JP2011113267A - Data center configuration management system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data center configuration management system that easily and quickly defines proper system configuration according to a state of a load, in a virtual data center in a cloud environment. <P>SOLUTION: The data center configuration management system has: a configuration storage part 110 which acquires snapshots 101 of system configuration of a data center and throughputs at the time when the data center operates, to store information on the snapshots 101 and information on the throughputs in association with each other in a configuration DB 102; a configuration retrieval part 120 which retrieves a snapshot 101 having proper system configuration from the configuration DB 102 on the basis of estimation states of the throughputs in the data center; and a configuration-altering part 130 which constructs the data center by using a programming interface, or alters the system configuration of the data center on the basis of the system configuration of the retrieved snapshot 101. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technology related to a data center that provides a service by a computer system, and is particularly effective when applied to a data center configuration management system that manages the configuration of a virtual server or the like in a data center virtualized by cloud computing. It is about technology.

  When various services are provided to users via the Internet or the like by a computer system, a data center that operates and manages a large number of devices such as servers and storages related to the provision of services is often used. ing. Usually, in data centers, using a lot of equipment and operation management software, etc., very complicated procedures and tools such as file transfer, OS (Operating System) command execution, application startup and setting are performed. It is used and operated.

  On the other hand, in recent years, due to the evolution of cloud computing technology, for example, “Amazon EC2 (registered trademark)” (Non-Patent Document 1), the data center itself can be virtualized without being aware of physical facilities and equipment. Can be handled. In such a data center, a user can start up or add an instance (virtual server including an OS) using an API (Application Programming Interface) or a command using the API. Access to the data center will be converted into APIs, and by using this from a software program, there is a possibility that operation and control with cooperation between servers, which has been difficult with conventional data centers, can be highly automated. There is.

“Amazon Elastic Compute Cloud (Amazon EC2)”, [online], Amazon Web Services LLC, Internet <URL: http://aws.amazon.com/ec2/>

  In a data center using a cloud technology such as “Amazon EC2”, the user is charged mainly by the number of instances that are activated and the usage time. Therefore, if an instance can be activated only when necessary, the cost is low. However, in an operation in which the instance is always activated, the cost may be higher than that of a conventional data center. Starting an instance can usually be done in a short time of several minutes, so it is desirable to perform detailed control according to the load when considering cost reduction.

  On the other hand, when fine control of starting and stopping of instances is performed, it is quite expensive to manually control the deployment of running applications, setting the status of services, and distributing the appropriate number of instances for each function and service. Cost. In addition, it is technically difficult to completely adjust the number and configuration of servers to be activated according to load conditions by autonomous control.

  Accordingly, an object of the present invention is to provide a data center configuration management system capable of easily and quickly constructing an appropriate system configuration according to the load status in a virtual data center in a cloud environment. The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

  Of the inventions disclosed in this application, the outline of typical ones will be briefly described as follows.

  A data center configuration management system according to a representative embodiment of the present invention is a data center configuration that changes a system configuration according to a predicted service load status for a virtual data center having a programming interface in a cloud environment. A management system having the following characteristics.

  That is, the data center configuration management system acquires a snapshot of the system configuration of the data center and the throughput during operation of the data center, and associates the snapshot information with the throughput information in the configuration database. A configuration storage unit to store, a configuration search unit to search for the snapshot having a system configuration that can be processed from the configuration database based on a predicted state of throughput in the data center, and the configuration search unit to search for the snapshot The data center is constructed using the programming interface based on a snapshot system configuration, or has a configuration changing unit that changes the system configuration of the data center.

  Among the inventions disclosed in the present application, effects obtained by typical ones will be briefly described as follows.

  According to a typical embodiment of the present invention, a snapshot of the system configuration of the data center is prepared in advance in association with throughput information during operation for configurations with proven operation and verified configurations. As a result, a mechanism that enables administrators to quickly select the optimal snapshot according to the load forecast status and easily adjust the data center system configuration based on the selected snapshot is realized at low cost. It becomes possible to do.

It is the figure which showed the outline | summary of the example of a structure of the data center structure management system which is one embodiment of this invention. It is the figure which showed the example of the structure pattern of the data center holding the information of the snapshot in one embodiment of this invention. It is the figure which showed the example of the data structure of the structure DB in one embodiment of this invention, and specific data.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiment, and the repetitive description thereof will be omitted.

  A data center configuration management system according to an embodiment of the present invention has a system configuration according to a load prediction situation for a virtual data center in a cloud environment having an API such as “Amazon EC2” described above. It is a system that realizes a mechanism that can be easily changed and adjusted.

  For example, in “Amazon EC2”, an AMI (Amazon Machine Image) that is a configuration image of an instance (virtual server) can be stored in a virtual storage of “Amason S3” that is a storage service in a cloud environment. Also, a desired one can be selected from the AMIs held in “Amason S3”, and an instance can be activated on “Amason EC2” using the selected AMI. “Amazon EC2” provides a so-called Web API for performing such instance management (starting / stopping of instances, resource management), storage management, and inquiries from client terminals.

  By using such an API, it is considered that the operation of a virtual data center in a cloud environment can be highly automated by applying a programming technique as described above. However, it is difficult to fully automate and adjust the number and configuration of virtual servers according to the predicted load status. As this factor, for example, there are predetermined rules in the system configuration (for example, the ratio of the number of servers for each type of server or the determination of the configuration of an even number of servers), which must be followed. . In addition, there is already a technology to predict the load automatically to some extent, but there is a limit in cases such as sudden occurrence of events, and there is often a need for some kind of judgment in the end. Also mentioned.

  Therefore, in the data center configuration management system according to the present embodiment, for example, each virtual server or the like for several configurations such as a configuration that has been used in the production environment in the past and a configuration in which processing performance has been verified by a performance test. The configuration information including information about the configuration image (such as AMI) is held as a snapshot. As an example of the snapshot, for example, a configuration corresponding to a characteristic load situation such as “a configuration that can withstand the highest traffic ever” and “a configuration that is assumed immediately after television broadcasting” can be considered. By preparing such a characteristic configuration in advance, for example, an administrator or a CIO (Chief Information Officer) can make a quick decision.

  One of the major advantages of a virtual data center in a cloud environment is that a test environment of various levels and configurations including a scale equivalent to that of a production environment can be flexibly constructed. Therefore, for example, a test environment equivalent to the production environment can be temporarily constructed, and a snapshot related to the configuration when a sufficient verification is performed by performing a performance test or the like can be obtained and held.

  By using the data center configuration management system of the present embodiment, for example, in a data center that provides a financial transaction service, it is expected that the load on a specific service will increase from the trend of the New York market on the previous day. Control the number of servers and adjust the balance of the number of servers by selecting a snapshot that can be handled according to the load status and changing the data center system configuration based on the selected snapshot. Can be performed easily, quickly and at low cost.

  In addition, for example, if the service is usually provided at the data center operated by the company, but sudden increase in load is expected, such as news of a major company's management failure, "Amazon EC2" etc. It is possible to temporarily control a virtual data center based on a cloud environment of an external vendor to increase the resources and distribute the load.

[System configuration]
FIG. 1 is a diagram showing an outline of a configuration example of a data center configuration management system according to an embodiment of the present invention. The data center (DC) configuration management system 100 is configured by a computer system such as a server or a PC (Personal Computer), and can access a virtual data center in the cloud 200 using “Amazon EC2” or the like via the Internet or the like. It is a system that can.

  The DC configuration management system 100 includes, for example, a configuration storage unit 110 implemented by a software program, a configuration search unit 120, a configuration change unit 130, and an approval unit 140, a database of the configuration DB 102, and a virtual in the cloud 200. It has a snapshot 101 of the data center system configuration.

  The configuration storage unit 110 further includes a throughput acquisition unit 111 and a snapshot acquisition unit 112, and has a function of acquiring and storing information on the data center system configuration and operation throughput in the cloud 200.

  The throughput acquisition unit 111 has a function of acquiring throughput information during operation from a data center in the cloud 200 (for example, production environment # 1 (201) or test environment # 1 (211) in the figure). Here, the throughput acquired by the throughput acquisition unit 111 is a basic level such as a throughput (for example, a CPU (Central Processing Unit) usage rate) that can be acquired from a virtual data center in the cloud 200 using the above-described API. ), Throughput at the application level running on the virtual server (for example, the number of screen displays per second on the Web server, the number of processing per second on the application (APL) server, 1 on the database (DB) server) Index that serves as a key for determining an appropriate data center configuration, such as the number of SQL processes per second).

  In order to acquire the throughput at the application level, for example, a throughput acquisition program may be operated on the virtual server to acquire the throughput from the program, or the throughput acquisition unit 111 may be on the virtual server. You may acquire by acquiring and analyzing an application log etc.

  The snapshot acquisition unit 112 has a function of acquiring a snapshot of the system configuration of the target data center from which the throughput acquisition unit 111 has acquired the throughput. The acquired snapshot is held in the system as the snapshot 101. Note that the snapshot 101 includes, as system configuration information, information such as the type and number of each virtual server that is running and an ID for a configuration image such as AMI for each virtual server. In addition, the application and service setting information includes information such as the type of application to be activated in each virtual server and the setting of the service state / mode. The snapshot 101 is held as a file or the like on the DC configuration management system 100, for example.

  Further, information of each snapshot 101 (ID etc. for identifying the snapshot 101) is stored in the configuration DB 102 in association with the throughput information of the corresponding data center. In the example of FIG. 1, for example, for the configuration of the production environment # 1 (201), a snapshot is acquired on September 1, 2009, and the throughput associated with the user access at that time is also acquired. Are stored in the snapshot 101 and the configuration DB 102. Similarly, on September 15, 2009, a test environment # 1 (211) is built on the cloud 200 and a performance test is performed. The snapshot and the configuration DB 102 are acquired together with the snapshot and the throughput at that time. Hold on. The contents of the configuration DB 102 will be described later.

  The configuration search unit 120 is suitable (can be processed) according to the situation of the load (throughput) expected in the data center by the cloud 200 from the snapshot 101 held in the configuration DB 102 and the corresponding throughput information. It has a function of searching for a snapshot 101 having a configuration. The search function is implemented as, for example, a Web application that runs on a Web server program (not shown), and is accessed via a network from a Web browser (not shown) on the operation person terminal 300 that is a client terminal such as a data center operator. Is done.

  For example, the operation person or the like predicts the load (throughput) status such as the traffic volume and the concentration level of the type based on the market trend of the previous day, various news, and the like, and stores each snapshot held in the configuration DB 102. An appropriate snapshot 101 is searched and selected based on the throughput information of 101. In addition, for example, when the load prediction is automatically possible (for example, when regularity such as “monthly xx day increases traffic by 30%” is known, or when the load traffic is estimated by the load prediction system etc. When a certain amount of load can be predicted automatically based on the above), an appropriate snapshot 101 may be automatically selected based on the load prediction information.

  The configuration changing unit 130 further includes a configuration correcting unit 131 and a configuration applying unit 132, and applies the configuration of the snapshot 101 searched and selected by the configuration searching unit 120 to the data center in the cloud 200. It has a function to change the configuration. The configuration correction unit 131 corrects the configuration information with respect to the configuration of the snapshot 101 searched and selected by the configuration search unit 120 based on an instruction from an operation person or the like automatically or via the operation person terminal 300. It has a function. For example, the configuration information (such as the IP address) in the snapshot 101 of the test environment cannot be applied as it is to the production environment and may need to be corrected.

  In the present embodiment, the snapshot 101 searched and selected by the configuration search unit 120 may be modified such as increasing or decreasing the number of servers or changing the type. At this time, for example, the configuration search unit 120 searches for the snapshot 101 serving as a base, and the configuration correction unit 131 indicates the number of increase / decrease of the server in accordance with the predicted load status for the configuration of the snapshot 101. It may be calculated automatically.

  The configuration application unit 132 uses the API or the like based on the configuration of the snapshot 101 corrected by the configuration correction unit 131 in the cloud 200 and has a configuration equivalent to that of the data center (for example, test environment # 1 (211)). Environment # 2 (202)) has a function of constructing or changing the configuration.

  As specific processing, for example, an instance necessary for the cloud 200 is started using an API based on the contents of the modified snapshot 101, and a startup script is transferred by logging into the virtual server of each instance. And execute. In the startup script, the provision of the service is started by mounting the virtual storage on the cloud 200, starting the stored application, and starting the service. When the configuration is reduced, an unnecessary instance is also stopped.

  It should be noted that the snapshot 101 searched and selected by the configuration search unit 120 is not corrected by the operation person or the like by the configuration correction unit 131 and applied as it is by the configuration application unit 132, but is applied or changed by an administrator or CIO. You may make it apply after obtaining approval. At this time, for example, the configuration correcting unit 131 does not allow the modified snapshot 101 to be transferred to the configuration applying unit 132 unless it has been approved by the approval unit 140.

  The approval unit 140 has a function of accepting application approval for the modified snapshot 101 from the administrator terminal 400 which is a client terminal such as an administrator. For the approval, for example, a mechanism such as a one-time password is used so that only an authorized person such as an administrator can perform the approval process. Note that the administrator terminal 400 is not necessarily located at the same site as the DC configuration management system 100, so it is desirable to enable access using a mobile terminal such as a mobile phone as well as a terminal such as a PC.

  As described above, in the DC configuration management system 100 according to the present embodiment, in the data center using the cloud 200, for example, the configuration storage unit 110 performs throughput and snapshot 101 for several configurations that are characterized by a load situation. And the information of the snapshot 101 is stored in the configuration DB 102 in association with the throughput. After the person in charge of operation searches and selects an appropriate snapshot 101 based on the load prediction status from the information of the snapshot 101 held in the configuration DB 102, and obtains approval from the administrator by the approval unit 140 Then, the configuration change unit 130 automatically changes the configuration of the data center of the cloud 200 using an API or the like based on the configuration information of the snapshot 101 searched and selected.

[Snapshot configuration pattern]
FIG. 2 is a diagram illustrating an example of a configuration pattern of a data center that holds information of the snapshot 101. The upper diagram shows an example of the minimum configuration in the data center using the cloud 200. In the example of FIG. 2, the data center operates a web system having a so-called three-tier configuration including a web server, an application (APL) server, and a database (DB) server, and in the minimum configuration, there are six web servers. , It shows that the APL server has three units and the DB server has one unit.

  The middle diagram shows a configuration example in which each server is augmented with respect to the minimum configuration, and the APL server and the DB server are separated into, for example, a business system and an information system to enhance the processing capability. In this business system / information system separation configuration, the number of Web servers is increased to nine, the number of APL servers is increased to three to separate the business system and the information system, and similarly, the number of DB servers is increased to one business system. Separated into information systems. In this way, instead of simply increasing the number of units to increase the minimum configuration, for example, prepare a configuration that separates and separately increases the APL server and DB server for each processing system according to the characteristics of the load. It is possible to keep.

  The lower diagram shows a configuration example when the Web server is augmented with respect to the minimum configuration and the processing capability for content reference is enhanced. In this reference strengthening configuration, the number of Web servers is doubled to 12 units. This configuration can be selected when it is assumed that there are many users who refer to static content, for example, after a new product is announced. In this way, it is possible to prepare a configuration that individually increases for each type of server according to the load characteristics.

[Data structure]
FIG. 3 is a diagram illustrating a data configuration of the configuration DB 102 and an example of specific data. The configuration DB 102 is a database that holds the information of the snapshot 101 of the configuration of the data center in association with the information of the throughput of the corresponding configuration. For example, the configuration ID, configuration name, snapshot, production operation throughput, test throughput , Each item such as remarks.

  The items of the configuration ID and the configuration name hold an ID and a name for identifying the configuration of the data center related to the snapshot 101, respectively. The configuration ID item is automatically numbered by the configuration storage unit 110, for example. In the item of the configuration name, an operation person or the like can set a name that makes it easy to recognize the characteristics of each configuration.

  The snapshot item is information (snapshot ID) for identifying the target snapshot 101. In the example of FIG. 3, it is possible to specify the date (for example, “January 1, 2008” in the case of “minimum configuration”) when the snapshot 101 (and throughput) is acquired by the content of the identification information itself. However, the date when the snapshot 101 is acquired may be included as an item in the configuration DB 102 separately. In the present embodiment, the snapshot 101 is stored as a file or the like in the DC configuration management system 100 and the identification information is stored in the configuration DB 102. However, the snapshot 101 itself may be stored in the configuration DB 102. .

  The items of the throughput at the time of production operation and the throughput at the time of test are respectively the measured value of the throughput when the data center configuration related to the target snapshot 101 is operated in the production environment and the throughput when the performance test is performed in the test environment. Holds measured value information.

  Here, in the example of FIG. 3, the throughputs are PV / s (Page View / sec: the number of screen displays per second by the Web server) and Trans / s (Transaction / sec: per second by the APL server). And three types of indexes are set: SQL / sec (SQL / sec: number of SQL processes per second by the DB server). In this embodiment, the PV / s, Tran / s, and SQL / s indexes are based on the assumption that they depend on the number of Web servers, APL servers, and DB servers, respectively. The number of servers and the like can be determined, but other indicators may be used as long as they are indices that can determine the number of servers and the like. In addition to these indices, for example, indices such as the CPU usage rate (system margin) and the average value / maximum value of TAT (Turn Around Time) may be further set.

  The item of remarks holds a description explaining the contents of each configuration (snapshot 101). With this information, the person in charge of operation can grasp the characteristics and the like of each snapshot 101, so that the search becomes easy.

  In the example of FIG. 3, no value is set in the item of “throughput during production operation” for the configuration of “enhancement of information system”, but this is confirmed by a performance test in the test environment. This indicates that the configuration has not yet been operated in a production environment.

  When the configuration retrieval unit 120 retrieves an appropriate configuration (snapshot 101) from the configuration DB 102 as illustrated in FIG. 3 according to the load prediction status, the operator in charge uses PV / s and PV / s as load predictions. Each of the index values of Tran / s is predicted and instructed to the configuration search unit 120. The configuration search unit 120 presents a search result having a configuration that clears all the instructed PV / s and Tran / s values. At this time, for example, the search results may be presented in priority order in ascending order of charges to be charged, and the person in charge may select from among them, or the structure in which the charges to be charged are the lowest, etc. May be selected automatically.

  As described above, according to the DC configuration management system 100 according to an embodiment of the present invention, in a virtual data center in the cloud 200 having an API, for example, there has been a past operation record in a production environment. The configuration storage unit 110 obtains the throughput and the snapshot 101 for some configurations with characteristic load conditions, such as the configuration and the configuration whose processing performance has been verified by the performance test. And stored in the configuration DB 102.

  Furthermore, from the information of the snapshot 101 held in the configuration DB 102, an operation person or the like searches for and selects an appropriate snapshot 101 based on the predicted load state, and configures the configuration based on the searched / selected configuration information. The changing unit 130 automatically changes the configuration of the data center of the cloud 200 using an API or the like.

  In this way, by preparing in advance the configuration with a characteristic load situation in the data center as a candidate, the operation staff, administrator, CIO, etc. can quickly and optimally snapshot according to the load forecast situation. 101 can be selected. In addition, a mechanism that can easily change and adjust the system configuration of the data center based on the snapshot 101 selected according to the predicted load state can be realized at low cost.

  As mentioned above, the invention made by the present inventor has been specifically described based on the embodiment. However, the present invention is not limited to the embodiment, and various modifications can be made without departing from the scope of the invention. Needless to say.

  The present invention is applicable to a data center configuration management system that manages the configuration of a virtual server or the like in a data center virtualized by cloud computing.

DESCRIPTION OF SYMBOLS 100 ... Data center (DC) configuration management system, 101 ... Snapshot, 102 ... Configuration DB, 110 ... Configuration storage part, 111 ... Throughput acquisition part, 112 ... Snapshot acquisition part, 120 ... Configuration search part, 130 ... Configuration change , 131 ... configuration correction unit, 132 ... configuration application unit, 140 ... approval unit,
200 ... Cloud, 201 ... Production environment # 1, 202 ... Production environment # 2, 211 ... Test environment # 1,
300 ... operation staff terminal,
400: Administrator terminal.

Claims (6)

A data center configuration management system that changes a system configuration according to a predicted service load status for a virtual data center having a programming interface in a cloud environment,
A configuration storage unit for acquiring a snapshot of the system configuration of the data center and throughput during operation of the data center, and storing the snapshot information and the throughput information in a configuration database;
A configuration retrieval unit that retrieves the snapshot having a system configuration that can be processed from the configuration database, based on a predicted state of throughput in the data center;
A configuration change unit configured to construct the data center using the programming interface or to change the system configuration of the data center based on the system configuration of the snapshot searched by the configuration search unit; Data center configuration management system. In the data center configuration management system according to claim 1,
The snapshot includes a configuration image of a server operating in the data center, information on the type and number of servers, and setting information on an application or service operating on the server. . In the data center configuration management system according to claim 1 or 2,
The data center configuration management system according to claim 1, wherein the throughput includes an application or a service running on a server of the data center. In the data center configuration management system according to claim 3,
The throughput includes at least one of the number of screen display times per unit time by the Web server in the data center, the number of processing cases per unit time by the application server, and the number of SQL processing cases per unit time by the DB server. A data center configuration management system characterized by that. In the data center configuration management system according to any one of claims 1 to 4,
The configuration change unit corrects the system configuration of the snapshot searched by the configuration search unit, and then constructs the data center or changes the system configuration of the data center. Center configuration management system. In the data center configuration management system according to any one of claims 1 to 5,
And an approval unit that receives approval for the construction of the data center or the change of the system configuration of the data center based on the system configuration of the snapshot searched by the configuration search unit by the configuration change unit. A data center configuration management system characterized by that.