a master server, Receiving a redistribution request from a first slave server; Receiving a first tree from the first slave server, the first tree comprising a plurality of first nodes individually comprising a first number of concurrent users indicating how many users are in a first zone or a portion of the first zone that; Receiving a second tree from the second slave server, the second tree comprising a plurality of second nodes individually including a second number of concurrent users indicating how many users are present in the second zone or a portion of the second zone; that; The first portion of the first region is adjacent to the second region, and the sum of the number of the second concurrent users of the second region and the number of the first concurrent users of the first portion of the first region is a server load. If lower than a threshold or a predetermined percentage of the server load threshold, determining that the first portion of the first region is controlled by the second slave server; And Transmitting a result indicating that a portion of the first region is controlled by the second slave server; Computer code set for managing a distributed virtual environment (DVE) that enables the exchange of control of the first portion of the first region between the first slave server and the second slave server.

Images