JP5563126B1 - Information processing apparatus and detection method - Google Patents

Abstract

It is determined whether or not a plurality of pages having the same contents are grouped together.
An information processing apparatus disclosed in the present application realizes a plurality of virtual machines and allocates a part of physical memory to the memory of each virtual machine. The first virtual machine includes a securing unit, a measurement unit, and a determination unit. The securing unit secures a measurement area in the physical memory, writes a random number sequence as first data to the first area of the measurement area, and uses a fixed character for the second area of the measurement area. Write the column as second data. The measurement unit reads the first data and the second data from the measurement region at predetermined intervals, and measures the first data read performance and the second data read performance. The determination unit determines whether the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data.
[Selection] Figure 1

Description

  The present invention relates to an information processing apparatus and a detection method.

  In recent years, as represented by cloud computing, information systems (service providing systems) that provide services via a network have been widely used. For example, a service providing system includes IaaS (Infrastructure as a Service) that provides a user with a virtual machine via a network.

  In such a service providing system, a plurality of virtual machines (hereinafter referred to as “VM (Virtual Machine)”) are realized, and physical resources such as a physical memory and a CPU (Central Processing Unit) are allocated to each VM. . Here, the total amount of physical resources allocated to each VM may be allocated more than the physical resources of the computer. This is a technique called overcommitment, and assumes that all VMs do not use all physical resources at the same time.

  As part of the overcommitment method, there is KSM (Kernel SamePage Merging) that efficiently allocates physical memory to VMs. In KSM, when the contents of pages stored in a plurality of memory areas in a VM are the same, or when the contents of pages stored in a plurality of memory areas are the same between VMs, the contents are the same. A plurality of pages are combined into one and stored in the physical memory area. As a result, for example, when realizing a plurality of VMs that start from the same image and hold pages having the same contents, physical memory can be efficiently allocated to each VM. In KSM, when a change occurs in the content of a saved page, a new memory area is allocated to the changed page as needed.

  In a service providing system that employs such memory overcommitment, a paging (swap) technique is used when the virtual machine is busy and lacks resources. Paging is a method for responding to a new memory usage request by saving (page out) a memory area with low usage frequency to an HDD (Hard Disk Drive) or the like.

Adai University, “Packing Algorithm for Virtual Machine Placement”, IPSJ SIG Technical Report, Vol.2010-EVA-33, No.4, 2010/11/29 Red Hat Customer Portal, Chapter 7 KSM, [online], 2013, [Search April 9, 2013], Internet (URL: https://access.redhat.com/knowledge/docs/en-US/ Red_Hat_Enterprise_Linux / 6 / html / Virtualization_Administration_Guide / chap-KSM.html)

  However, the conventional technique has a problem that it cannot be determined whether or not a plurality of pages having the same content are grouped together. Specifically, when the management subject of a service providing system such as a host OS (Operating System) or a hypervisor is different from the user of the virtual machine, the user of the virtual machine determines whether or not KSM is enabled. Cannot be identified.

  An embodiment of the disclosure has been made in view of the above, and an object thereof is to determine whether or not a plurality of pages having the same contents are combined into one.

  The information processing apparatus disclosed in the present application realizes a plurality of virtual machines and allocates a part of physical memory to the memory of each virtual machine. The first virtual machine includes a securing unit, a measurement unit, and a determination unit. The securing unit secures a measurement area in the physical memory, writes a random number sequence as first data to the first area of the measurement area, and uses a fixed character for the second area of the measurement area. Write the column as second data. The measurement unit reads the first data and the second data from the measurement region at predetermined intervals, and measures the first data read performance and the second data read performance. The determination unit determines whether the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data.

  According to one aspect of the information processing device to be disclosed, there is an effect that it can be determined whether or not a plurality of pages having the same contents are grouped into one.

FIG. 1 is a diagram illustrating an example of the overall configuration of a computer system according to the first embodiment. FIG. 2 is a diagram illustrating an example of information stored in the physical memory management DB according to the first embodiment. FIG. 3 is a diagram illustrating an example of information stored in the page-out management DB according to the first embodiment. FIG. 4 is a diagram illustrating an example of a data structure stored in the reference value storage unit according to the first embodiment. FIG. 5 is a diagram for explaining the processing operation performed by the VM execution unit according to the first embodiment. FIG. 6 is a flowchart illustrating the procedure of the detection process performed by the VM execution unit according to the first embodiment. FIG. 7 is a diagram for explaining the processing operation performed by the VM execution unit according to the second embodiment. FIG. 8 is a diagram showing that information processing by a detection program for executing processing by a computer system is specifically realized using a computer.

  Hereinafter, embodiments of an information processing apparatus and a detection method to be disclosed will be described in detail based on the drawings. The invention disclosed by this embodiment is not limited. Each embodiment can be appropriately combined as long as the processing contents do not contradict each other.

(First embodiment)
FIG. 1 is a diagram illustrating an example of the overall configuration of a computer system according to the first embodiment. As illustrated in FIG. 1, the computer system includes a client terminal 1 a, a client terminal 1 b, and an information processing apparatus 10. In the example illustrated in FIG. 1, in the computer system, a client terminal 1 a, a client terminal 1 b, and an information processing apparatus 10 are connected via an arbitrary network 2. Further, the number of client terminals is not limited to the illustrated number.

  The client terminal 1a uses a service provided by the information processing apparatus 10. For example, the client terminal 1 a uses resources provided by the information processing apparatus 10. Here, the resource provided by the information processing apparatus 10 is, for example, a virtual machine (hereinafter referred to as “VM (Virtual Machine)”). The client terminal 1a is an information processing apparatus such as a PC (Personal Computer) or a server apparatus. The client terminal 1b is a terminal that functions similarly to the client terminal 1a. For this reason, the detailed description about the client terminal 1b is abbreviate | omitted. Moreover, in the following description, when not distinguishing the client terminals 1a and 1b, they are called the client terminals 1. In the first embodiment, a case where the user of the client terminal 1 is a user of a service contracted with an IaaS (Infrastructure as a Service) provider will be described.

  The information processing apparatus 10 is, for example, a PC, a server apparatus or the like, and operates with a service request from the client terminal 1 as a trigger, and provides a specific service such as a VM to the client terminal 1 via the network 2. For example, the information processing apparatus 10 realizes a plurality of virtual machines and allocates a part of the physical memory to the memory of each virtual machine. In the following description, the total capacity of the physical memory included in the information processing apparatus 10 is also referred to as “physical resource”.

  In the information processing apparatus 10, the first VM secures a measurement area in the physical memory, writes a random number sequence as first data to the first area of the measurement area, and also measures the measurement area. A fixed character string is written as second data in the second area. The first VM reads the first data and the second data from the measurement area at a predetermined interval, and measures the first data read performance and the second data read performance. Then, the first VM determines whether the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data.

  Next, details of the information processing apparatus 10 will be described with reference to FIG. As illustrated in FIG. 1, the information processing apparatus 10 includes a communication control unit 11, a swap area 12, a storage unit 13, a control unit 14, and a virtual area 15.

  The communication control unit 11 is a processing unit that controls communication with the client terminal 1. For example, the communication control unit 11 receives an execution request or operation information from the client terminal 1 and outputs it to the VM management unit 14a. In addition, the communication control unit 11 transmits the VM screen information received by the VM management unit 14 a to the client terminal 1. The communication control unit 11 is a network interface card, for example.

  The swap area 12 is an area realized by a storage medium such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). In this swap area 12, data in a memory area with low usage frequency is saved (also referred to as “page-out”) when physical resources are insufficient.

  The storage unit 13 is a storage unit that stores a program and data executed by the processor and includes a physical memory management DB (Data Base) 13a and a page-out management DB 13b.

  The physical memory management DB 13a is a database that stores “physical memory address” and “machine address” in association with each other. FIG. 2 is a diagram illustrating an example of information stored in the physical memory management DB 13a according to the first embodiment. As illustrated in FIG. 2, the physical memory management DB 13 a stores information in which “physical memory address” and “machine address” are associated with each other.

  Here, the “physical memory address” stored in the physical memory management DB 13 a indicates an address in the physical memory of the information processing apparatus 10. For example, data values such as “0x400 to 0x24000” and “0x24000 to 0x44000” are stored in the “physical memory address”. “Machine address” indicates an address of a memory (also referred to as a virtual memory) allocated to each VM. In other words, the “machine address” indicates an address of physical memory that can be seen from the VM. For example, data values such as “0x000 to 0x20000” and “0x20000 to 0x40000” are stored in the “machine address”.

  As an example, the physical memory management DB 13a illustrated in FIG. 2 indicates that “0x400 to 0x24000” in “physical memory address” corresponds to “0x000 to 0x20000” in “machine address”.

  The page-out management DB 13b is a database that stores “machine address” and “swap area address” in association with each other. FIG. 3 is a diagram illustrating an example of information stored in the page-out management DB 13b according to the first embodiment. As illustrated in FIG. 3, the page-out management DB 13b stores “machine address” and “swap area address” in association with each other.

  Here, the “swap area address” stored in the page-out management DB 13 b indicates an address in the swap area 12. For example, data values such as “0x800 to 0x1000” and “0x1000 to 0x1200” are stored in the “swap area address”. The “machine address” stored in the page-out management DB 13b is the same as the “machine address” stored in the physical memory management DB 13a.

  For example, the page-out management DB 13b illustrated in FIG. 3 indicates that “0x1000 to 0x1200” in “machine address” corresponds to “0x800 to 0x1000” in “swap area address”.

  The control unit 14 includes a VM management unit 14a, a physical memory control unit 14b, a paging control unit 14c, and a KSM control unit 14d.

  The VM management unit 14 a is a processing unit that operates a plurality of VM execution units 16, VM execution units 17, and VM execution units 18 in which different operating environments are constructed in the virtual area 15. In addition, the VM management unit 14a allocates a virtual processor and a virtual memory to an OS (Operating System) operating on each VM. As a result, each VM can execute various processes independently. The virtual memory is a virtual memory realized by allocating a predetermined area in the physical memory of the information processing apparatus 10 as a memory used by the VM OS. The virtual processor is a virtual processor realized by allocating a predetermined processing capability in the processor of the information processing apparatus 10 as a processor used by the VM OS.

  The physical memory control unit 14b controls writing of data to the physical memory and reading of data from the physical memory in response to a request from each VM. For example, when the physical memory control unit 14b receives an instruction to write a predetermined character string in the physical memory from the VM execution unit 16 (the VM execution unit 17, the VM execution unit 18), which will be described later, the physical memory at the specified address Write data to. Further, when the physical memory control unit 14b receives an instruction to read data from the physical memory from the VM execution unit 16 (the VM execution unit 17, the VM execution unit 18), the physical memory control unit 14b reads the data from the physical memory at the specified address. put out.

  For example, when the physical memory control unit 14b receives an instruction to write a predetermined character string in the physical memory from the securing unit 18b described later, the physical memory control unit 14b writes the data to the physical memory at the designated address. When the physical memory control unit 14b receives an instruction to read data from the physical memory from the measurement unit 18c described later, the physical memory control unit 14b reads the data from the physical memory at the designated address.

  When the physical memory control unit 14b is requested to write data to the physical memory and there is no free space in the physical memory, the physical memory control unit 14b requests the paging control unit 14c described later to perform page-out processing. Further, when the requested data is paged out when the physical memory control unit 14b is requested to read data from the physical memory, the physical memory control unit 14b requests the paging control unit 14c described later to perform page-in processing.

  The paging control unit 14c pages out any data to the swap area 12 when a page-out process is requested from the physical memory control unit 14b. For example, the paging control unit 14c saves data in a memory area that is not frequently used in a swap area. Further, when the page-in process is requested from the physical memory control unit 14b, the paging control unit 14c reads the requested data from the swap area 12 and pages it into the free area of the physical memory.

  When KSM is enabled, the KSM control unit 14d stores pages having the same contents stored in a plurality of memory areas into one physical memory. For example, the KSM control unit 14d scans the physical memory and determines whether there is a page having the same content. When the KSM control unit 14d determines that pages with the same content exist, the KSM control unit 14d combines the pages with the same content into one.

  In addition, when the collected pages are updated, the KSM control unit 14d allocates new physical memory to the updated pages. The control unit 14 may be configured without the KSM control unit 14d. In such a case, KSM is not activated in the information processing apparatus 10.

  The virtual area 15 is an area managed by the VM management unit 14a, and an arbitrary number of VMs can be operated. For example, in the virtual area 15, a plurality of VM execution units 16, VM execution units 17, and VM execution units 18 in which different operation environments are constructed operate under the control of the VM management unit 14 a. The VM execution unit 16 is a virtual machine to which a virtual processor and virtual memory are allocated. Here, it is assumed that the virtual memory area allocated to the VM execution unit 16 is a machine address “0x000 to 0x20000”. Similarly, the VM execution unit 17 is a virtual machine to which a virtual processor or virtual memory is allocated. Here, it is assumed that the virtual memory area allocated to the VM execution unit 17 is a machine address “0x20000 to 0x40000”.

  The VM execution unit 18 is a virtual machine to which a virtual processor and virtual memory are allocated. The VM execution unit 18 has the following functions in addition to the functions of the VM execution unit 16 and the VM execution unit 17. That is, the VM execution unit 18 includes a reference value storage unit 18a, a securing unit 18b, a measurement unit 18c, and a determination unit 18d. The VM execution unit 18 is also referred to as a “first virtual machine”. It is assumed that the virtual memory area allocated to the VM execution unit 18 is a machine address “0x40000 to 0x60000”.

  The reference value storage unit 18a stores a predetermined reference value for determining whether or not KSM is enabled. In other words, the reference value storage unit 18a stores a predetermined reference value for determining whether or not a plurality of pages having the same content are grouped into one. FIG. 4 is a diagram illustrating an example of a data structure stored in the reference value storage unit 18a according to the first embodiment. As illustrated in FIG. 4, the reference value storage unit 18 a stores information in which “machine address”, “character string”, and “reference value” are associated with each other.

  Here, the “machine address” stored in the reference value storage unit 18a is the same as the “machine address” stored in the physical memory management DB 13a. The “character string” stored in the reference value storage unit 18a indicates a character string written as the first data or the second data. For example, in the “character string”, a character string such as a random number string “agjedois...” Written as the first data and a fixed character string “000... 00” written as the second data is stored. Is done. The “random number sequence” referred to here is a random character string determined by an encryption method, and indicates a character string that is difficult to be combined into one when KSM is enabled. "Is a character string that has a high possibility of appearing in applications and data used by the other VM execution units 16 and 17 and the VM execution unit 18, and is easily combined into one when KSM is enabled. Indicates a character string.

  Further, the “reference value” stored in the reference value storage unit 18a indicates a predetermined reference value for determining whether or not KSM is activated. The reference value is set based on the read performance of the first data measured immediately after the first data is written and the read performance of the second data measured immediately after the second data is written. Is done. For example, the “reference value” stores a data value indicating a time such as “5.0 (seconds), 8 (MB / s)” and a data value indicating a speed. This reference value is set by the measuring unit 18c. Details of the process for setting the reference value will be described later. In addition, in the “reference value” of the reference value storage unit 18a, either a data value indicating time or a data value indicating speed may be stored.

  As an example, the reference value storage unit 18a shown in FIG. 4 has a character string “agjedois...” Written in a memory area whose machine address is “0x40000-0x40200”. 5.0 (seconds), 8 (MB / s) ". In the reference value storage unit 18a shown in FIG. 4, the character string “000... 00” is written in the memory area having the machine address “0x40200 to 0x40400”, and the reference value of the reading performance is “5. 5 (seconds), 7.5 (MB / s) ".

  The securing unit 18b secures a measurement area in the physical memory, writes a random number sequence as first data to the first area of the measurement area, and fixes it to the second area of the measurement area. Write a character string as second data. For example, the securing unit 18b secures machine addresses “0x40000-0x40200” and “0x40200-0x40400” as measurement areas when receiving an instruction to start detection processing from the user. Then, the securing unit 18b writes the random number sequence “agjedois...” As the first data to the machine address “0x40000 to 0x40200”, and the fixed character string “000... 00” to the machine address “0x40200 to 0x40400”. Write as second data.

  More specifically, when 2 GB is secured as a measurement area, the securing unit 18 b divides 2 GB into two equal parts, one 1 GB is written to the first data, and the other 1 GB is written to the second data. Assign to a region. Here, one page is 4 KB, and 4 KB is written in the memory area as one unit. For example, the securing unit 18b writes, as first data, a random number sequence determined using an encryption method so that the same page does not appear repeatedly in a 1 GB memory area. The securing unit 18b repeatedly writes the same page as the second data in the 1 GB memory area. For example, the securing unit 18b writes, as the second data, a fixed character string in which all character strings in the page are configured with “0”.

  The measurement unit 18c reads the first data and the second data from the measurement area at predetermined intervals, and measures the first data read performance and the second data read performance. For example, the measurement unit 18c measures the time from when a data read request is issued until the requested data is read and the read transfer rate as the read performance. Here, the measurement unit 18c measures the read performance of the first data and the read performance of the second data periodically after the standby time has elapsed.

  Specifically, the measurement unit 18c secures 2 GB as a measurement area, and reads the 1 GB memory area in which the first data is written and the 1 GB memory area in which the second data is written, respectively. Measure read performance. The measurement unit 18c transfers the measured first data read performance and second data read performance to the determination unit 18d.

  The measurement unit 18c measures the first data read performance and the second data read performance when the first data and the second data are written by the securing unit 18b. Then, the measurement unit 18c determines the first reference value based on the measured read performance of the first data, and determines the second reference value based on the read performance of the second data. Subsequently, the measurement unit 18c sets the first reference value and the second reference value in the reference value storage unit 18a.

  For example, the measurement unit 18c measures the read performance of the first data and the read performance of the second data. Then, the measurement unit 18c determines the first reference value and the second reference value by adding a measurement error to each measurement value. Here, 10% is set as the measurement error, and a value that is 10% lower than the measured value is determined as a reference value. The measurement unit 18c stores the determined reference value in association with the machine address and character string of the first data stored in the reference value storage unit 18a. Note that the measurement unit 18c may measure, as the read performance, either the time from when the data read request is issued until the requested data is read or the read transfer rate. In such a case, the measurement unit 18c causes the reference value storage unit 18a to store the reference value determined based on the measured data value among the data value indicating time or the data value indicating speed.

  The determination unit 18d determines whether the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data. For example, the determination unit 18d determines that the same page is stored in the physical memory when the read performance of the first data is less than a predetermined reference value and the read performance of the second data is equal to or higher than the predetermined reference value. It is determined that it is shared.

  Specifically, the determination unit 18d acquires the reading performance of the first data and the reading performance of the second data from the measurement unit 18c. Further, the determination unit 18d reads out the first reference value and the second reference value from the “reference value” stored in the reference value storage unit 18a. Then, the determination unit 18d compares the first data read performance and the second data read performance with the first reference value and the second reference value, respectively.

  Here, when the reading performance of the second data is equal to or higher than the second reference value and the reading performance of the first data is less than the first reference value, the determination unit 18d determines that the second data is paged out. It is not assumed that the first data is paged out. In other words, when the read performance is measured, the second data that is a fixed character string is not paged out even though the first data and the second data are accessed at the same frequency. 18d infers that the second data is shared on the physical memory and accessed by the other VM execution units 16 and 17. Therefore, in such a case, the determination unit 18d determines that the same page is shared in the physical memory. That is, the determination unit 18d determines that KSM is enabled. If the reading performance of the second data is equal to or higher than the second reference value and the reading performance of the first data is less than the first reference value, the first data is accidentally selected as a page-out target. Sometimes it is done. The processing of the determination unit 18d in such a case will be described later.

  In addition, when the first data read performance is equal to or higher than the first reference value and the second data read performance is lower than the second reference value, the determination unit 18d pages out the second data. And the first data is estimated not to be paged out. In other words, when the read performance is measured, the first data is not paged out even though the first data and the second data are accessed at the same frequency, but the second data is paged out. Therefore, the determination unit 18d guesses that the second data is accidentally selected as a page-out target. For this reason, in such a case, the determination unit 18d determines that it cannot be determined that KSM is currently enabled. In such a case, it is considered that the second data is not shared on the physical memory and is unlikely to be accessed by other VM execution units 16 and 17.

  In addition, when the reading performance of the first data is less than the first reference value and the reading performance of the second data is less than the second reference value, the determination unit 18d also sets the first data to the second value. I guess the data is also paged out. In such a case, the determination unit 18d determines that it cannot be determined that KSM is currently enabled. In such a case, even if KSM is enabled and the second data is shared, the second data is paged out, so the physical memory may be tight.

  In addition, when the first data read performance is equal to or higher than the first reference value and the second data read performance is equal to or higher than the second reference value, the determination unit 18d also sets the first data to the second reference value. It is determined that the data is not paged out and it cannot be determined that KSM is currently enabled.

  FIG. 5 is a diagram for explaining the processing operation performed by the VM execution unit 18 according to the first embodiment. The horizontal axis in FIG. 5 indicates the memory space, and the vertical axis indicates time. In FIG. 5, it is assumed that the measurement unit 18c measures the reading speed as the reading performance after the waiting time has elapsed at regular intervals.

  As shown in FIG. 5, in the VM execution unit 18, the securing unit 18b writes the fixed character string in the area A (step S1). Subsequently, the measuring unit 18c reads the fixed character string written in the area A and measures the reading speed (step S2). Further, the securing unit 18b writes a random number sequence in the region B (step S3). Subsequently, the measuring unit 18c reads the random number sequence written in the region B and measures the reading speed (step S4). Then, the measurement unit 18c enters a standby time during which access is not performed (step S5). The waiting time here is 1 minute.

  After the standby time elapses, the measuring unit 18c reads the fixed character string written in the area A, measures the reading speed (step S6), reads the random number sequence written in the area B, and measures the reading speed (step S6). Step S7). Then, the determination unit 18d determines whether or not KSM is enabled based on the reading speed measured by the measurement unit 18c. Here, when the determination unit 18d determines that KSM is not enabled, the measurement unit 18c enters a standby time during which access is not performed (step S8). The waiting time here is also 1 minute.

  After the standby time elapses, the measuring unit 18c reads the fixed character string written in the area A, measures the reading speed (step S9), reads the random number sequence written in the area B, and measures the reading speed ( Step S10). Then, the determination unit 18d determines whether or not KSM is enabled based on the reading speed measured by the measurement unit 18c. Here, when the determination unit 18d determines that KSM is not enabled, the measurement unit 18c enters a standby time during which access is not performed. In such a case, the waiting time is likewise 1 minute. As described above, the measuring unit 18c sets the waiting time after measuring the reading performance of the first data and the reading performance of the second data to a constant interval (1 minute).

  Next, a detection process procedure performed by the VM execution unit 18 will be described with reference to FIG. FIG. 6 is a flowchart illustrating a procedure of detection processing by the VM execution unit 18 according to the first embodiment. For example, the VM execution unit 18 executes the process when receiving an instruction to start the detection process from the user.

  As shown in FIG. 6, the securing unit 18b secures a plurality of measurement areas in the physical memory (step S101). Then, the securing unit 18b writes a fixed character string in one of the secured measurement areas and writes a random number sequence in the other (step S102).

  Subsequently, the measuring unit 18c reads out each area, and measures a read delay and a read transfer rate (step S103). Moreover, the measurement part 18c sets a reference value based on a measurement result (step S104).

  Then, the measurement unit 18c enters a standby time (step S105) and determines whether or not a predetermined time has elapsed (step S106). Here, when it is determined that the predetermined time has elapsed (Yes in Step S106), the measurement unit 18c reads out each area, and measures the read delay and the read transfer rate (Step S107). In addition, the measurement part 18c repeats the determination process of step S106, when it determines with predetermined time not having passed (step S106, No).

  The determination unit 18d determines whether or not the read performance of the first data is less than the first reference value and the read performance of the second data is equal to or higher than the second reference value. In other words, the determination unit 18d determines whether only the read performance of the first data is less than the reference value (step S108). Here, when the determination unit 18d determines that only the first data read performance is not less than the reference value (step S108, No), it is determined that it is currently unknown whether or not KSM is enabled. (Step S109). In this case, in the VM execution unit 18 for measurement, the measurement unit 18c moves to step S105 and enters a standby time.

  On the other hand, when the determination unit 18d determines that only the read performance of the first data is less than the reference value (Yes in step S108), the determination unit 18d determines that KSM is activated (step S110). The VM execution unit 18 ends the detection process after step S110. When the VM execution unit 18 determines that KSM is enabled, the VM execution unit 18 may notify the user that KSM is enabled.

  Even if KSM is not activated, the first data may be paged out and the second data may not be paged out. In such a case, the determination unit 18d determines that only the read performance of the first data is less than the reference value. That is, the determination unit 18d may make an erroneous determination. For this reason, when the determination unit 18d determines that only the first data read performance is less than the reference value (step S108, Yes), the VM execution unit 18 does not end the detection process. You may make it transfer to step S105 and repeat a detection process. In this case, for example, after it is determined that only the read performance of the first data is less than the reference value, the VM execution unit 18 repeatedly executes the detection process a predetermined number of times, and only the read performance of the first data is obtained. If it is continuously determined that the value is less than the reference value, it is determined that KSM is activated. As a result, the VM execution unit 18 can prevent accidental misjudgment and more accurately determine whether or not KSM is activated.

  As described above, in the information processing apparatus 10 according to the first embodiment, the VM execution unit 18 secures a measurement area in the physical memory, and stores a random number sequence for the first area of the measurement area. 1 is written, and a fixed character string is written as second data to the second area of the measurement area. In addition, the VM execution unit 18 reads the first data and the second data from the measurement area at a predetermined interval, and measures the read performance of the first data and the read performance of the second data. Then, the VM execution unit 18 determines whether or not the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data. For example, if the read performance of the first data is less than the first reference value and the read performance of the second data is greater than or equal to the second reference value, the VM execution unit 18 displays the same page. It is determined that the physical memory is shared. Thereby, the information processing apparatus 10 according to the first embodiment can determine whether or not a plurality of pages having the same contents are combined into one. In other words, the information processing apparatus 10 according to the first embodiment can determine whether or not KSM is activated.

  In addition, when KSM is enabled, each VM executes various processes after the VM is started, so that a page having the same content is changed, and a new memory area is allocated to the changed page. It is done. That is, when KSM is enabled, the shared memory area may decrease according to the VM startup time. Here, in the conventional technology, when the management subject of the service providing system such as the host OS or the hypervisor is different from the user of the virtual machine, the user of the virtual machine identifies whether or not KSM is enabled. Can not do it. In such a case, the user of the virtual machine could not predict the VM performance degradation even if the free memory is tight and the resources are insufficient.

  Specifically, when resources are insufficient, CPU resources are shared among a plurality of VMs, and VM processing performance is lower than normal. Further, when the resources are seriously short, since the CPU is shared in a time division manner, the processing performance of the VM is further deteriorated due to the frequently occurring switching overhead. Further, in a state where resources are insufficient, the frequency of copying between the swap area and the physical memory increases as the switching frequency of the VM corresponding to the time division sharing of the CPU increases. In the swap area, data reading and data writing are slower than the physical memory. Therefore, the memory access time viewed from the VM becomes longer with the switching frequency of the VM. That is, the memory access speed viewed from the VM is greatly reduced.

  On the other hand, in the information processing apparatus 10 according to the first embodiment, it is possible to determine whether or not KSM is activated. As a result, the user of the first VM can predict the possibility of a memory shortage in the future. Further, the user of the first VM can avoid continuing to use the VM in a state where the performance is deteriorated when the resource is insufficient.

(Second Embodiment)
Although the embodiments of the present invention have been described so far, the present invention may be implemented in other embodiments besides the above-described embodiments. Therefore, other embodiments will be described below.

(System configuration)
In the first embodiment, the case where the user is a service user contracted with an IaaS provider has been described. However, the present invention is not limited to this. For example, it is also possible to use the server device on a trial basis before making a contract with an IaaS provider. In such a case, for example, the user can select an optimum IaaS provider by determining whether KSM for services provided by a plurality of IaaS providers is enabled.

  Further, for example, when 2 GB is secured as a measurement area, the securing unit 18 b divides 2 GB into two equal parts, one 1 GB is used as a first data writing area, and the other 1 GB is used as a second data writing area. Although described as what is allocated, it is not limited to this. For example, when 2 GB is secured as the measurement area, the securing unit 18 b may divide 2 GB into four equal parts. In this case, the securing unit 18b writes different first data (referred to as first data A and first data B) to each of the two 0.5 GB memory areas, and each of the two 0.5 GB memory areas. Are written with different second data (referred to as second data A and second data B). Then, each time the reading performance is measured, the measuring unit 18c may read data from four areas, or may read data from two areas. When reading data from two areas, for example, the measurement unit 18c alternately executes reading of the first data A and the second data A and reading of the first data B and the second data B. .

  In the first embodiment, the measurement unit 18c has been described as assuming that the standby time is a constant interval, and the read performance is measured after the standby time has elapsed. However, the measurement unit 18c is not limited thereto. For example, the measurement unit 18c may increase the standby time after measuring the read performance of the first data and the read performance of the second data exponentially according to the number of times of measurement of the read performance. In other words, the measurement unit 18c lengthens the predetermined interval exponentially according to the number of reading performance measurements. FIG. 7 is a diagram for explaining the processing operation performed by the VM execution unit 18 according to the second embodiment. The horizontal axis in FIG. 7 indicates the memory space, and the vertical axis indicates time.

As shown in FIG. 7, in the VM execution unit 18, the securing unit 18b writes a fixed character string in the area A (step S31). Subsequently, the measuring unit 18c reads the fixed character string written in the area A and measures the reading speed (step S32). Further, the securing unit 18b writes a random number sequence in the area B (step S33). Subsequently, the measuring unit 18c reads the random number sequence written in the region B and measures the reading speed (step S34). Then, the measurement unit 18c enters a standby time during which access is not performed (step S35). It is assumed here 1 minute waiting time (= ^{2 0} min).

After the standby time elapses, the measuring unit 18c reads the fixed character string written in the area A, measures the reading speed (step S36), reads the random number sequence written in the area B, and measures the reading speed ( Step S37). Then, the determination unit 18d determines whether or not KSM is enabled based on the reading speed measured by the measurement unit 18c. Here, when the determination unit 18d determines that KSM is not enabled, the measurement unit 18c enters a standby time during which access is not performed (step S38). The waiting time here is 2 minutes (= 2 ¹ minutes).

After the standby time elapses, the measuring unit 18c reads the fixed character string written in the area A, measures the reading speed (step S39), reads the random number sequence written in the area B, and measures the reading speed ( Step S40). Then, the determination unit 18d determines whether or not KSM is enabled based on the reading speed measured by the measurement unit 18c. Here, when the determination unit 18d determines that KSM is not enabled, the measurement unit 18c enters a standby time during which access is not performed (step S41). The waiting time here is 4 minutes (= ²² minutes).

After the standby time elapses, the measuring unit 18c reads the fixed character string written in the area A, measures the reading speed (step S42), reads the random number sequence written in the area B, and measures the reading speed (step S42). Step S43). Then, the determination unit 18d determines whether or not KSM is enabled based on the reading speed measured by the measurement unit 18c. Here, when the determination unit 18d determines that KSM is not enabled, the measurement unit 18c enters a standby time during which access is not performed. In such a case, the standby time is 8 minutes (= ²³ minutes). As described above, the measuring unit 18c increases the standby time after measuring the read performance of the first data and the read performance of the second data exponentially as the number of times of measurement of the read performance is repeated. As described above, in the example illustrated in FIG. 7, when the number of measurements counted after measuring the reference value is n, the standby time is 2 ⁿ⁻¹ minutes.

  Further, the measurement unit 18c decreases the access frequency to the first data and the second data by measuring the read performance by increasing the standby time exponentially as the read performance measurement is repeated. Can be made. This increases the probability that the first data will be paged out, and more accurately determines whether or not KSM is enabled.

  Further, when the read transfer rate is measured, the determination unit 18d may determine whether or not KSM is enabled using the read transfer rate. For example, when KSM is not activated, it is assumed that the time required for read transfer from a 1 GB memory area is 5 (seconds), for example. Here, when the second data is written in the 1 GB memory area so that the same page is repeated in units of 4 KB, it is only necessary to transfer the data for one page. Milliseconds). For this reason, the determination unit 18d determines that KSM is enabled when the read transfer rate of the second data is significantly higher than the second reference value. In addition, since the securing unit 18b equally divides the measurement area and writes the first data and the second data, the first reference value and the second reference value are approximately approximate values. Therefore, the determination unit 18d can determine that KSM is enabled even when the read transfer rate of the second data is significantly higher than the read transfer rate of the first data.

  When the first reference value and the second reference value are approximately approximate, the measurement unit 18c is based on the measured average value of the first data read performance and the second data read performance. Thus, the first reference value and the second reference value may be determined. In such a case, the first reference value and the second reference value are the same.

  Also, among the processes described in this embodiment, all or part of the processes described as being performed automatically can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, control procedures, specific names, and information including various data and parameters shown in the above-mentioned documents and drawings (for example, FIGS. 1 to 7) are arbitrary unless otherwise specified. Can be changed.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or a part of the distribution / integration may be functionally or physically distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured.

(program)
FIG. 8 is a diagram showing that information processing by a detection program for executing processing by a computer system is specifically realized using a computer. As illustrated in FIG. 8, the computer 3000 includes, for example, a memory 3010, a CPU 3020, and a network interface 3070. Each part of the computer 3000 is connected by a bus 3100.

  The memory 3010 includes a ROM 3011 and a RAM 3012 as illustrated in FIG. The ROM 3011 stores a boot program such as BIOS (Basic Input Output System).

  Here, as illustrated in FIG. 8, the hard disk drive 3080 stores, for example, an OS 3081, an application program 3082, a program module 3083, and program data 3084. In other words, the detection program according to the disclosed technology is stored in, for example, the hard disk drive 3080 as the program module 3083 in which instructions executed by the computer are described. More specifically, the hard disk drive 3080 stores a program module 3083 in which each procedure for executing the same information processing as each unit of the VM execution unit 18 described in the above embodiment is described.

  Data used for information processing by the detection program is stored as program data 3084 in, for example, the hard disk drive 3080. The CPU 3020 reads the program module 3083 and program data 3084 stored in the hard disk drive 3080 to the RAM 3012 as necessary, and executes various procedures.

  Note that the program module 3083 and the program data 3084 related to the detection program are not limited to being stored in the hard disk drive 3080. For example, the program module 3083 and the program data 3084 may be stored in a removable storage medium. In this case, the CPU 3020 reads data via a removable storage medium such as a disk drive. Similarly, the program module 3083 and the program data 3084 related to the detection program may be stored in another computer connected via a network (LAN (Local Area Network), WAN (Wide Area Network), etc.). . In this case, the CPU 3020 reads various data by accessing another computer via the network interface.

(Other)
Note that the detection program described in the present embodiment can be distributed via a network such as the Internet. The detection program can also be executed by being recorded on a computer-readable recording medium such as a hard disk, a flexible disk (FD), a CD-ROM, an MO, or a DVD, and being read from the recording medium by the computer.

1, 1a, 1b Client terminal 10 Information processing device 11 Communication control unit 12 Swap area 13 Storage unit 13a Physical memory management DB
13b Page-out management DB
14 control unit 14a VM management unit 14b physical memory control unit 14c paging control unit 14d KSM control unit 15 virtual region 16, 17, 18 VM execution unit 18a reference value storage unit 18b securing unit 18c measurement unit 18d determination unit

Claims (6)

In an information processing apparatus that realizes a plurality of virtual machines and allocates a part of physical memory to the memory of each virtual machine,
The first virtual machine is
A measurement area is secured in the physical memory, a random number sequence is written as first data to the first area of the measurement area, and a fixed character string is written to the second area of the measurement area A securing unit for writing as second data,
A measurement unit that reads the first data and the second data from the measurement area at predetermined intervals, and measures the read performance of the first data and the read performance of the second data;
A determination unit that determines whether or not the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data. apparatus.   The determination unit has the same page when the first data read performance is less than a first reference value and the second data read performance is equal to or higher than a second reference value. The information processing apparatus according to claim 1, wherein the information processing apparatus determines that the physical memory is shared. The measurement unit measures the read transfer rate as the data read performance,
The determination unit determines that the same page is shared in the physical memory when a read transfer rate of the second data significantly exceeds a read transfer rate of the first data. The information processing apparatus according to claim 1 or 2.   The information processing apparatus according to claim 1, wherein the measurement unit lengthens the predetermined interval exponentially according to the number of reading performance measurements. The measuring unit measures and measures the reading performance of the first data and the reading performance of the second data when the first data and the second data are written by the securing unit. The first reference value is determined based on the read performance of the first data, and the second reference value is determined based on the read performance of the second data, and the first reference value and The information processing apparatus according to claim 2 , wherein the second reference value is set in the storage unit. A detection method executed by an information processing apparatus that realizes a plurality of virtual machines and allocates a part of physical memory to the memory of each virtual machine,
The first virtual machine is
A measurement area is secured in the physical memory, a random number sequence is written as first data to the first area of the measurement area, and a fixed character string is written to the second area of the measurement area Securing process for writing as second data;
A measurement step of reading the first data and the second data from the measurement area at a predetermined interval to measure the read performance of the first data and the read performance of the second data;
And a determination step for determining whether the same page is shared in the physical memory based on the read performance of the first data and the read performance of the second data. Method.

Images