JP2011186805A - Method for configuring information processing system and physical machine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for configuring an information processing system which copes with the movement of a virtual machine between different subnets in a fixed network and to provide a physical machine. <P>SOLUTION: When a virtual machine (5) moves as a virtual machine (B) from a physical machine 1 accommodated in a first subnet in a fixed network to a physical machine 2 accommodated in a second subnet which is different from the first subnet, a layer 2 tunnel extending from the physical machine 1 to the physical machine 2 is constructed. Traffic to the virtual machine (5) of the physical machine 1 is distributed to the layer 2 tunnel by means of the physical machine 1 and a network interface to reach the virtual machine (B) of the physical machine 2. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an information processing system configuration method and a physical machine, and more particularly, to an information processing system configuration method and a physical machine that can cope with movement of a virtual machine between different subnets in a fixed network.

  There is a mobility support technology in an IP network as a technology for continuously using an application even if a terminal on the network is physically moved to another network. This is described in Non-Patent Documents 1 and 2.

  The mobility support technologies described in Non-Patent Documents 1 and 2 are based on a mobile network constructed for so-called mobile phone use in which the network itself supports mobility. It does not support mobility on existing fixed networks built for video distribution to top boxes.

  In Patent Document 1, the virtual machine at the destination controls the source device in order to absorb the difference in the device driver environment that occurs when the virtual machine that was running on the source host is moved to the destination host. A computer system is described in which a front-end driver and a back-end driver are prepared so that an application can continue operation without being aware of the difference between the drivers. Using this computer system, when moving a virtual machine that was running on a physical server to another physical server, the differences in the device driver environment due to the difference between the source and destination physical servers are hidden. be able to.

JP 2010-9396 A

Standard document: RFC3344 Standard document: RFC3775

  However, the mobility support technologies described in Non-Patent Documents 1 and 2 are based on the premise of so-called mobile networks, and are applied to existing fixed networks constructed for video distribution to fixed telephones and set-top boxes. It is not possible.

  If the computer system disclosed in Patent Document 1 is used, it is possible to conceal the driver difference due to the difference between the migration source and the migration destination physical server. However, this does not solve the problem of moving virtual machines across different subnets in a fixed network. Even if a virtual machine that was running on a physical machine in a subnet is moved to a physical machine in a different subnet, the IP address used by the virtual machine does not change. As a result, there is a problem in that it becomes impossible to access the data held in the migration source physical machine from the migration destination virtual machine, and the execution of the application cannot be continued. The computer system disclosed in Patent Document 1 does not solve such a problem.

  In order to realize mobility support technology on an existing fixed network, it is also possible to move various data used by the virtual machine together with the virtual machine. However, in this case, a new problem arises, that is, the data is usually enormous, and a large load is applied to the network along with the movement of the data, and it takes time to move the data. There arises a problem that the application itself cannot be started quickly at the destination.

  In addition, when moving data, it is necessary to dynamically manage and change the IP address along with the movement, and to change the application environment in conjunction with the change of the IP address. The original purpose of moving a virtual machine is to avoid changing the IP address dynamically by managing it or changing the application environment in conjunction with the IP address change. Then that purpose will not be achieved.

  Therefore, when moving a virtual machine that was running on a physical machine to another physical machine, only the virtual machine that is the application execution environment (application image) is moved, and various data used by the virtual machine (for example, In the case of a spreadsheet application, the form data input to the table itself is usually left on the physical machine that is the source of movement.

  As described above, until now, virtual servers can only move within the same subnet (same layer 2 broadcast domain), and a method for moving virtual servers across different subnets has been realized. It has not been. In order to achieve this in a fixed network that does not support terminal mobility, it is necessary to solve the problem caused by the network itself that the subnets are different.

  An object of the present invention is to provide an information processing system configuration method and a physical machine that can cope with movement of virtual machines between different subnets in a fixed network.

  In order to solve the above-mentioned problem, the configuration method of the information processing system according to the present invention is such that a virtual machine is different from the first subnet from the first physical machine accommodated in the first subnet in the fixed network. When moving to a second physical machine accommodated in the second subnet, a layer 2 tunnel from the first physical machine to the second physical machine is constructed, and the virtual machine of the first physical machine is established. The traffic is distributed to the layer 2 tunnel.

  In addition, in the configuration method of the information processing system according to the present invention, when a virtual machine moves from the first physical machine to the second physical machine, the virtual machine stored in the database of the first physical machine The application image is copied to the database of the second physical machine and stored in the database of the second physical machine, and the virtual machine acquires the application image stored in the database of the second physical machine. Furthermore, various data stored in the database of the first physical machine is acquired through a layer 2 tunnel, and the various data is processed according to the acquired application image.

  In the configuration method of the information processing system according to the present invention, the first physical machine acquires configuration information of a physical machine in the network stored in the database of the first physical machine, and the configuration information It is characterized by constructing a layer 2 tunnel based on this.

  In the information processing system configuration method according to the present invention, the first physical machine acquires configuration information of a physical machine in the network from an external operation management system, and creates a layer 2 tunnel based on the configuration information. It is characterized by construction.

  In the physical machine according to the present invention, the virtual machine is accommodated in a second subnet different from the first subnet from the own physical machine accommodated in the first subnet in the fixed network. When moving to the second physical machine, create a layer 2 tunnel from the local physical machine to the second physical machine and distribute the traffic to the virtual machine on the first physical machine to the layer 2 tunnel It is characterized by having.

  Conventionally, virtual machines can only be moved within a network that is closed within the same layer 2 broadcast domain (for example, within one data center). You can move the virtual machine. As a result, a larger number of physical machines can be shared, the operating efficiency of the physical machines can be increased by the statistical multiplexing effect, and a reduction in capital investment and power consumption can be expected.

  In addition, in the prior art RFC3344, RFC3775 is a mechanism for realizing the movement of the terminal inside the IP network, tunnel processing (IP in IP, GRE, etc.) at the layer 3 level is necessary, but in the present invention, Tunnel processing at the layer 2 level that can be realized on the physical machine side is sufficient. In general, since processing at a high layer is more complicated, according to the present invention, processing can be simplified.

It is a block diagram which shows the information processing system to which this invention is applied. It is a figure which shows notionally the structure method of the information processing system which concerns on this invention. It is a figure which shows the internal structure of a physical machine. It is a figure which shows an example of the operation | movement sequence in the structure method of the information processing system which concerns on this invention.

  The present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing an information processing system to which the present invention is applied. This information processing system is configured by connecting a physical machine and a user terminal to an IP network which is an existing fixed network. Here, the physical machines 1 and 2 and the user terminal 3 are illustrated, but it is normal that more physical machines and user terminals are connected to the IP network. The physical machines 1 and 2 are configured as physical servers, for example.

  The physical machine 1 is connected to the IP network via the router 4, the physical machine 2 is connected to the IP network via the router 5, and the user terminal 3 is connected to the IP network via the router 6. Databases 7 and 8 are connected to the physical machine 1 and the physical machine 2, respectively. For example, the physical machines 1 and 2 execute various processes according to access from the user terminal 3, and return the results as responses.

  Since the physical machine 1 and the physical machine 2 are connected to the IP network via separate routers 4 and 5, the physical machine 1 and the physical machine 2 are accommodated in different subnets in the IP network.

  Virtual machines (1) to (5) are operating on the physical machine 1, and virtual machines (A) and (C) to (E) are operating on the physical machine 2. Here, it is assumed that the virtual machine (5) of the physical machine 1 moves to the physical machine 2 as a virtual machine (B). The virtual machine (5 (B)) is, for example, a virtual server that executes a spreadsheet application.

  Before the virtual machine (5) is moved from the physical machine 1 to the physical machine 2, the database 7 includes (1) application images of the virtual machines (1) to (5), (2) virtual machines (1) to (5) (3) Configuration information (for example, IP addresses) of physical machines 1 and 2 in the IP network, and database 8 stores (1) virtual machines (A), (C) to (E ) Application image, (2) various data used by virtual machines (A), (C) to (E), and (3) configuration information of physical machines 1 and 2 in the IP network.

  The user can use the virtual machine (5) by using the IP address of the virtual machine (5) of the physical machine 1 (S1). Here, when the virtual machine (5) moves (S2) to the physical machine 2 as the virtual machine (B), the virtual machines (A) to (E) exist in the physical machine 2. Even when the virtual machine (5) is moved to become the virtual machine (B) of the physical machine 2, the user is not aware of the configuration change such as the movement of the virtual machine and the virtual machine (B ) Can be used continuously (S3).

  The present invention satisfies such a need. FIG. 2 is a diagram conceptually showing a method of configuring the information processing system according to the present invention. Note that the present invention is also realized as a physical machine constituting such an information processing system.

  When the virtual machine (5) of the physical machine 1 moves to the physical machine 2 as a virtual machine (B), a layer 2 tunnel is constructed between the physical machine 1 and the physical machine 2. Specifically, although it depends largely on the implementation, a layer 2 tunnel is constructed between the network interface of the physical machine 1 and the hypervisor of the virtual machine B. The layer 2 tunnel can be constructed by a layer 2 tunneling protocol, for example, L2TPv3 (Layer 2 Tunneling Protocol version 3).

  In addition, the application image of the virtual machine 5 stored in the database 7 is copied and stored in the database 8 as the application image of the virtual machine (B). Further, the databases 7 and 8 are connected (mounted) to the virtual machine (B).

  As a result, the virtual machine (B) uses the application image of the virtual machine B stored in the database 8, and further uses various data stored in the database 9 (the application image of the virtual machine (5) was used. Various data) can be used through the layer 2 tunnel.

  Traffic from the user terminal 3 to the virtual machine (5) is distributed at the network interface of the physical machine 1, and an IP address is added (encapsulation) with the transmission source as the physical machine 1 and the transmission destination as the physical machine 2 Through the layer 2 tunnel to the virtual machine (B).

  The virtual machine (B) acquires the application image of the virtual machine (B) from the database 8, and acquires various data used by the application image from the database 7 through the layer 2 tunnel. Then, various data is processed according to the application image, and the processing result is returned to the user terminal 3. Further, various data of the processing results are stored in the database 7. By using the layer 2 tunnel in this way, the user receives service continuously using the virtual machine (B) on the physical machine 2 as if the user continues to communicate with the virtual machine (5). be able to.

  FIG. 3 is a diagram illustrating an internal configuration of the physical machine 1. The physical machine 2 has the same configuration.

  As shown in the figure, the physical machine 1 includes a network interface 31, a hypervisor 32, and virtual machines (1) to (5). The network interface 31 connects the physical machine 1 to the IP network via the router 4. The hypervisor 32 is configured by software, and operates on the network interface 31 to construct a virtual machine. The virtual machines (1) to (5) operate under the management of the hypervisor 32.

  Each of the virtual machines (1) to (5) has the same configuration, but only the configuration of the virtual machine (5) is illustrated here. The virtual machine (5) includes a network interface 33, a guest OS 34, and an application image 35. The network interface 33 is interposed between the hypervisor 32 and the guest OS 34 and provides a virtual interface. The guest OS 34 is a virtual OS. The application image 35 operates under the management of the guest OS 34.

  The layer 2 tunnel is actually constructed between the network interface (network interface 33) of the virtual machine (B) and the network interface 31 of the physical server 1 itself.

  FIG. 4 is a diagram showing an example of an operation sequence in the method of configuring the information processing system according to the present invention.

  Before the virtual machine (5) moves, the virtual machine (5) accesses the database 7 according to a request from the user terminal 3, acquires an application image and various data used by the database, and acquires various data according to the application image. Process (S41) and return the processing result to the user terminal 3.

  Here, when the physical machine 1 (specifically, the hypervisor) receives a movement request for the virtual machine (5) (S42), the physical machine 1 obtains the configuration information of the physical machine 2 in the IP network from the database 7. Obtain (S43) and build a layer 2 tunnel from the physical machine 1 to the physical machine 2 (S44). The layer 2 tunnel from the physical machine 1 to the physical machine 2 can be understood from the configuration information of the physical machine 2 in the IP network. That is, a layer 2 tunnel from the physical machine 1 IP address to the physical machine 2 IP address may be constructed.

  On the other hand, on the physical machine 2 (hypervisor) side, the virtual machine (B) is registered in response to the request to move the virtual machine (5) on the physical machine 1 side (S45), and the virtual machine (B) Reserve space.

  After constructing the layer 2 tunnel, the physical machine 1 rewrites the route information in the hypervisor 32 (S46). That is, traffic from the user terminal 3 to the virtual machine (5) is distributed at the network interface 31, and the route information is rewritten so as to reach the hypervisor of the physical machine 2 through the layer 2 tunnel.

  Next, the application image of the virtual machine (5) stored in the database 7 is copied and stored in the database 8 as the application image of the virtual machine (B) (S47).

  Next, the virtual machine (B) is connected (mounted) to the database 8 (S48). Further, the virtual machine (5) is disconnected (unmounted) from the database 7 (S49), and then the virtual machine (B) is connected (mounted) to the database 7 (S50). The reason why the virtual machine (5) is separated from the database 7 is to eliminate contention for the database 7.

  Traffic from the user terminal 3 to the virtual machine (5) is distributed at the network interface 31 of the physical machine 1, reaches the hypervisor of the physical machine 2 through the layer 2 tunnel, and is further sent to the virtual machine (B). .

  The virtual machine (B) accesses the database 8 according to a request from the user terminal 3 to acquire the application image of the virtual machine (B) (S51), and further accesses the database 7 using a layer 2 tunnel. Then, various data used by the application image of the virtual machine (B) (various data used when the virtual machine (B (5)) was operating on the physical server 1) is acquired (S52). The virtual machine (B) processes various data according to the application image and returns the processing result to the user terminal 3. Further, various data of the processing results are stored in the database 7 using a layer 2 tunnel.

  Although the embodiments have been described above, the present invention is not limited to the above embodiments, and includes various modifications. For example, in the above embodiment, the configuration information of the physical machines 1 and 2 is stored in the databases 7 and 8, and the physical machine 1 (hypervisor) acquires the configuration information of the physical machine 2 from the database 7. The configuration information of the machines 1 and 2 may be separately stored in an external operation management system or the like, and the physical machine 1 (hypervisor) may acquire the configuration information of the physical machine 2 from the outside.

1,2 ... physical machine, 3 ... user terminal, 4,5,6 ... router, 7,8 ... database, 31,33 ... network interface, 32 ... hypervisor, 34 ... Guest OS, 35 ... Application image, (1)-(5), (A)-(E) ... Virtual machine

Claims (5)

A method of configuring an information processing system having a movable virtual machine,
When the virtual machine moves from the first physical machine accommodated in the first subnet in the fixed network to the second physical machine accommodated in the second subnet different from the first subnet, A configuration of an information processing system that constructs a layer 2 tunnel from the first physical machine to the second physical machine and distributes traffic to the virtual machine of the first physical machine to the layer 2 tunnel Method.   When a virtual machine is moved from the first physical machine to the second physical machine, an application image of the virtual machine stored in the database of the first physical machine is stored in the database of the second physical machine. And the virtual machine acquires the application image stored in the database of the second physical machine and further stores it in the database of the first physical machine. 2. The method of configuring an information processing system according to claim 1, wherein the acquired various data is acquired through a layer 2 tunnel, and the various data is processed in accordance with the acquired application image.   The first physical machine acquires configuration information of a physical machine in a network stored in a database of the first physical machine, and builds a layer 2 tunnel based on the configuration information. The method for configuring the information processing system according to claim 1 or 2.   The first physical machine acquires configuration information of a physical machine in a network from an external operation management system, and constructs a layer 2 tunnel based on the configuration information. Configuration method of the information processing system. A physical machine used in an information processing system having a movable virtual machine,
When the virtual machine moves from the own physical machine accommodated in the first subnet in the fixed network to the second physical machine accommodated in the second subnet different from the first subnet, A physical machine having a function of constructing a layer 2 tunnel from its own physical machine to a second physical machine and distributing traffic to a virtual machine on the first physical machine to the layer 2 tunnel.

Images