JP5556507B2 - Virtual machine management program, virtual machine management method, and virtual machine management apparatus - Google Patents

Description

  The present invention relates to a virtual machine management program, a virtual machine management method, and a virtual machine management apparatus that manage virtual machines.

  Conventionally, an information processing system (multiple hierarchy system) in which a plurality of computers share processing hierarchically has been used. As a multi-tier system, for example, a three-tier system having a Web server that provides an interface for using the system, an App (Application) server that executes processing on the system, and a DB (Database) server that manages data is known. . Each server executes processing in cooperation with a processing request from a user, and responds to the processing request. In this way, the reliability and responsiveness of the system can be improved by sharing the processing among the servers.

  In a multi-tier system such as the Web three-tier system, the server functions of each tier can be realized by a virtual machine (VM). The VM is a virtual computer realized by virtualizing resources such as a CPU (Central Processing Unit) and a storage device of the computer. By virtualizing computer resources, it becomes possible to simultaneously operate a plurality of VMs having different OSs (Operating Systems) on one physical computer. A computer system using such a virtual machine is called a virtual system.

  By the way, in a multi-tier system, when the response time of an end user increases, it is important as a first step in dealing with a failure to identify a server in which a problem has occurred. For this reason, a method of determining the presence or absence of a problem by measuring the processing time in the server of each hierarchy and monitoring the transition is widely adopted. For example, there is a technique for capturing a packet transmitted / received via a network and analyzing a processing time at a server in each layer based on the captured packet.

JP 2006-011683 A

  However, in the virtual system, a packet communicated between VMs operating on one computer cannot be captured because it does not pass through the switch. As a result, in the virtual system, communication information such as packets that can be captured by the switch is limited to a part of all packets communicated between VMs, and the accuracy of analysis using the captured communication information decreases. It was.

  In one aspect, the present invention provides a virtual machine management program capable of increasing the proportion of communication information that can be acquired on a network connecting servers that execute virtual machines, among virtual communication information between virtual machines, An object is to provide a machine management method and a virtual machine management apparatus.

  In one proposal, the computer refers to a storage unit in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the virtual path indicated in each communication path information is indicated. For a pair of machines, a server computer that executes each of the plurality of virtual machines is determined so that each of the two virtual machines that form the pair is executed by different server computers, and a server computer that executes each of the plurality of virtual machines is determined. A virtual machine management program is provided that outputs processing information and executes processing.

  In another plan, the computer refers to a storage unit in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and each of the communication path information For the pair of virtual machines shown, the server computer that executes each of the plurality of virtual machines is determined and each of the plurality of virtual machines is executed so that each of the two virtual machines of the pair is executed by different server computers A virtual machine management method is provided that outputs information indicating a server computer.

  In yet another alternative, the storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered is referred to, and virtual An execution server determination means for determining a server computer that executes each of the plurality of virtual machines, and each of the plurality of virtual machines so that each of the two virtual machines that form the pair is executed by different server computers. There is provided a virtual machine management device comprising output means for outputting information indicating a server computer to be executed.

  Of the communication information between virtual machines, it is possible to increase the proportion of communication information that can be acquired on a network connecting servers that execute virtual machines.

It is a figure which shows the system which concerns on 1st Embodiment. It is a figure which shows the system configuration example of 2nd Embodiment. It is a figure which shows the example of an internal structure of an operation rack and a visualization rack. It is a figure which shows the structural example of the hardware of a visualization apparatus. It is a figure which shows the outline | summary of a virtual system visualization process. It is a block diagram which shows the function of a visualization apparatus. It is a figure which shows the data structure example of an address memory | storage part. It is a figure which shows the example of a data structure of a path | route storage part. It is a flowchart which shows the procedure of a virtual system visualization process. It is a figure which shows the outline | summary of VM rearrangement for communication path | route detection. It is a figure which shows the outline | summary of the presence or absence investigation process of communication. It is a figure which shows the example of a data structure of VM arrangement | positioning list | wrist. It is a figure which shows the example of a data structure of a routing table. It is a flowchart which shows the procedure of a communication path | route detection process. It is a figure which shows an example of a communication path | route detection process. It is a figure which shows the outline | summary of the rearrangement process for visualization based on the path | route list input from the outside. It is a figure which shows the example of a data structure of VM arrangement | positioning list | wrist. It is a figure which shows the example of a data structure of the arrangement | positioning possible list | wrist. It is a flowchart which shows the procedure of the rearrangement process for visualization. It is a figure which shows the 1st example of arrangement | positioning pattern selection. It is a figure which shows the 2nd example of arrangement | positioning pattern selection. It is a figure which shows an example of a visualization process.

Hereinafter, the present embodiment will be described with reference to the drawings.
[First Embodiment]
FIG. 1 is a diagram illustrating a system according to the first embodiment. The virtual machine management device 1 manages virtual machines 7a, 7b, 7c, and 7d executed by a plurality of server computers 2 and 3. For example, the virtual machine management device 1 determines server computers that execute the virtual machines 7a, 7b, 7c, and 7d. The virtual machine arrangement device 6 causes the server computer determined by the virtual machine management device 1 to execute each of the virtual machines 7a, 7b, 7c, and 7d. In this way, causing the server computer to execute the virtual machine is sometimes referred to as arranging the virtual machine on the server computer.

  The server computers 2 and 3 are both connected to the switch 4. The server computers 2 and 3 communicate via a network using the switch 4. The switch 4 has a port mirroring function. The port mirroring function is a function of outputting (capturing) a copy of a packet transmitted / received via a predetermined port from a mirror port. The port mirroring function is sometimes referred to as port monitoring. The analyzer 5 is connected to the mirror port of the switch 4. The analysis device 5 analyzes the processing status by the virtual machines 7a, 7b, 7c, and 7d based on the communication information captured by the port mirroring of the switch 4.

  In a system having such a configuration, it is effective to capture more communication information in the switch 4 in order to improve the analysis accuracy of the analysis device 5. For example, in the pre-change arrangement of the virtual machines 7a, 7b, 7c, 7d, the virtual machine 7a and the virtual machine 7b are executed by the server computer 2, and the virtual machine 7c and the virtual machine 7d are executed by the server computer 3. Yes. Further, communication is performed between the virtual machine 7a and the virtual machine 7b, communication is performed between the virtual machine 7b and the virtual machine 7c, and communication is performed between the virtual machine 7c and the virtual machine 7d. . In this case, the communication information between the virtual machine 7b and the virtual machine 7c can be captured by the switch 4. However, the communication information between the virtual machine 7a and the virtual machine 7b is performed via the host OS in the server computer 2 and cannot be captured. Similarly, communication information between the virtual machine 7a and the virtual machine 7b cannot be captured. If it does so, the communication information which can be acquired with the analyzer 5 will become limited, and the analysis precision of the processing condition by virtual machine 7a, 7b, 7c, 7d will fall. For example, although the processing time of the cooperation processing between the virtual machine 7c and the virtual machine 7d can be obtained, how much time each of the virtual machine 7c and the virtual machine 7d executed within the processing time I can't ask.

Therefore, the virtual machine management apparatus 1 includes a storage unit 1a, an execution server determination unit 1b, and an output unit 1c as functions for improving analysis accuracy.
The storage means 1a stores at least one piece of communication path information 9a, 9b, 9c. The communication path information 9a, 9b, and 9c indicate a pair of two virtual machines that communicate with each other among the plurality of virtual machines 7a, 7b, 7c, and 7d. In the example of FIG. 1, three pieces of communication path information 9a, 9b, and 9c are stored. The communication path information 9a indicates a pair of a virtual machine 7a with an identification number “1” and a virtual machine 7b with an identification number “2”. The communication path information 9b indicates a pair of the virtual machine 7b with the identification number “2” and the virtual machine 7c with the identification number “3”. The communication path information 9c indicates a pair of the virtual machine 7c with the identification number “3” and the virtual machine 7d with the identification number “4”.

  The execution server determination unit 1b refers to the storage unit 1a and determines a server computer that executes each of the plurality of virtual machines 7a, 7b, 7c, and 7d. At this time, the execution server determination unit 1b performs the virtual machine 7a so that the two virtual machines forming the pair are executed by different server computers with respect to the virtual machine pairs indicated in the communication path information 9a, 9b, and 9c. , 7b, 7c, 7d are determined.

For example, the execution server determination unit 1b includes an arrangement pattern generation unit 1b-1, an availability determination unit 1b-2, a selection unit 1b-3, and a determination unit 1b-4.
The arrangement pattern generation means 1b-1 includes arrangement patterns 8a, 8b, which indicate combinations of a plurality of virtual machines and server computers that execute the virtual machines when a plurality of virtual machines are executed by a predetermined number or less of server computers. 8c,... Are generated. For example, the arrangement pattern generation unit 1b-1 generates a plurality of arrangement patterns in which server computers that execute at least some of the plurality of virtual machines 7a, 7b, 7c, and 7d are different from each other. At this time, if the maximum number of server computers to be used is designated in advance, for example, the arrangement pattern generation means 1b-1 increases the number of servers one by one from two servers to the maximum number of server computers to be used. An arrangement pattern is generated for each number of servers.

  The availability determining unit 1b-2 determines whether the generated arrangement patterns 8a, 8b, 8c,. For example, the availability determination unit 1b-2 can apply an arrangement pattern in which two virtual machines constituting a pair are executed by different server computers with respect to the virtual machine pairs indicated in all the communication path information 9a, 9b, 9c. Is determined.

  When there are a plurality of applicable arrangement patterns, the selection unit 1b-3 selects an arrangement pattern that meets a selection condition specified in advance. For example, when the selection condition places importance on performance, the selection unit 1b-3 selects an arrangement pattern having the largest number of server computers to be used among applicable arrangement patterns. In addition, when the selection condition is saving the number, the selection unit 1b-3 selects an arrangement pattern with the smallest number of server computers to be used among applicable arrangement patterns.

The determining unit 1b-4 determines a server computer that executes each of the plurality of virtual machines according to the applicable arrangement pattern selected by the selecting unit 1b-3.
The output unit 1c outputs information (arrangement information) indicating a server computer that executes each of the plurality of virtual machines 7a, 7b, 7c, and 7d determined by the execution server determination unit 1b. The placement information is output to the virtual machine placement device 6, for example.

  According to such a system, a server computer that executes the virtual machines 7a, 7b, 7c, and 7d is determined by the execution server determination unit 1b of the virtual machine management apparatus 1. At this time, for the virtual machine pairs indicated in the communication path information 9a, 9b, and 9c, the virtual machines 7a, 7b, 7c, and 7d are set so that the two virtual machines that form the pair are executed by different server computers. A server computer to be executed is determined. For example, it is determined that the server computer 2 executes the virtual machines 7a and 7c and the server computer 3 executes the virtual machines 7b and 7d. Arrangement information indicating the contents determined by the execution server decision unit 1b is output to the virtual machine arrangement apparatus 6 by the output unit 1c.

  The virtual machine placement device 6 rearranges the virtual machines 7a, 7b, 7c, and 7d according to the placement information. In the example of FIG. 1, the virtual machine placement apparatus 6 instructs the server computer 2 to move (migrate) the virtual machine 7 b to the server computer 3. Then, the server computer 2 stops the execution of the virtual machine 7b and transfers a program and data necessary for the execution of the virtual machine 7b to the server computer 3. The server computer 3 executes the virtual machine 7b based on the program and data acquired from the server computer 2. Similarly, the virtual machine placement apparatus 6 instructs the server computer 3 to move the virtual machine 7c to the server computer 2. As a result, the virtual machines 7 a and 7 c are executed by the server computer 2, and the virtual machines 7 b and 7 d are executed by the server computer 3.

  Then, all communication between the virtual machines 7a, 7b, 7c, and 7d is performed via the switch 4. That is, all communication information passes through the switch 4. As a result, all communication information between the virtual machines 7 a, 7 b, 7 c and 7 d is captured by the switch 4 and input to the analyzer 5. As a result, the analysis device 5 can analyze the processing status of the virtual machines 7a, 7b, 7c, and 7d with high accuracy. For example, when a single transaction is executed by the virtual machines 7a, 7b, 7c, and 7d in cooperation, it is possible to analyze how much time each virtual machine 7a, 7b, 7c, and 7d has spent on processing.

  As described above, according to the system according to the first embodiment, the ratio of the communication information that can be acquired on the network that connects the server computers that execute the virtual machines is increased among the communication information between the virtual machines. be able to. As a result, it is possible to improve the analysis accuracy of the processing status of the virtual machine using the communication information.

  In the example of FIG. 1, the virtual machines 7a, 7b, 7c, and 7d are rearranged so that all the communication information between the virtual machines 7a, 7b, 7c, and 7d can be captured. Information may not need to be captured. For example, among transactions executed in cooperation with the virtual machines 7a, 7b, 7c, and 7d, there is a case where it is desired to know the processing time for the processing executed with the virtual machine 7c and the virtual machine 7d. In that case, the virtual machine 7b and the virtual machine 7c may be executed by different server computers, and the virtual machine 7c and the virtual machine 7d may be executed by different server computers.

[Second Embodiment]
In the second embodiment, a server group for system visualization is prepared separately from the operation server group, and packets are captured using the server group for system visualization. In the following description, the virtual server is abbreviated as “VM”. A server computer that executes a VM is called a physical server. Furthermore, analyzing the processing status by a plurality of VMs is called “visualization”.

  In principle, all communication paths can be captured if the communicating VMs are arranged in different physical servers. For example, as shown in FIG. 1, if VMs that generate communication are arranged in different physical servers, all communication can be visualized. However, fixing with this arrangement at all times is not efficient because flexible operation, which is an advantage of the virtual system, is not possible. On the other hand, in a virtual system, it is possible to dynamically arrange VMs using live migration technology. Live migration is a technique for moving a VM running on one physical server to another physical server without stopping. Therefore, in the second embodiment, when a problem occurs in the system, a physical server group used for investigation (visualization) of whether or not the system functions are operating as designed is provided. Then, a VM that realizes a service to be visualized is arranged in a physical server group used for system visualization, and system visualization is performed.

  FIG. 2 is a diagram illustrating a system configuration example according to the second embodiment. In the second embodiment, an operation rack A and a visualization rack B are provided. The operation rack A has a built-in physical server in which a VM that realizes a service not targeted for visualization is arranged. The visualization rack B has a built-in physical server in which a VM that realizes a service to be visualized is arranged. The operation rack A and the visualization rack B are connected to the network 10.

  Further, a VM placement device 11 and a plurality of terminal devices 21, 22,... Are connected to the network 10. The VM placement device 11 is a computer that instructs the placement of VMs to physical servers built in the operation rack A and the visualization rack B, respectively. Each terminal device 21, 22,... Is a computer used by a user who uses a service provided by a VM arranged in a physical server in the operation rack A or the visualization rack B.

  FIG. 3 is a diagram illustrating an example of the internal structure of the operation rack and the visualization rack. In the operation rack A, a plurality of physical servers 31, 32, 33,... And a switch 51 are mounted. Each of the physical servers 31, 32, 33,... Is an individual computer. Each physical server 31, 32, 33,... Is connected to a switch 51. The physical servers 31, 32, 33,... Communicate with each other via the switch 51. The switch 51 is also connected to the network 10.

  In the visualization rack B, a plurality of physical servers 41, 42, 43,..., A switch 52, and a visualization device 100 are mounted. Each of the physical servers 41, 42, 43,... Is an individual computer. Each physical server 41, 42, 43,... Is connected to a switch 52. The physical servers 41, 42, 43,... Communicate with each other via the switch 52.

  The switch 52 is a switching hub having a port mirroring function. The physical servers 41, 42, 43,... Are connected to normal ports other than the mirror port 52a of the switch 52. The switch 52 is further connected to the network 10 via a port other than the mirror port 52a. The visualization device 100 is connected to the mirror port 52 a of the switch 52.

  The visualization apparatus 100 receives a packet output from the mirror port, and analyzes a message transmitted / received by a VM arranged in the physical servers 41, 42, 43,. Further, the visualization apparatus 100 can appropriately determine the arrangement of VMs. For example, the visualization apparatus 100 detects a pair of VMs that are communicating with each other based on the received packet, and determines the placement of the VM so that the ratio of packets that can be captured is high. When determining the VM arrangement, the visualization apparatus 100 transmits arrangement information indicating the determined VM arrangement to the VM arrangement apparatus 11 via the switch 52.

  FIG. 4 is a diagram illustrating a hardware configuration example of the visualization apparatus. The entire visualization apparatus 100 is controlled by the CPU 101. A random access memory (RAM) 102 and a plurality of peripheral devices are connected to the CPU 101 via a bus 106.

  The RAM 102 is used as a main storage device of the visualization device 100. The RAM 102 temporarily stores at least a part of OS programs and application programs to be executed by the CPU 101. The RAM 102 stores various data necessary for processing by the CPU 101.

Peripheral devices connected to the bus 106 include a hard disk drive (HDD) 103 and a plurality of communication interfaces 104 and 105.
The HDD 103 magnetically writes and reads data to and from the built-in disk. The HDD 103 is used as a secondary storage device of the visualization device 100. The HDD 103 stores an OS program, application programs, and various data. Note that a semiconductor storage device such as a flash memory can also be used as the secondary storage device.

  The communication interface 104 is connected to the normal communication port 52 b of the switch 52. The communication interface 104 transmits / receives data to / from other communication devices such as the VM placement device 11 via the switch 52.

  The communication interface 105 is connected to the mirror port 52 a of the switch 52. The communication interface 105 receives a packet output from the mirror port 52 a of the switch 52 and transfers it to the CPU 101. Even if the destination of the received packet is an apparatus other than the visualization apparatus 100, the communication interface 105 acquires the packet without discarding it.

  The visualization device 100 may further include a graphic processing device, an input interface, an optical drive device, and the like. For example, when the visualization apparatus 100 has a graphic processing apparatus, a monitor is connected to the graphic processing apparatus. Then, the graphic processing apparatus displays an image on the monitor screen in accordance with a command from the CPU 101. Examples of the monitor include a display device using a CRT (Cathode Ray Tube) and a liquid crystal display device.

  When the visualization apparatus 100 has an input interface, for example, a keyboard and a mouse are connected to the input interface. In this case, the input interface transmits a signal sent from the keyboard or mouse to the CPU 101. Note that the mouse is an example of a pointing device, and other pointing devices can be used. Examples of other pointing devices include a touch panel, a tablet, a touch pad, and a trackball.

  Further, when the visualization apparatus 100 has an optical drive device, the optical drive device reads data recorded on the optical disk using a laser beam or the like. An optical disc is a portable recording medium on which data is recorded so that it can be read by reflection of light. Optical disks include DVD (Digital Versatile Disc), DVD-RAM, CD-ROM (Compact Disc Read Only Memory), CD-R (Recordable) / RW (ReWritable), and the like.

  With the hardware configuration as described above, the processing functions of the present embodiment can be realized. 4 shows an example of the hardware configuration of the visualization apparatus 100. However, the physical servers 31, 32, 33,... In the operation rack A and the physical servers 41, 42, 43 in the visualization rack B are shown. ,..., The VM placement device 11 can be realized by similar hardware. The terminal devices 21, 22,... Have a graphic processing device and an input interface in addition to the hardware configuration shown in FIG. The terminal devices 21, 22,... Receive user operation inputs to input devices such as a keyboard and a mouse connected to the input interface via the input interface. Further, the terminal devices 21, 22,... Display various information such as processing results by the VM on a monitor connected to the graphic processing device.

Next, an overview of virtual system visualization processing in the second embodiment will be described.
FIG. 5 is a diagram showing an outline of the virtual system visualization process. In the example of FIG. 5, VMs 81 to 86 are executed on the physical servers 31, 32, 33,. Each VM 81 to 86 is set with an identifier (VM number) that can be uniquely identified in the system. The VM number of the VM 81 is “1”. The VM number of the VM 82 is “2”. The VM number of the VM 83 is “3”. The VM number of the VM 84 is “4”. The VM number of the VM 85 is “5”. The VM number of the VM 86 is “6”.

  The operations of the physical servers 31, 32, 33,... Are controlled by the host OSs 31a, 32a, 33a,. Similarly, the operations of the physical servers 41, 42, 43,... In the visualization rack B are controlled by the host OSs 41a, 42a, 43a,. In the virtual system, the VM arrangement can be dynamically changed.

  VMs arranged on different physical servers communicate via a switch that connects the physical servers. For example, the VM 82 in the physical server 31 and the VM 83 in the physical server 32 perform communication via the switch 51. On the other hand, VMs arranged on the same physical server communicate with each other via the host OS of the physical server. For example, the VMs 81 and 82 in the physical server 31 communicate via the host OS 31a.

  In the second embodiment, among VMs executed on the physical servers 31, 32, 33,... In the operation rack A, a VM that executes a service to be visualized (for example, a Web service based on a Web three-tier system). Only relocation is done. In the example of FIG. 5, it is assumed that services realized by the four VMs 81 to 84 are objects to be visualized.

  When the VMs 81 to 84 are rearranged, it is possible to capture all communication paths if the communicating VM pairs are arranged in different physical servers. If the switch 51 is provided with a port mirroring function and a visualization device is mounted in the operation rack A, VMs can be rearranged in the operation rack A and virtual system visualization can be realized. However, it is not efficient to fix the VM that executes the service to be visualized in the operation rack A with the arrangement for visualization. Further, in the physical server in the operation rack A, VMs 85 and 86 that are not related to the visualization target service are also operating. When a packet is captured in the operation rack A, a packet related to communication between the VM 85 and the VM 86 is also captured. If there are too many packets to be captured, the visualization apparatus 100 may not be able to receive the packets in time, and packets that cannot be received may occur. Therefore, in the second embodiment, VMs that execute the services to be visualized are arranged on the physical servers 41, 42, 43,... In the visualization rack B. As a result, it is possible to prevent the packet to be captured from being missed and to prevent adverse effects on other services being operated by executing the virtual system visualization process.

  As described above, in the second embodiment, dedicated physical servers 41, 42, 43,... Are provided in the visualization rack B for investigation when a problem occurs in the system. Then, the VMs 81 to 84 that execute the investigation target service are relocated to the physical servers 41, 42, and 43.

  When rearranging the VMs 81 to 84, the visualization apparatus 100 determines the placement destination of the VMs 81 to 84 so that the VMs communicating with each other are placed on different physical servers. In the example of FIG. 5, it is determined that the VM 81 is arranged on the physical server 41, the VM 82 is arranged on the physical server 42, the VM 83 is arranged on the physical server 41, and the VM 84 is arranged on the physical server 42. Then, based on the placement information indicating the placement determined by the visualization device 100, the VM placement device 11 executes a rearrangement process.

  In the VMs 81 to 84 rearranged in the physical servers 41 and 42 in the visualization rack B, all communication between the VMs is performed via the switch 52. Accordingly, all communication-related packets are captured by the switch 52 and input to the visualization apparatus 100. As a result, in the visualization apparatus 100, it is possible to analyze the coordinated processing by the plurality of VMs 81 to 84 with high accuracy.

Next, functions of the visualization apparatus 100 for realizing the virtual system visualization process illustrated in FIG. 5 will be described.
FIG. 6 is a block diagram illustrating functions of the visualization apparatus. The visualization apparatus 100 includes an analysis unit 110, an address storage unit 120, a route detection unit 130, a route storage unit 140, and an arrangement determination unit 150.

  Based on the packet captured from the switch 52, the analysis unit 110 analyzes the processing status of a process that takes a multi-layer system configuration by a plurality of VMs. For example, the analysis unit 110 analyzes the time required for processing by each VM when executing one transaction. For example, the analysis unit 110 transmits the analysis result to a terminal device used by the system administrator.

  The address storage unit 120 stores an address for identifying each VM. In the second embodiment, the address storage unit 120 stores the IP (Internet Protocol) address of each VM. For example, a part of the storage area of the RAM 102 or the HDD 103 is used as the address storage unit 120.

  The path detection unit 130 detects a communication path between VMs that are communicating based on the packet captured from the switch 52. For example, the path detection unit 130 determines whether communication has occurred for a VM pair obtained by selecting two VMs from VMs that execute the analysis target service. The presence or absence of communication in the VM pair can be determined by analyzing the header information of the packet. For example, the path detection unit 130 refers to the address storage unit 120 and extracts the addresses of the transmission source and the destination from the header information of the communicated packet. Then, the path detection unit 130 determines that communication has occurred between the VM pair corresponding to the extracted address.

  Further, the path detection unit 130 generates the arrangement information 61 such that the VM pair for which the presence / absence of communication is not determined is arranged on a different physical server. Then, the path detection unit 130 transmits the generated arrangement information 61 to the VM arrangement apparatus 11. The VM placement apparatus 11 places VMs that execute the services to be visualized on the physical servers 41, 42, 43,... In the visualization rack B according to the placement information 61. Thereafter, the placed VM is loaded. For example, a request message from the terminal device is input to a VM arranged in the physical servers 41, 42, 43,. Then, in response to the request message, an operation linked with a plurality of VMs is executed, and the packet used for the VM linked operation is captured by the switch 52. Then, the path detection unit 130 can determine the presence / absence of communication for a VM pair that has not determined the presence / absence of communication based on the captured packet. Note that the communication path is represented by, for example, the VM number of a pair of VMs that performed communication. The route detection unit 130 stores communication route information indicating the detected communication route in the route storage unit 140.

The path storage unit 140 stores communication path information indicating a pair of VMs that perform communication. For example, a part of the storage area of the RAM 102 or the HDD 103 is used as the path storage unit 140.
The placement determination unit 150 refers to the path storage unit 140 and determines the placement of each VM such that communication between as many VMs as possible is performed via the switch 52. For example, the arrangement determining unit 150 determines to arrange two VM pairs that are performing communication on different physical servers as much as possible. The placement determining unit 150 transmits placement information 62 indicating the determined placement to the VM placement device 11. The VM placement apparatus 11 places VMs that execute the services to be visualized on the physical servers 41, 42, 43,... In the visualization rack B according to the placement information 62. Thereafter, the placed VM is loaded. For example, a request message from the terminal device is input to a VM arranged in the physical servers 41, 42, 43,. Then, according to the request message, an operation in which a plurality of VMs cooperate with each other is executed, and a packet used by the VM for cooperation is captured by the switch 52. When the analysis unit 110 analyzes the captured packet at this time, for example, the processing time per transaction of the VM that executes the service to be visualized is calculated.

Next, information stored in the visualization device 100 will be described.
FIG. 7 is a diagram illustrating an example of a data structure of the address storage unit. The address storage unit 120 stores an address correspondence table 121. In the address correspondence table 121, a set of a VM number and an IP address is set for each VM. By referring to such an address correspondence table 121, it is possible to identify the transmission source VM and the destination VM of the captured packet.

  FIG. 8 is a diagram illustrating an example of the data structure of the path storage unit. The route storage unit 140 stores a route list 141. In the route list 141, pairs of VM numbers of two VMs that communicate with each other are registered as communication route information.

  The visualization apparatus 100 configured as shown in FIG. 6 to FIG. 8 makes it possible to collect communication packets between VMs that execute services to be visualized with a high probability in a multi-tier system using a virtual system. The processing status can be analyzed.

FIG. 9 is a flowchart illustrating a procedure of virtual system visualization processing. In the following, the process illustrated in FIG. 9 will be described in order of step number.
[Step S11] The route detection unit 130 detects a communication route between VMs that execute the services to be visualized. For example, the path detection unit 130 repeats the rearrangement of VMs, and checks the presence / absence of communication between VMs every time the VMs are rearranged. Thereby, the communication path between VMs is automatically detected. The detected communication path is registered in the path list 141 in the path storage unit 140 as communication path information.

  [Step S12] The placement determining unit 150 determines the placement of the VMs for visualization so that the VM pairs indicated by the communication route information in the route list 141 are placed on different physical servers. Then, the placement determination unit 150 transmits placement information indicating the determined placement to the VM placement device 11. As a result, VMs that execute services are rearranged on the physical servers 41, 42, 43,... In the visualization rack B.

  [Step S13] The analysis unit 110 visualizes the operation status of the rearranged VM. That is, the analysis unit 110 applies a load to the rearranged VM and captures packets input / output to / from the VM, thereby analyzing the service processing status. And the analysis part 110 transmits an analysis result to the terminal device which an administrator uses.

  Note that the communication path detection process shown in step S11 is unnecessary when the communication path between VMs is known in advance. When the communication path between VMs is known in advance, for example, the administrator inputs an operation input of communication path information to the terminal device. The route detection unit 130 acquires communication route information from the terminal device used by the administrator and stores it in the route storage unit 140.

Hereinafter, each process shown in FIG. 9 will be described in detail.
<Communication path detection processing>
In the route detection process, one or more VMs are rearranged. The number of times VM rearrangement is performed depends on the number of physical servers in the visualization rack B and the number of VMs that execute the services to be visualized. The physical server number of each physical server in the visualization rack B and the VM number of each VM that executes the service to be visualized are set in advance in the visualization device 100 by, for example, an operation input by an administrator using a terminal device. Is done.

  FIG. 10 is a diagram showing an outline of VM rearrangement for communication path detection. In the example of FIG. 10, it is assumed that M physical servers 41, 42, 43,..., 4m are mounted in the visualization rack B (M is an integer of 2 or more). In addition, it is assumed that n (n is an integer of 2 or more) VMs 91, 92, 93,.

  It is assumed that the number M of physical servers in the visualization rack B and the number n of VMs that execute the services to be visualized are set in advance in the path detection unit 130 before the virtual system visualization process is started. For example, when the system administrator inputs an operation using the terminal device, the number M of physical servers in the visualization rack B and the number n of VMs that execute the services to be visualized via the terminal device are determined by the route detection unit 130. Set to

  The path detection unit 130 arranges the VMs 91, 92, 93,..., 9n on the physical servers 41, 42, 43,. And the path | route detection part 130 detects a communication path | route by capturing the communication between VM91,92,93, ..., 9n. In order to detect all communication paths between VMs, the path detection unit 130 places two VMs that form a pair on different physical servers at least once for all pairs of VMs.

  For example, in the first VM rearrangement, the path detection unit 130 equally distributes n VMs 91, 92, 93,..., 9n to M physical servers 41, 42, 43,. Deploy. If the number of VMs (n) ≦ the number of physical servers (M), all the VMs 91, 92, 93,..., 9n can be arranged in different physical servers. In this case, all paths can be detected by one VM rearrangement.

  On the other hand, when the number of VMs (n)> the number of physical servers (M), the investigation needs to be performed in multiple times. In the first investigation, the VMs are evenly arranged on the physical servers to detect the path. In the first investigation, the path of the VM arranged on the same physical server cannot be detected. Thus, in the second investigation, the path detection unit 130 arranges the VMs arranged on the same physical server in the first rearrangement to be different physical servers.

When “the number of VMs (n) ≦ the square of the number of physical servers (M ² )”, all routes can be detected by two investigations. If there are more, perform the third and subsequent surveys in the same way. If all VM pairs are examined and all paths are detected, the process is complete. For example, if there are 10 physical servers, it is possible to investigate 100 VMs in 2 investigations and 1000 VMs in 3 investigations.

  In this way, by performing VM relocation, all of the VM pairs can be allocated to different physical servers at least once. Then, each time a VM is rearranged, the VM can execute the process related to the service to be visualized, whereby the presence / absence of communication can be investigated for each pair of VMs.

  FIG. 11 is a diagram showing an outline of the communication presence / absence investigation process. In FIG. 11, VMs arranged in each physical server are represented by circles. In the circle, the VM number of the VM corresponding to the circle is shown. Similarly, in FIGS. 15, 16, 20, 21, and 22, the VM is represented by a circle having a VM number inside.

  The path detection unit 130 of the visualization apparatus 100 includes a VM arrangement list 131 and a path table 132. The VM placement list 131 and the route table 132 are created by the route detection unit 130. Further, the VM arrangement list 131 and the path table 132 are stored in a storage area managed by the path detection unit 130 among the storage areas in the RAM 102, for example.

  In the VM arrangement list 131, VM arrangement patterns are registered. The VM placement pattern is information indicating on which physical server each VM to be rearranged is placed. In the VM arrangement list 131, for each VM arrangement pattern, the number of paths that can be newly investigated for the presence or absence of communication in the corresponding arrangement pattern is set. The path detection unit 130 refers to, for example, the VM arrangement list 131 and preferentially selects an arrangement pattern having a large number of paths that can be investigated, and transmits arrangement information 61 corresponding to the selected arrangement pattern to the VM arrangement apparatus 11. To do.

  In the route table 132, for each pair of VMs, an investigation status as to whether there is communication on the communication route between the corresponding VMs is set. The route detection unit 130 can determine an uninvestigated pair among the VM pairs with reference to the route table 132.

  In the example of FIG. 11, VMs with VM numbers “1”, “2”, and “3” are allocated to the physical server 41 with physical server number “1”, and the physical server with physical server number “2” is allocated. Thus, the arrangement information 61 indicating that the VMs having the VM numbers “4”, “5”, and “6” are arranged is generated. The placement information 61 is sent from the path detection unit 130 to the VM placement device 11. Then, the VM placement apparatus 11 places the VM on the physical servers 41 and 42 according to the placement information 61.

  Thereafter, the arranged VM is caused to execute the service to be visualized. Then, the packet 63 accompanying the cooperation process between VMs is communicated via the switch 52. For example, a destination IP address and a source IP address are set in the header portion of the packet 63.

  The packet 63 is captured by the switch 52 and passed to the route detection unit 130 of the visualization device 100. The path detection unit 130 acquires the VM number corresponding to the IP address of the destination and transmission source of the acquired packet 63 from the address storage unit 120. In the example of FIG. 11, the destination IP address is “192.168.0.4”, and the corresponding VM number is “4”. The source IP address is “192.168.0.1”, and the corresponding VM number is “1”.

  The path detection unit 130 recognizes that communication has been performed between the destination VM of the captured packet 63 and the transmission source VM, and generates communication path information 65 indicating the VM number pair of the VM. . Then, the route detection unit 130 registers the generated communication route information 65 in the route list 141 in the route storage unit 140.

In this way, the presence or absence of communication for each communication path can be investigated.
Hereinafter, the data structure of the VM arrangement list 131 and the route table 132 will be described.
FIG. 12 is a diagram illustrating a data structure example of the VM arrangement list. A plurality of arrangement patterns are registered in the VM arrangement list 131. An arrangement pattern number is assigned to the arrangement pattern. When n VMs are allocated to M physical servers, “M raised to the nth power (M ⁿ )” layout patterns are generated.

  In the VM placement list 131, the physical server number of the physical server that is the placement destination of the VM is set in association with the VM number of each VM for each placement pattern. In the VM arrangement list 131, the number of surveyable paths is set for each arrangement pattern. The number of searchable paths is the number of unsearched communication paths that can be checked by the corresponding arrangement pattern. The number of surveyable paths is updated every time the presence / absence of communication according to one arrangement pattern is investigated.

FIG. 13 is a diagram illustrating an example of the data structure of the routing table. In the route table 132, a check flag for each VM pair is set. Investigation flags corresponding to the number of brute force combinations ( _n C ₂ ways) for selecting two VMs that form a pair from n VMs are set. The VM number in the row direction and the VM number in the column direction of each cell in the routing table 123 indicates a VM pair corresponding to the investigation flag set in the cell.

  The investigation flag takes one of the values “0”, “1”, and “2”. The investigation flag “0” indicates that the presence or absence of communication has not been investigated for the corresponding VM pair. The investigation flag “1” indicates that the investigation of the presence / absence of communication is completed for the corresponding VM pair, and it is determined that the route is a communication path (communication detection path). The investigation flag “2” indicates that the investigation of the presence / absence of communication has been completed for the corresponding VM pair, and it has been determined that the route is a route where communication is not performed (communication non-detection route).

Next, the procedure of the communication path detection process will be described.
FIG. 14 is a flowchart illustrating a procedure of communication path detection processing. In the following, the process illustrated in FIG. 14 will be described in order of step number.

  [Step S <b> 21] The path detection unit 130 generates all arrangement patterns for assigning VMs to physical servers, and creates a VM arrangement list 131 in which the generated arrangement patterns are registered. At this time, the field of the number of searchable paths in the VM arrangement list 131 is blank.

  [Step S22] The route detection unit 130 creates a route table 132 indicating brute force combinations between VMs. At this time, the route detection unit 130 initializes all the investigation flags in the route table 132 to “0”.

  [Step S23] The path detection unit 130, for all the arrangement patterns in the VM arrangement list 131, is the number of two VMs corresponding to the paths having the investigation flag “0” in the route table 132 arranged in different physical servers (investigable) Path number). Then, the path detection unit 130 sets the number of paths that can be investigated for each layout pattern in the VM layout list 131.

  [Step S24] The route detection unit 130 arranges VMs according to an arrangement pattern having the maximum number of routes that can be investigated. Specifically, the path detection unit 130 selects an arrangement pattern having the maximum number of searchable paths from the arrangement patterns in the VM arrangement list 131. In addition, when there are a plurality of arrangement patterns having the maximum number of surveyable routes, the route detection unit 130 selects one of them. The path detection unit 130 generates arrangement information according to the selected arrangement pattern and transmits it to the VM arrangement apparatus 11. Then, the VM placement apparatus 11 rearranges the VM according to the placement information.

  [Step S25] The path detection unit 130 causes the arranged VMs to execute processing. For example, if the system is in operation, the request message from the terminal device for the VM in the highest hierarchy in the multi-tier system is arranged in the physical servers 41, 42, 43,... In the visualization rack B. Input to VM. Then, transaction processing is executed by a plurality of VMs according to the request message. Between VMs that execute transaction processing, processing is called by a request message and a processing result is responded by a response message. Request messages and response messages are communicated by, for example, IP packets. For example, the route detection unit 130 proceeds to step S26 after a predetermined time has elapsed.

  [Step S26] The path detection unit 130 analyzes a packet captured during a period in which the VM arranged in Step S25 executes processing, and determines a communication path used for communication. Specifically, the route detection unit 130 extracts a set of a destination IP address and a transmission source IP address from the captured packet. Next, the path detection unit 130 refers to the address correspondence table 121 in the address storage unit 120, identifies a VM number pair corresponding to the extracted set of IP addresses, and sets it as a communication detection path.

  [Step S27] The route detection unit 130 updates the investigation flag in the route table 132. Specifically, the route detection unit 130 sets “1” to the investigation flag corresponding to the VM number pair determined as the communication detection route in step S26 among the routes newly investigated by the arrangement pattern selected in step S23. Set. In addition, the route detection unit 130 sets “2” in the investigation flag corresponding to the VM number pair that has not been determined as the communication detection route in step S26 among the routes newly investigated by the arrangement pattern selected in step S23. To do.

  [Step S28] The path detection unit 130 determines whether there is an uninvestigated communication path. If there is a communication route (a pair of VM numbers) with the investigation flag “0” in the route table 132, it is determined that there is an uninvestigated communication route. If there is an uninvestigated communication path, the path detection unit 130 proceeds with the process to step S23. In addition, the route detection unit 130 advances the process to step S29 when all the communication routes have been investigated.

  [Step S29] The route detection unit 130 outputs the communication route information of the communication detection route to the route list 141 in the route storage unit 140. Specifically, the route detection unit 130 refers to the route table 132, extracts the VM number pair of the communication route whose investigation flag is “1”, and generates communication route information. Then, the route detection unit 130 registers the generated communication route information in the route list 141.

In this way, the route list 141 indicating the communication route through which communication has been performed is generated.
FIG. 15 is a diagram illustrating an example of the communication path detection process. The example of FIG. 15 is an example of communication path detection in the case of two physical servers and six VMs. There are 15 combinations that select two VMs that form a pair from six VMs. In FIG. 15, among the investigated communication paths, communication paths in which communication is detected are indicated by solid lines, and communication paths in which communication is not detected are indicated by wavy lines.

  In the first VM placement, VMs with VM numbers “1”, “2”, and “3” are placed on the physical server 41, and VMs with VM numbers “4”, “5”, and “6” are physical servers. 42. In the first arrangement, nine communication paths are investigated, and communication is detected in six communication paths.

  In the second VM placement, VMs with VM numbers “1”, “2”, “4”, and “5” are placed on the physical server 41, and VMs with VM numbers “3” and “6” are physical servers. 42. In the second arrangement, four communication paths are investigated, and communication is detected on two of the communication paths.

  In the third VM placement, VMs with VM numbers “1” and “4” are placed on the physical server 41, and VMs with VM numbers “2” and “5” are placed on the physical server 42. In the third arrangement, two communication paths are investigated, and no communication is detected in all the investigated communication paths. In the third VM placement, VMs with VM numbers “3” and “6” may be placed on either physical server.

In this way, 15 combinations can be investigated in three investigations, and all communication paths 66 can be detected.
<Relocation for visualization>
Next, the visualization rearrangement process will be described. The visualization rearrangement process may be executed based on the route list 141 calculated by the route detection unit 130 or may be executed based on the route list input from the outside.

  FIG. 16 is a diagram showing an overview of the visualization rearrangement process based on a route list input from the outside. A route list 142 is input to the visualization device 100. For example, the route list 142 is transmitted from the terminal device used by the administrator to the visualization device 100. The input route list 142 is stored in the route storage unit 140, for example. Thereafter, the visualization apparatus 100 calculates how to arrange the VMs on the physical servers 41, 42, 43,... Of the visualization rack B. Specifically, the arrangement determining unit 150 of the visualization device 100 selects an arrangement pattern in which VM pairs of communication detection paths indicated in the path list 142 are arranged on different physical servers as much as possible. At that time, the placement determining unit 150 generates VM placement lists 151a, 151b, 151c,... For each number of placement destination physical servers. In the VM placement lists 151a, 151b, 151c,..., VM placement patterns when VMs are rearranged on a predetermined number of physical servers are set. The placement determination unit 150 is a physical server in which all of the VM pairs of the communication detection paths indicated in the path list 142 are different from each other among the placement patterns set in the VM placement lists 151a, 151b, 151c,. The arrangement pattern arranged in is extracted. Next, the arrangement determining unit 150 registers the extracted arrangement pattern in the arrangement possible list 152. Then, the arrangement determining unit 150 selects an arrangement pattern from the arrangement patterns registered in the arrangement possible list 152 according to a selection condition specified in advance. In the selection condition designated in advance, for example, it is designated whether to select an arrangement pattern that emphasizes performance or to select an arrangement pattern that saves the number.

  The placement determination unit 150 sends the placement information of the selected placement pattern to the VM placement device 11. Then, the VM placement apparatus 11 executes the rearrangement by moving the VM from the physical servers 31, 32, 33 in the operation rack A to the physical servers 41, 42, 43 in the visualization rack B.

  Although FIG. 16 shows the visualization rearrangement process based on the route list input from the outside, the visualization rearrangement process can also be performed based on the route list 141 generated by the route detection unit 130. In that case, when the communication path detection processing by the path detection unit 130, the VM that executes the service to be visualized is already arranged in the physical servers 41, 42, 43,... In the visualization rack B. Therefore, the rearrangement instruction by the VM placement device 11 is issued to the physical servers 41, 42, 43,... In the visualization rack B.

Further, the VM arrangement lists 151a, 151b, 151c,... And the arrangement possible list 152 are stored in a storage area in the RAM 102 managed by the arrangement determining unit 150, for example.
FIG. 17 is a diagram illustrating an example of a data structure of the VM arrangement list. In the placement determination unit 150, VM placement lists 151a,..., 151l are generated for each number of physical servers used as VM placement destinations. In the VM placement list 151a,..., 151l, a placement pattern for placing a VM on the corresponding number of physical servers is set. For example, in the VM arrangement list 151a, 2 ⁿ (2 ⁿ ) arrangement patterns when the number of physical servers is two are set. In the VM arrangement list 151l, arrangement patterns of M to the nth power (M ⁿ ) when the number of physical servers is M are set. The arrangement pattern is an array of physical server numbers corresponding to the VM numbers of the VMs to be arranged.

  In addition, determination results for each arrangement pattern are set in each VM arrangement list 151a,. The determination result is “possible” or “impossible”. With respect to all VM pairs shown in the path list, an arrangement pattern in which VMs constituting a pair are arranged in different physical servers is determined to be arrangeable, and a determination result “OK” is set. On the other hand, with respect to at least one of the VM pairs shown in the path list, an arrangement pattern in which the VMs forming a pair are arranged in the same physical server is determined to be impossible, and the determination result “impossible” is set.

  FIG. 18 is a diagram illustrating an example of the data structure of the arrangement possible list. In the arrangement possible list 152, the arrangement pattern number of the arrangement pattern whose determination result is “possible” in the VM arrangement list among the arrangement patterns of the number of physical servers is set in association with the number of physical servers. .

Next, the procedure of the visualization rearrangement process will be described.
FIG. 19 is a flowchart illustrating a procedure of the visualization rearrangement process. In the following, the process illustrated in FIG. 19 will be described in order of step number.

  [Step S31] The placement determination unit 150 creates a VM placement list for each number of physical servers. For example, the placement determination unit 150 creates a VM placement list when n VMs are assigned to 2 to M physical servers. The placement determination unit 150 stores the created VM placement list in, for example, the RAM 102.

  [Step S32] The placement determination unit 150 determines whether or not each placement pattern registered in each of the created VM placement lists can be placed. The determination of whether or not placement is possible is a determination of whether or not the condition that two VMs at both ends of a communication path are placed on different physical servers is satisfied for all the communication detection paths shown in the path list. . The placement determination unit 150 sets the determination result of whether or not placement is possible in the VM placement list. For example, the arrangement determination unit 150 sets the determination result “permitted” as the arrangement pattern determined to be available for arrangement. Further, the arrangement determining unit 150 sets a determination result “impossible” for the arrangement pattern determined to be impossible to arrange.

  [Step S <b> 33] The arrangement determining unit 150 lists the arrangement patterns whose arrangement result is “OK”, and generates the arrangement possible list 152. Specifically, the placement determination unit 150 selects VM placement lists one by one. Next, the arrangement determining unit 150 acquires an arrangement pattern number of an arrangement pattern whose evaluation result is “OK” among the arrangement patterns registered in the selected VM arrangement list. Then, the arrangement determining unit 150 registers the arrangement pattern number of the arrangement pattern having the evaluation result “OK” in the arrangement possible list 152 in association with the number of physical servers corresponding to the selected VM arrangement list.

  [Step S34] The arrangement determining unit 150 determines whether or not there is at least one arrangement pattern having the determination result “OK”. If there is an arrangement pattern with the determination result “OK”, the arrangement determining unit 150 advances the process to step S36. If there is no arrangement pattern with the determination result “OK”, the arrangement determining unit 150 advances the process to step S35.

  [Step S35] The placement determining unit 150 deletes some of the communication paths in the path list. For example, the arrangement determining unit 150 transmits the route list to the terminal device used by the administrator. The administrator deletes a part of the communication route registered in the route list using the terminal device. The route list updated by the administrator is transmitted from the terminal device to the arrangement determining unit 150. The arrangement determining unit 150 updates the route list in the route storage unit 140 with the route list received from the terminal device. After deleting some communication paths in the path list, the placement determining unit 150 advances the process to step S32.

  [Step S36] When there is an arrangement pattern with the determination result “OK”, the arrangement determining unit 150 determines an arrangement pattern selection condition. Selection conditions are set in advance. If the selection condition is “number saving”, the arrangement determining unit 150 advances the process to step S37. If the selection condition is “performance-oriented”, the arrangement determining unit 150 advances the processing to step S38.

  [Step S37] When the selection condition is “number saving”, the arrangement determining unit 150 selects an arrangement pattern having the smallest number of physical servers among the arrangement patterns registered in the arrangement available list 152. When there are a plurality of arrangement patterns with the smallest number of physical servers, the arrangement determining unit 150 selects one of them. Thereafter, the arrangement determining unit 150 proceeds with the process to step S39.

  [Step S38] When the selection condition is “performance-oriented”, the arrangement determining unit 150 selects an arrangement pattern having the largest number of physical servers among the arrangement patterns registered in the arrangement available list 152. When there are a plurality of arrangement patterns with the largest number of physical servers, the arrangement determining unit 150 selects one of them.

  [Step S39] The placement determination unit 150 instructs the VM placement device 11 to place a VM on a physical server according to the selected placement pattern. Then, the VM placement apparatus 11 rearranges the VM.

  In this way, VM pairs in the communication path indicated in the path list can be arranged in different physical servers. As a result, communication between VMs via the communication path is performed via the switch 52, and can be captured by port mirroring.

  FIG. 20 is a diagram illustrating a first example of arrangement pattern selection. In the example of FIG. 20, it is assumed that four VMs 201 to 204 are arranged on a maximum of four physical servers. The VM number of the VM 201 is “1”, the VM number of the VM 202 is “2”, the VM number of the VM 203 is “3”, and the VM number of the VM 204 is “4”. Four communication paths are registered in the path list 143 indicating the communication paths between the VMs 201 to 204. The registered communication path is a communication path between the VM 201 and the VM 203, a communication path between the VM 202 and the VM 203, a communication path between the VM 202 and the VM 204, and a communication path between the VM 203 and the VM 204. .

  When four VMs 201 to 204 are arranged on a maximum of four physical servers, 14 arrangement patterns 211 to 224 can be generated. In the arrangement patterns 211 to 224 in FIG. 20, the physical server is represented by a rectangle. The VMs arranged in each physical server are indicated by circles with VM numbers in a rectangle representing the physical server.

  Of the 14 arrangement patterns 211 to 224, the three arrangement patterns 222 to 224 are determined to be arrangeable. That is, the arrangement patterns 222 to 224 are arranged in physical servers having different VMs at both ends of the communication path for all the communication paths indicated in the path list 143.

  Among the arrangement patterns 222 to 224 that can be arranged, the VMs 201 to 204 are arranged on three physical servers in the arrangement patterns 222 and 223. In the arrangement pattern 224, the VMs 201 to 204 are arranged on four physical servers.

  Here, if the selection condition is “number priority”, one of the arrangement patterns 222 and 223 is selected. On the other hand, if the selection condition is “performance-oriented”, the arrangement pattern 224 is selected.

FIG. 21 is a diagram illustrating a second example of arrangement pattern selection. In the example of FIG. 21, it is assumed that four VMs 201 to 204 are arranged in a maximum of two physical servers.
When four VMs 201 to 204 are arranged on a maximum of two physical servers, seven arrangement patterns 231 to 237 can be generated. Similar to FIG. 20, in the arrangement patterns 231 to 237 of FIG. 21, the physical server is represented by a rectangle. The VMs arranged in each physical server are indicated by circles with VM numbers in a rectangle representing the physical server.

  In the example of FIG. 21, the generated arrangement patterns 231 to 237 do not have an arrangement pattern that satisfies the conditions shown in the route list 143. Therefore, the communication route information is deleted from the route list 143. For example, a communication path related to a process with a low visualization priority is deleted. In the example of FIG. 21, the communication between the VM 201 and the VM 203 is performed using IIOP (Internet Inter-ORB Protocol). Communication between the VM 203 and the VM 204 is performed using SQL (Structured Query Language). Communication between the VM 202 and the VM 203 is performed using IIOP. Communication between the VM 202 and the VM 204 is performed using ssh (Secure SHell).

  For example, when the administrator determines that the communication of ssh is less necessary to be visualized, the communication path information indicating the ssh communication path is deleted from the path list 143. Then, a route list 144 including the remaining three communication route information is newly created. Since the communication path information has been deleted, the conditions for determining that arrangement is possible are relaxed. As a result, when it is determined whether or not each of the arrangement patterns 231 to 237 can be arranged based on the path list 144 after the communication path is deleted, the arrangement pattern 233 that can be arranged is detected. Therefore, an arrangement pattern 233 that can be arranged is selected.

Thus, by deleting predetermined communication path information from the path list, it is possible to select an appropriate arrangement pattern.
In the above example, the deletion of the communication path information from the path list is performed by an operation input from the administrator. However, the deletion of the communication path information can be automated. For example, for each protocol used for communication, the priority of investigation is set in the arrangement determination unit 150 in advance. In the example of FIG. 21, the priority of “ssh” is set lower than “IIOP” and “SQL”. When creating a route list, a protocol used in each communication route is set in association with each communication route. In the process of step S35 of FIG. 19, the arrangement determining unit 150 deletes one of the communication paths for performing communication using the protocol with the lowest priority from the path list. In this way, it is possible to automate the process of deleting communication path information from the path list.

<System visualization>
Next, system visualization processing will be described.
FIG. 22 is a diagram illustrating an example of the visualization process. FIG. 22 shows an example of analyzing the processing status of the Web three-tier system. In the Web 3-level system, the highest level is configured by a Web server. The Web server interprets a request message in accordance with, for example, HTTP (HyperText Transfer Protocol), and executes the processing indicated in the request message. The second layer is composed of application servers. For example, the application server interprets a request message according to IIOP and executes the processing indicated in the request message. The third layer is composed of DB servers. The DB server interprets a request message according to, for example, SQL, and executes the processing indicated in the request message.

  In the example of FIG. 22, a Web three-tier system is constructed using six VMs 71 to 76. The VMs 71 to 73 have a Web server function. The VMs 74 and 75 have an application server function. The VM 76 has a DB server function.

  In the Web three-tier system, communication occurs between each of the VMs 71 to 73 having a Web server function and each of the VMs 74 and 75 having an application server function. Further, communication occurs between each of the VMs 74 and 75 having the application server function and the VM 76 having the DB server function.

  The VM placement device 11 places the VMs 71 to 76 on the physical servers 41 to 43 in the visualization rack B in accordance with, for example, an instruction from the visualization device 100. For example, VMs 71 to 76 operating on any of the physical servers 31, 32, 33,... In the operation rack A are relocated to the physical servers 41 to 43. In this case, the VM placement apparatus 11 instructs the physical server in the operation rack A in which the VMs 71 to 76 were operating to perform live migration of the VM to one of the physical servers in the visualization rack B. For example, the physical server of the placement destination is specified by the IP address of the OS of the physical server. The physical server in the operation rack A that has received the live migration instruction transmits information on the VM to be rearranged to the physical server of the placement destination. The information to be transmitted includes, for example, a program and data for realizing a VM function. In the example of FIG. 22, VMs 71 to 73 are arranged in the physical server 41. VMs 74 and 75 are arranged in the physical server 42. The VM 76 is arranged on the physical server 43.

  After the rearrangement, the request message from the client addressed to the Web 3-tier system is passed to any VM 71 to 73 that functions as the Web server in the physical server 41. Note that the clients for the Web three-tier system are, for example, terminal devices 21, 22,. When the function of the application server is used for processing according to the request message, the VM that has received the request message from the client transmits the request message to the VMs 74 and 75 that function as the application server. The VM that has received the request message from the VM that functions as the Web server transmits the request message to the VM 76 that functions as the DB server when the function of the DB server is used for processing according to the request message. The VM 76 executes processing according to the received request message, and returns a response message to the VM that transmitted the request message. The VM that has received the response message from the VM 76 responds to the VM functioning as a Web server with the response message after completion of predetermined processing. Then, the VM that has received the response message from the VM functioning as the application server transmits a response message to the client after completion of predetermined processing.

  In this way, a coordinated process by a plurality of VMs is executed. In the example of FIG. 22, VMs that generate communication with each other are arranged in different physical servers. Therefore, all request messages and response messages communicated between VMs are captured by the switch 52. That is, the packets constituting all the request messages and response messages are output from the mirror port 52a.

  Based on the packet output from the mirror port 52a, the analysis unit 110 of the visualization device 100 reconfigures the process (transaction) executed in the Web three-tier system. For example, the analysis unit 110 stores a transaction model in advance. The transaction model is information that defines, for each transaction type, processing contents executed by a VM functioning as a server in each hierarchy in a transaction of the corresponding type, and a calling relationship between the processes.

  The analysis unit 110 analyzes the captured packet and reconfigures the communicated request message and response message. At this time, the analysis unit 110 gives the time information (time stamp) for the communication to the request message and the response message. The request message and the response message include the source IP address and the destination IP address. Therefore, the analysis unit 110 can recognize the VM that is the caller of the process and the VM that is the callee based on the request message and the response message. The process call source VM is the VM of the request message transmission source. Further, the VM that called the process is also the destination VM of the response message. The process call destination VM is the destination VM of the request message. Further, the VM that is the call destination of the process is also the VM that is the transmission source of the response message.

  The analysis unit 110 creates a pair (message pair) of a request message and a response message in which a process calling VM and a calling VM match. A message pair is associated with one process in any VM. Then, the analysis unit 110 reproduces the executed transaction 77 by combining message pairs that satisfy the constraint conditions regarding the call between processes. The constraint condition regarding the call between processes is, for example, a condition that “the time zone of the call source process includes the time zone of the call destination process”. The time zone of each process is a time zone from when a process request message is output until a process corresponding to the request message outputs a response message. The analysis unit 110 checks the reproduced transaction 77 against the transaction model, and determines the type of the executed transaction 77. Further, the analysis unit 110 calculates the processing time in each VM based on the time information of each message included in the reproduced transaction 77. Thereby, the processing time in VM which functions as a server of each hierarchy for every transaction classification can be obtained as an analysis result. An analysis technique for visualizing processing contents using such captured packets is disclosed in detail in, for example, Japanese Patent Application Laid-Open No. 2006-011683.

  As described above, if VMs communicating with each other are arranged in different physical servers, it is possible to visualize the processing of the entire transaction 77 in the multi-tier system based on the captured packet. That is, it is possible to analyze in detail the processing status of a multi-tier system by a plurality of VMs.

[Other Embodiments]
In the first and second embodiments described above, the analysis is performed using the communication information (packet) captured via the switch, but the packet can be simply recorded as a communication log.

  In the first and second embodiments described above, communication information (packets) is captured using the port mirroring function of the switch, but may be captured by other methods. For example, instead of a switch, a packet can be captured by connecting each physical server and a visualization device using a repeater bubble. That is, in the repeater hub, all packets are output from all ports other than the input port. Therefore, the packet addressed to each physical server can be acquired in the visualization device.

  The above processing functions can be realized by a computer. In that case, a program describing processing contents of functions that the virtual machine management apparatus 1 and the visualization apparatus 100 should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic storage device, an optical disk, a magneto-optical recording medium, and a semiconductor memory. Examples of the magnetic storage device include a hard disk device (HDD), a flexible disk (FD), and a magnetic tape. Optical discs include DVD, DVD-RAM, CD-ROM / RW, and the like. Magneto-optical recording media include MO (Magneto-Optical disc).

  When distributing the program, for example, a portable recording medium such as a DVD or a CD-ROM in which the program is recorded is sold. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

  In addition, at least a part of the above processing functions can be realized by an electronic circuit such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or a PLD (Programmable Logic Device).

  As mentioned above, although embodiment was illustrated, the structure of each part shown by embodiment can be substituted by the other thing which has the same function. Moreover, other arbitrary structures and processes may be added. Further, any two or more configurations (features) of the above-described embodiments may be combined.

The techniques disclosed in the above embodiments include the techniques shown in the following supplementary notes.
(Supplementary note 1)
Reference is made to a storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the pair of virtual machines indicated in each of the communication path information is determined as the pair. Determining a server computer to execute each of the plurality of virtual machines so that each of the two virtual machines to be executed is executed by a different server computer;
Outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management program characterized by causing a process to be executed.

(Supplementary Note 2) When determining a server computer that executes each of the plurality of virtual machines,
Generating an arrangement pattern indicating a server computer that executes each of the plurality of virtual machines;
Among the generated arrangement patterns, for the virtual machine pairs indicated in all the communication path information, it is determined that the arrangement patterns executed by different server computers in each of the two virtual machines constituting the pair are applicable,
The virtual machine management program according to appendix 1, wherein a server computer that executes each of the plurality of virtual machines is determined according to an applicable arrangement pattern.

(Supplementary Note 3) When generating an arrangement pattern, a server computer that executes at least a part of the plurality of virtual machines generates a plurality of different arrangement patterns,
When there are multiple applicable arrangement patterns, select an arrangement pattern that meets the selection conditions specified in advance.
The virtual machine management program according to appendix 2, wherein a server computer that executes each of the plurality of virtual machines is determined according to the selected applicable arrangement pattern.

  (Supplementary note 4) The virtual machine management program according to supplementary note 3, wherein, when the selection condition is performance-oriented, an arrangement pattern having the largest number of server computers to be used is selected from among applicable arrangement patterns.

  (Additional remark 5) When the said selection conditions are number saving, the arrangement pattern with the few number of server computers to be used is selected from the applicable arrangement patterns, The additional description 3 or 4 characterized by the above-mentioned Virtual machine management program.

  (Additional remark 6) When there is no applicable arrangement pattern, a part of communication path information in the said memory | storage means is deleted, and the virtual machine pair shown by each of all the communication path information left in the said memory | storage means after deletion is deleted The virtual machine management program according to any one of appendices 3 to 5, wherein an arrangement pattern that is executed by different server computers in each of the two virtual machines that form the pair is determined to be applicable.

(Appendix 7) Further to the computer,
Each of the plurality of virtual machines is executed by a plurality of server computers connected by a switch, and information communicated between the virtual machines via the switch in accordance with execution of processing using the plurality of virtual machines is transmitted from the switch. Acquired,
Generating communication path information indicating a pair of a destination virtual machine and a transmission source virtual machine of the acquired information, and storing the generated communication path information in the storage unit;
The virtual machine management program according to any one of appendices 1 to 6, wherein the virtual machine management program executes processing.

(Supplementary Note 8) When each of the plurality of virtual machines is executed by a plurality of server computers connected by a switch,
For all of the virtual machine pairs obtained by selecting two virtual machines of the plurality of virtual machines, the virtual machines that are paired are executed on different server computers at least once. Generating at least one arrangement pattern indicating a combination of each machine and a server computer executing a virtual machine;
Each time a placement pattern is generated, information indicating a server computer that executes each of the plurality of virtual machines is output according to the placement pattern.
The virtual machine management program according to appendix 7, wherein

(Supplementary note 9)
Reference is made to a storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the pair of virtual machines indicated in each of the communication path information is determined as the pair. Determining a server computer to execute each of the plurality of virtual machines so that each of the two virtual machines to be executed is executed by a different server computer;
Outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management method.

(Additional remark 10) With reference to the memory | storage means to which the at least 1 communication path information which shows the pair of two virtual machines which mutually communicate among several virtual machines is registered, About the virtual machine pair shown by each communication path information Execution server determining means for determining a server computer for executing each of the plurality of virtual machines so that each of the two virtual machines in the pair is executed by different server computers;
Output means for outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management apparatus comprising:

DESCRIPTION OF SYMBOLS 1 Virtual machine management apparatus 1a Storage | storage means 1b Execution server determination means 1b-1 Arrangement pattern generation means 1b-2 Availability determination means 1b-3 Selection means 1b-4 Determination means 1c Output means 2, 3 Server computer 4 Switch 5 Analysis apparatus 6 Virtual machine placement device 7a, 7b, 7c, 7d Virtual machine 8a, 8b, 8c,... Placement pattern 9a, 9b, 9c Communication path information

Claims (7)

On the computer,
Reference is made to a storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the pair of virtual machines indicated in each of the communication path information is determined as the pair. Determining a server computer to execute each of the plurality of virtual machines so that each of the two virtual machines to be executed is executed by a different server computer;
Outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management program characterized by causing a process to be executed. When determining a server computer for executing each of the plurality of virtual machines,
Generating an arrangement pattern indicating a server computer that executes each of the plurality of virtual machines;
Among the generated arrangement patterns, for the virtual machine pairs indicated in all the communication path information, it is determined that the arrangement patterns executed by different server computers in each of the two virtual machines constituting the pair are applicable,
The virtual machine management program according to claim 1, wherein a server computer that executes each of the plurality of virtual machines is determined according to an applicable arrangement pattern. When generating the arrangement pattern, a server computer that executes at least a part of the plurality of virtual machines generates a first plurality of arrangement patterns different from each other,
When there are multiple applicable arrangement patterns, select an arrangement pattern that meets the selection conditions specified in advance.
3. The virtual machine management program according to claim 2, wherein a server computer that executes each of the plurality of virtual machines is determined in accordance with the selected applicable arrangement pattern.   When there is no applicable arrangement pattern, a part of the communication path information in the storage unit is deleted, and the virtual machine pair indicated in each of all the communication path information remaining in the storage unit after the deletion is 4. The virtual machine management program according to claim 3, wherein it is determined that an arrangement pattern that is executed by different server computers in each of the two virtual machines that constitute a virtual machine is applicable. In addition to the computer,
Among a plurality of server computers connected by the switch, in accordance with the respective second plurality of arrangement patterns indicating the server computer to perform each of the plurality of virtual machines, repeated each of the plurality of virtual machines in the plurality of server computers Each time the plurality of virtual machines are executed on the plurality of server computers, information communicated between the virtual machines via the switch is executed from the switch along with execution of the process using the plurality of virtual machines. Acquired,
Generating communication path information indicating a pair of a destination virtual machine and a transmission source virtual machine of the acquired information, and storing the generated communication path information in the storage unit;
The virtual machine management program according to claim 1, wherein the virtual machine management program executes processing.
Computer
Reference is made to a storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the pair of virtual machines indicated in each of the communication path information is determined as the pair. Determining a server computer to execute each of the plurality of virtual machines so that each of the two virtual machines to be executed is executed by a different server computer;
Outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management method. Reference is made to a storage means in which at least one communication path information indicating a pair of two virtual machines that communicate with each other among a plurality of virtual machines is registered, and the pair of virtual machines indicated in each of the communication path information is determined as the pair. Execution server determining means for determining a server computer for executing each of the plurality of virtual machines, such that each of the two virtual machines is executed by different server computers;
Output means for outputting information indicating a server computer that executes each of the plurality of virtual machines;
A virtual machine management apparatus comprising:

Images