JP2012069056A - Cloud service relocation system and method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable services and IT resources in a cloud computing environment to be optimally allocated in a distributed manner.SOLUTION: A server meter 102 monitors a processing operation of each server resource 101 to collect load information and stores policy definition information 108, which is predetermined for each SaaS, containing at least a function condition, price condition, environmental performance condition, evaluation condition, and priority of each condition which are required to provide a SaaS. A locator processing part 112 checks the load information for each server resource, which is collected by the server meter, against condition information contained in the stored policy definition; extracts server resources that satisfy the conditions; and determines one server resource from the extracted server resources that is most appropriate to a condition having the highest priority in the policy definition information that is predetermined for the SaaS. A relocation processing part 110 relocates the SaaS to the server resource determined by the locator processing part.

Description

  The present invention relates to an operation technology of a computer processing service used in a cloud computing environment, and is particularly suitable for optimizing the arrangement of the service in an IT resource (computer device, software application) that provides the service. It is about technology.

  Cloud computing is a technology for using each IT resource as a service through a network, rather than owning IT resources of various layers such as software, middleware, and hardware.

  This cloud computing (hereinafter simply referred to as “cloud”) facilitates the ownership of IT resources with reduced initial investment and the capacity control of IT resources when the amount of requests increases.

  Services provided in the cloud (cloud services) are SaaS (Software as a Service), PaaS (Platform as a Service) and IaaS (Infrastructure as a Service) / Haas (Hardware as a Service) It is classified into three types.

  SaaS uses a software application as a service, PaaS uses a software application development environment, execution environment, platform for operation / monitoring as a service, and IaaS / Haas Uses server resources (server, HDD, network, etc.) as a service.

  Also, SaaS, PaaS, and IaaS that are only disclosed to a limited extent within a company are called private clouds, and SaaS, PaaS, and IaaS that can be shared with many users via the Internet are called public clouds. .

  Portability and dispersibility are the most important features of IT resources deployed on the cloud. That is, the IT resource on the cloud can be operated by freely selecting any IT resource existing in the lower layer. For example, a server resource that operates SaaS can be selected from any IaaS, and can be switched to another IaaS according to a request such as price and performance.

  When operating SaaS or a general application, it is extremely difficult to estimate (optimize) the minimum server resources necessary for operation. For example, Patent Document 1, This is proposed in Patent Document 2 and Patent Document 3.

  In these technologies, a certain amount of server resources are prepared in advance, and the problem is solved by taking a method of increasing or decreasing the server resources according to the load. By using these technologies, server resources can be used without waste.

  However, since the absolute amount of server resources to be secured is estimated based on the maximum load, the above-mentioned technologies cannot reduce the maintenance cost of server resources. In addition, it is difficult to cope with a decrease in server resource prices and replacement of server resources when a need for more environmentally friendly server resources is required.

  As a technique for solving such a problem, the above-described cloud computing can be cited. With the realization of this cloud computing, it has become possible to secure server resources necessary for SaaS operations as needed from IaaS existing on the cloud.

  However, the server resource optimization problem at the time of SaaS operation shown below still needs to be handled manually by the operator, and it becomes a very difficult task as the choice of server resources increases due to the IaaS. ing.

  That is, (1) when expanding server resources when the load of SaaS increases, or when shifting to high-performance server resources, (2) reducing server resources when the load of SaaS decreases, When migrating to a low-cost but inexpensive server resource, (3) If the server has the same performance (performance / security), select a server resource with another evaluation axis (price / environmental performance) In any of the cases (1) to (3), it is necessary to select a server resource that can satisfy the SaaS operation policy (performance / security), which is a very difficult task.

JP 2001-167071 A JP 2008-217285 A JP-A-9-321759

  The problem to be solved is that the conventional technology cannot easily select the IT resource that is truly optimal for the service from the enormous IT resources in the cloud computing environment.

  An object of the present invention is to solve these problems of the prior art, to enable an optimal distribution and arrangement of services and IT resources in a cloud computing environment, and to provide an efficient service.

In order to achieve the above object, the present invention automatically determines a server resource that is truly optimal for the service and rearranges the server resources that provide the service. Specifically, by a programmed computer process, a service provided by cloud computing is rearranged to a server resource that executes the process of the service, and as a function of executing the programmed computer process, It has a navigation processing unit and a relocation processing unit. The navigation processing unit includes a locator processing unit and a reputation database, and stores the following three pieces of information for server resources in the reputation database.
1. Load information of each server resource acquired from a server meter set in advance in each server resource.
2. Server profile information such as the performance and price of each server resource that the server resource provider inputs in advance.
3. Policy definition information that is set by an arbitrary third party and consists of evaluation conditions such as reliability and customer satisfaction for each server resource.
Then, the navigation processing unit inputs policy definition information in which operating conditions (CPU performance condition, memory capacity condition, communication speed condition, etc.) of the cloud service (SaaS) prepared in advance are described in the locator processing unit, Check the policy definition information against the information about each server resource stored in the reputation database, obtain a list of server resources that can satisfy the policy definition, obtain a list of server resources, Select the server resource that best suits the selected selection conditions, and if the selected server resource is different from the server resource that is currently operating SaaS, identify the server resource as a relocation destination and relocate it SaaS information and server resource information specified as relocation destination In addition, to notify the relocation processing unit. The relocation processing unit copies the service (SaaS) from the currently active server resource that is the relocation source to the relocation destination together with the server information, and after the copy is completed, the server information is relocated to the relocation destination server resource. Start the server from, and delete the server information from the relocation source. Through the series of operations described above, the service (SaaS) can be arranged and operated on an appropriate server resource.

  According to the present invention, information such as performance / price required for the cloud service (SaaS) and operating state information of server resources that operate each cloud service (SaaS) are used. It is possible to easily perform efficient optimum distributed arrangement.

It is a block diagram which shows the structural example of the cloud service rearrangement system which concerns on this invention. It is explanatory drawing which shows the example of a table structure in the server profiler (103) set by the operator (104) of FIG. It is explanatory drawing which shows the table structural example of the load information of the server which the server meter (102) of FIG. 1 collected and stored in the reputation database (113). It is explanatory drawing which shows the table structural example of the server evaluation information input from the operator (114) of FIG. 1 and stored in the reputation database (113). It is explanatory drawing which shows the example of a table structure in the policy definition (108) input from the operator (107) of FIG. It is explanatory drawing which shows the example of a table structure of the monitoring information of SaaS which the SaaS meter (109) of FIG. 1 collects. It is a flowchart which shows the 1st processing operation example which concerns on the cloud service rearrangement method of this invention by the cloud service rearrangement system of FIG. It is a flowchart which shows the 2nd processing operation example which concerns on the cloud service rearrangement method of this invention by the cloud service rearrangement system of FIG.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. The cloud service rearrangement system according to the present invention in FIG. 1 includes a server resource 101, an optimal distributed arrangement system 105, and SaaS 106, and each of the server resource 101, the optimal distributed arrangement system 105, and SaaS 106 includes a CPU, main memory, It consists of a computer equipped with a display device, an input device, an external storage device, etc., and after installing a program or data recorded in a storage medium such as a CD-ROM via the optical disc drive device etc. in the external storage device, Each processing is executed by reading from the external storage device into the main memory and processing by the CPU.

  For example, the server resource 101 has a server meter 102 as a function for executing programmed computer processing, and the optimum distributed arrangement system 105 has a relocation processing unit 110 ("Relocation" in the figure as a function for executing programmed computer processing). ”And a navigation processing unit (described as“ navigation unit ”in the figure) 111, and this navigation processing unit 111 is provided with a locator processing unit (described as“ locator ”in the figure) 112 and a reputation database 113. The SaaS 106 has functions of a policy definition 108 and a SaaS meter 109.

  Reference numerals 104, 107, and 114 denote operators who operate the respective devices via input devices. The operator 104 that is a cloud provider provides the server resource 101, and the operator 107 that is a SaaS developer SaaS 106 (cloud service) is deployed on the server resource 101, and the operator 114 as an evaluator adds non-quantitative evaluation information of each server resource to the information in the reputation database 113.

  In addition, the operator (cloud provider) 104 creates a server profiler 103 that shows the performance specification information of the server resource defined for each server resource (102) as a server profiler 103a in FIG. Store in the reputation database 113.

  In this reputation database 113, load information in each server resource (101), which is collected as an example of server meter information 102a in FIG. Evaluation information added by the operator 114 is stored, and each piece of information is stored in the content of reputation information exemplified as server evaluation information 113a in FIG.

  The operator (SaaS developer) 107 inputs a policy definition 108, which is an example of the policy definition 108 a in FIG. 5, indicating requirements regarding functions, prices, and the like when operating the SaaS 106. The SaaS meter 109 is set and the state of the SaaS 106 is monitored. The information monitored and collected by the SaaS meter 109 has the contents exemplified as the monitoring information 109a in FIG.

  In the optimum distributed arrangement system 105, the locator processing unit 112 in the navigation processing unit 111 reads the policy definitions 108 and 108a in the SaaS 106 and the monitoring result information (monitoring information 109a) by the SaaS meter 109, and the locator processing unit 112 reads the read information. And information stored in the reputation database 113 is used to determine whether or not the cloud service according to the present invention needs to be relocated, and when relocation is necessary, the relocation processing unit 110 is requested to perform relocation. In response to a request from the locator processing unit 112, the relocation processing unit 110 performs relocation of the server resource 101 with respect to the SaaS 106.

  In this way, in the cloud service relocation system of this example, the virtual servers on which SaaS (cloud service) is operating are relocated to other physical servers by looking at the free state of each physical server.

  That is, the cloud service rearrangement system of this example rearranges a cloud service provided by cloud computing to a server resource that executes the processing of the cloud service. The processing operation is monitored and load information of the server resource 101 is collected, and in the storage device (policy definition information storage means) (not shown), the SaaS provided in advance for each SaaS (cloud service) is provided. Policy definition information 108 and 108a including at least the function conditions and price conditions of the necessary server resources, environmental performance conditions and evaluation conditions, and the priority of each condition is stored.

  Then, the locator processing unit 112 collates the load information of each server resource collected by the server meter 102 with the condition information included in the stored policy definition, and extracts all the server resources that satisfy the condition. From all the extracted server resources, one server resource most suitable for the condition with the highest priority in the policy definition information predetermined for the SaaS (cloud service) is specified, and the relocation processing unit 110 uses the locator The SaaS is rearranged on the server resource specified by the processing unit 112.

  Alternatively, the reputation database 113 stores profile information that is set in advance for each server resource and includes at least function information and provision price information of the server resource, and the locator processing unit 112 performs the relocation target SaaS The value of profile information that satisfies each condition information in the policy definitions 108 and 108a by comparing the condition information in the set policy definitions 108 and 108a with each information in the profile information of each server resource stored in the reputation database 113 Are extracted, and the load information of each server resource collected by the server meter 102 is compared with each condition information included in the policy definition information for each extracted server resource. Satisfied All the server resources to be extracted, and one server resource that is predetermined for the SaaS and is most suitable for the highest priority condition in the policy definition information is identified from all the extracted server resources, and the relocation processing The unit 110 rearranges the SaaS to the server resource specified by the locator processing unit 112.

  The service meter 109 monitors the operation status of the SaaS and collects operational load information including at least the CPU usage amount, the memory usage amount, the disk usage amount, and the communication amount of each SaaS. When the value of the operational load information of the SaaS to be rearranged collected by exceeds the threshold predetermined for the SaaS, the processing operation by the locator processing unit 112 may be executed, or in advance The processing operation by the locator processing unit 112 may be executed at a predetermined timing.

  Further, in the policy definition 108, as shown in the policy definition 108a of FIG. 5, the server resource function conditions include the CPU performance, the memory capacity, the disk capacity, and the communication speed of the server resource, and the server resource price condition includes the server resource Includes resource CPU usage, memory usage, disk usage, and network usage, includes server resource power consumption as server resource environmental conditions, and server resource evaluation conditions are determined in advance for each server resource. In addition, the locator processing unit 112 includes customer satisfaction for the server resource, provider evaluation, and reliability, and the locator processing unit 112 is configured based on a threshold that is determined in advance corresponding to the priority of priority information in the policy definition 108.108a Whether to satisfy the value set in the condition Performing determination, select each server resource that satisfies all the conditions, from the server resource selected, identifies one of the server resources that best satisfies the criteria of priority highest in priority information.

  Note that the locator processing unit 112 extracts a server resource that satisfies a preset condition when there are a plurality of highest priority conditions.

  Further, if the server resource specified by the locator processing unit 112 is one that is providing the cloud service, the relocation processing by the relocation processing unit 110 is not performed.

  The relocation processing unit 110 expands the server image on the server resource being processed to the server resource specified by the locator processing unit 112, activates the cloud service on the server resource, and applies the cloud to the activated server resource. Replicate the session information of the service, and after the replication of the session information is completed, the access destination of the user of the cloud service is changed to the server resource specified by the locator processing unit 112, and the server image in the server resource being processed Stop and delete.

  Hereinafter, the details of the configuration and processing contents of the cloud service rearrangement system of this example will be described with reference to FIGS. In FIG. 1, only one server resource 101, SaaS 106, etc. are shown, but there are a plurality of server resources, SaaS, etc. to be processed in the cloud service relocation system of this example.

  The cloud provider (104) providing the server resource (101) defines a server profiler (103) for each server resource (101) under management.

  This server profiler (103) holds server resource profiles such as CPU, memory, network, and disk capacity information, unit price for each resource, and power consumption, and the reputation database 113 in the optimal distributed arrangement system 105. Saved in. An example of a stored server profiler is shown in FIG.

  The cloud provider (104) places a server meter (102) in each server resource (101). This server meter (102) is an application for monitoring the load of the arranged server resource (101), and alive monitoring, CPU usage measurement, memory usage measurement, disk I / O for each server resource (101) Monitor and acquire measurement, disk space monitoring, network I / O measurement, process status, power consumption, etc. An example of the monitored / acquired information is shown in FIG.

  Information acquired by the server meter (102) (server meter information 102a in FIG. 2) is transmitted to the optimal distributed arrangement system 105 via the network at a predetermined timing, and the server Similar to the profiler (103), it is held in the reputation database 113.

  Based on these two pieces of information (server meter information 102a and server profilers 103 and 103a), in the optimum distributed arrangement system 105, the profile (server profiler 103a) of the server existing on the cloud and the load state (server meter 102a) are evaluated. Centralized management is performed in the base 113.

  In addition to these, the operator (evaluator) 114 adds non-quantitative server resource (101) evaluation information (reliability, evaluation from a SaaS user, cloud provider financial information, etc.) As evaluation information for the server resource (101).

  As illustrated in FIG. 4, in the reputation base 113, for each server (server ID; I001), a system corresponding to server meter information 102 a that is a measurement result of the server meter 102 for the server and the server profile 103. The server evaluation information 113a including the reputation 103a and the human reputation 114a, which is evaluation information for the server input from the operator 114, is set. The human reputation 114a in this evaluation information is not limited to the operator 114 as an evaluator, but is a SaaS developer (107) using the server resource (101), an operator of this system, an arbitrary third Input by the user.

  Next, setting by a SaaS developer (operator 107) who deploys SaaS 106 on the server resource 101 on the cloud will be described.

  For the SaaS 106 for which the server resource 101 is to be optimized and operated, the SaaS developer (operator 107) defines in advance a policy definition 108 that describes functional and price requirements for operating the SaaS 106.

  In addition to the system ID (SaaS id) that uniquely identifies the SaaS 106, the policy definition 108 defines various requirements such as CPU requirements, memory requirements, required network speed, and required disk capacity. It is held in the SaaS 106 in the form. An example of the policy definition 108 is shown in FIG.

  In addition, the SaaS developer (operator 107) sets the SaaS meter 109 for the SaaS 106. The SaaS meter 109 is an application for monitoring the state of the SaaS 106, and performs monitoring and collection of monitoring result information such as CPU usage measurement, memory usage measurement, disk I / O measurement and the like in the SaaS 106. An example of information acquired by the SaaS meter 109 is shown in FIG.

  Based on the two SaaS information of the policy definition 109 and the information acquired by the SaaS meter, the optimum distributed arrangement system 105 can manage the policy definition of the server resource 101 necessary for the operation of the SaaS 106 and the operation state of the SaaS 106 itself. .

  Then, the optimum distributed arrangement system 105 automatically selects the server resource (101) that is optimal for the operation of the SaaS 106 by using the information from the server meter 102 of the server resource 101 and the information from the SaaS 106.

  As shown in FIG. 1, the optimal distributed arrangement system 105 of this example includes a relocation processing unit 110 and a navigation processing unit 111. The navigation processing unit 111 includes the above-described reputation database 113 and a locator processing unit 112 that provides an optimal distributed arrangement function.

  Processing operations in the optimal distributed arrangement system 105 including the processing operation of the locator processing unit 112 (“Seeing the free state of each physical server, the entire virtual server on which SaaS is running is relocated to another physical server. )) Will be described with reference to FIG.

  A polling period is set for each SaaS including SaaS 106, and when the set polling period elapses (step 701), the SaaS (106) is displayed in the SaaS meter (109) between the previous polling time and the present time. Monitoring data (for example, monitoring information 109a in FIG. 6) that is continuously collected and a predefined policy definition (for example, policy definition 108a in FIG. 5) are transmitted to the locator processing unit 112 of the optimal distributed arrangement system 105. (Steps 702 to 704).

  The locator processing unit 112 of the optimum distributed arrangement system 105 that has received the monitoring data (monitoring information 109a) and the policy definition (108a) from the SaaS (106) sequentially executes the following four operations A to D (steps). 705-712).

<A (Steps 705 to 706)>
For example, the reference values (A, B, C,...) Are acquired from the “priority-function” column in the policy definition 108a in FIG. 5, and the reference values (A, B, C,...) Are stored in the locator processing unit 112 using the acquired reference values as keys. With reference to the threshold value table, the “function threshold value” corresponding to the reference value (A, B, C,...) Is acquired. The “function threshold” is information defining how much load is allowed for each function item (CPU, Mem, N / W, etc.).

  In this example, a strict threshold is set for the function in order from the reference value B for the reference value A and from the reference value C for the reference value B. For example, taking the CPU (AVG) as an example, when the reference value is A, the CPU (AVG) load is 60% or more, and when the reference value is B, the CPU (AVG) load is 65% or more. The CPU (AVG) in the monitoring data holds a threshold that should not be exceeded.

  This threshold is set for each item of the functional requirement in the policy definition 108a, and the locator processing unit 112 uses the threshold and monitoring data for all the items of the functional requirement, and the server evaluation information 113a (system) in the reputation database 113. Performance predictions calculated from (reputation information) are compared, and servers that can satisfy the threshold values for all items are obtained in a list.

  This will be described in more detail below. First, the locator processing unit 112 acquires the “functional requirement” column in the policy definition 108a of FIG. This column stores static information (CPU rank, memory capacity, etc.) necessary for the server required for SaaS.

  Next, the locator processing unit 112 acquires a reference value set corresponding to the previously acquired “functional requirement” in the “priority-function” column in the policy definition 108 a. Here, it is assumed that the reference value is “A”. Note that the threshold value of “priority-function” includes a threshold value for each item in the policy definition 108a.

  The locator processing unit 112 refers to an internal threshold table using the acquired reference value (“A”) as a key, and acquires a corresponding function threshold (information defining how much load is allowed). This threshold value is unique in units of system units, server meter information items, and threshold level (A, B, C,...) Units. Specifically, “CPU (AVG): A: 20% or less” , B: 22% or less, ..., Z: No limit (values A to Z are unique in the system) "," CPU (MAX): A: 90% or less, B: 93% or less, ..., Z: No limit (The values of A to Z are unique in the system).

  Based on the information in the “functional requirement” column acquired in this way (static information required for the server required for SaaS; CPU rank, memory capacity, etc.) and the threshold of “priority-function” In addition, the locator processing unit 112 searches the server evaluation information 113a (reputation information) stored in the reputation database 113 shown as an example in FIG.

  Then, the locator processing unit 112 acquires information corresponding to the “functional requirement” column from the system reputation information 103a existing in the reputation database 113 (the same information as the server profiler information 103a in FIG. 2). The information corresponding to the threshold of “degree-function” is acquired from the server meter information 102a existing in the reputation database 113 (the same information as the server load information in FIG. 3), and “information corresponding to the“ functional requirement ”column (server (Meter information value) <"priority-function threshold value" is specified.

  That is, the locator processing unit 112 assigns each server resource having performance satisfying information corresponding to the “functional requirement” column (static information necessary for the server required for SaaS) to the server profiler information 103a in FIG. Further, with respect to each extracted server resource, the server load information 102a in FIG. 3 is referred to, and the operating status (load state) of the server meter is acquired. The load is “priority-function”. If the threshold is exceeded, the server resource is excluded from the relocation candidates. For example, when CPU (AVG) = A, the server resource is excluded if the latest server operating status exceeds the threshold.

  Specifically, in the present example, in the policy definition information 108a of FIG. 5, regarding the service whose SaaS identifier (id) is S001, as the functional requirements, the CPU rank is “8”, the memory capacity is “512 MB”, the uplink The communication speed is “10.00 Mbps”, the downlink communication speed is “15.00 Mbps”, the disk capacity is “320,000 MB”, the price condition of the CPU is “8 yen / second”, and the memory usage price is “ 512 yen / MB ”, the upstream communication unit price is“ 10.00 yen / GB ”, the downstream communication unit price is“ 10.00 yen / GB ”, and the disk usage amount is“ 10.00 yen / GB ”. It is registered. The environmental performance condition is “60.00W × m” for power consumption, the evaluation condition is “three stars” for reliability, “four stars” for provider evaluation, and “three stars” for environmental evaluation. It is registered as a request value.

  In the server profiler of FIG. 2, regarding the server resource whose server identifier (id) is I001, as the functional requirements, the CPU rank is “25”, the memory capacity is “8192 MB”, the uplink communication speed is “100.00 Mbps”, Downlink communication speed is “1000.00 Mbps”, disk capacity is “2,340,000 MB”, CPU usage fee is “10 yen / second”, memory usage fee is “8,500 yen”, upstream communication The usage fee is “0.023 yen”, the downlink communication fee is “0.000 yen”, the disk usage fee is “10.000 yen”, and the power consumption is “60.00 W × m” as an environmental condition. It is registered as a value. Information relating to SLA (Service Level Agreement) is also registered, and the operation time is “99.9999”, the refund type is “transfer”, and the refund policy is “Rule008”.

  The server resource with the server identifier (id) I001 shown in FIG. 2 first satisfies the conditions of the policy definition information shown in FIG. 5 with respect to its functional requirements (conditions), and is extracted as a server resource to be relocated. Is done.

  In this manner, by collating the server profiler information shown in FIG. 2 with the functional conditions of the policy definition information shown in FIG. 5, the server resources that can be functionally moved (relocated) SaaS are narrowed down.

  Further, SaaS can be moved in a performance manner by collating the information of the server meter 102a shown in FIG. 3 and the conditions of the policy definition information shown in FIG. Narrow down possible server resources.

  For example, in the server meter 102a shown in FIG. 3, with respect to the server resource whose extracted server identifier (id) is I001, the standard operation rate and maximum operation rate of the CPU, the standard use rate and maximum use rate of the memory, and the up and down The monitoring result data such as the communication speed and communication volume, the remaining capacity of the disk, the amount of writing / reading, the power consumption, the power consumption estimation, the electricity bill, etc. are registered, and this content is the policy definition information shown in FIG. In particular, the standard operation rate and maximum operation rate of the CPU, the standard use rate and maximum use rate of the memory, the communication speed and communication volume of the up and down, and the monitoring result data of the remaining capacity of the disk The server resource with the server identifier (id) I001 is extracted as a SaaS relocatable server resource in terms of performance.

  Thus, by collating the server profiler information shown in FIG. 2 with the conditions of the policy definition information shown in FIG. 5, the server resources that can be relocated in SaaS are narrowed down functionally, and further shown in FIG. The function of the policy definition information shown in FIG. 5 is performed after narrowing down the server resources that can be functionally rearranged by collating the information of the server meter 102a with each condition of the policy definition information shown in FIG. Identify the relocation destination server resource that best suits the conditions.

  Similarly, the locator processing unit 112 executes the following three operations B to D in order for other conditions (price conditions, environmental performance conditions, evaluation conditions) (steps 707 to 712). Thus, it becomes possible to narrow down the server resources that satisfy all the requirements in the policy definition 108a shown in FIG.

  <B = Steps 707 to 708> The locator processing unit 112 acquires the reference value from the “priority-price” column of the policy definition 108a, and uses the reference value as a key from the threshold value table held in the locator processing unit 112. Get the price threshold. The price threshold is information that defines how much excess is allowed for the price of each item (CPU, Mem, N / W, etc.).

  As the reference value is A, a stricter threshold is required for the price. Taking the CPU price as an example, if the CPU price in the policy definition is 10.00 yen, the threshold value when the reference value is A is 10.00 yen, and the threshold value when the reference value is B is 10.50 yen In addition, a threshold that the CPU fee must not exceed is held.

  This threshold is set for each item of the price requirement of the policy definition 108a, and the locator processing unit 112 acquires the threshold for all items of the policy definition 108a, and server resources that can satisfy the threshold for all items Get in a list. At this time, the locator processing unit 112 calculates an assumed price from the monitoring information, and determines the rearrangement priority in ascending order of the assumed price.

  <C = Steps 709 to 710> The locator processing unit 112 acquires the reference value from the “priority-environment” column in the policy definition 108a, and sets the environment threshold value using the reference value as a key from the threshold value table in the locator processing unit 112. get. The environmental threshold is information defining how much excess is permitted for each item (power consumption and environmental performance) of the environment.

  As the reference value is A, a stricter threshold is set for the environment. Taking power consumption as an example, when the reference value is A, the power consumption is 60.0 W × m or less, and when the reference value is B, the power consumption exceeds 65.0 W × m. The threshold that should not be held is held.

  This threshold value is set for each item of the environmental performance of the policy definition 108a, and the locator processing unit 112 acquires the threshold value for all items of the policy definition 108a, and server resources that can satisfy the threshold value for all items Get in a list. At this time, the locator processing unit 112 determines the rearrangement priority in descending order of evaluation regarding environmental performance.

  <D = Steps 711 to 712> The locator processing unit 112 acquires the reference value from the “priority-evaluation” column in the policy definition 108a, and uses the reference value as a key from the threshold value table in the locator processing unit 112 to set the evaluation threshold value. get. The evaluation threshold is information that defines how much shortage is permitted for each evaluation item (reliability, customer satisfaction, provider evaluation).

  As the reference value is A, a stricter threshold is set for evaluation. Taking reliability as an example, if the reference value is A, the power consumption should not be less than 4.5 when the reference value is B, and if the reference value is B, it should be 4 or more. The threshold is held.

  This threshold value is set for each item in the evaluation of the policy definition 108a, and the locator processing unit 112 acquires the threshold value for all items of the policy definition 108a, and sets server resources that can satisfy the threshold value for all items. Get in the list. At this time, the locator processing unit 112 determines the rearrangement priority in descending order of evaluation.

  After that, the locator processing unit 112 merges the list of all server resources acquired by the above processes A to D, and if there are no server resources in the list or the server resource on which SaaS is currently operating. If there is only one (step 713), it is determined that SaaS is not a relocation target, and the relocation processing is not performed (step 717).

  If one or more server resources exist in the list (step 713), the relocation destination is selected after relocation according to the priority of the policy definition 108a (steps 714 to 716). For example, when the priority is a function A, a price B, an environment C, and an evaluation C, the server resource located at the highest level in the “function” is set as the relocation destination. When there are a plurality of priorities A such as function A and price A, it is determined in advance by the system whether any priorities are prioritized, and the priorities are followed.

  If the highest priority relocation destination acquired above is the currently active server resource (step 715), it is determined that SaaS is not a relocation target and the relocation processing is not performed (step 717). If it is different from the currently active server resource, the server resource is determined to be a relocation destination (step 716), and the processing of the locator processing unit 112 is terminated.

  In this way, after the server resource determined to be optimal for operating SaaS by the locator processing unit 112 is selected, the relocation processing unit 110 performs relocation of SaaS.

  The processing operation of the optimal distributed arrangement system 105 including the relocation processing operation of the SaaS 106 performed by the relocation processing unit 110 will be described below with reference to FIG.

  When the relocation determination by the locator processing unit 112 is completed and there is a server resource to be relocated, a relocation instruction notification (relocation notification) from the locator processing unit 112 to the relocation processing unit 110 occurs (steps 801 to 801). 802).

  When the relocation notification is generated from the locator processing unit 112, the optimal distributed arrangement system 105 copies the server image existing in the relocation source server resource (101) to the self relocation processing unit 110 (step 803). The server image is data such as a file that defines the entire server under the OS.

  The relocation processing unit 110 expands the copied server image on the server resource of the relocation destination (step 804), restores the server having the same configuration as the relocation source, and starts SaaS (step 805).

  Furthermore, since it is necessary for the relocation processing unit 110 to shift user access from the relocation source to the relocation destination in a stepwise manner, the relocation processing unit 110 replicates the relocation source session information to the relocation destination (step 806).

  Then, after the replication is completed, the relocation processing unit 110 changes the access destination of the user to the relocation destination, performs communication confirmation (step 807), stops the relocation source server after the communication confirmation is completed. The server resource of the rearrangement source is deleted for each server image (step 808).

  Through the above processing, relocation of SaaS to an appropriate server resource by the optimal distributed arrangement system 105 is completed.

  As described above with reference to FIGS. 1 to 8, the cloud service relocation system of this example includes the navigation processing unit 111 and the relocation processing unit 110 as functions for executing programmed computer processing. The navigation processing unit 111 includes a locator processing unit 112 and a reputation database 113, and stores the following three pieces of information for the server resource 101 in the reputation database 113.

  1. Load information of each server resource acquired from a server meter (102) set in advance in each server resource (101).

  2. Server profile information such as the performance and price of each server resource (101) that is input in advance by the server resource provider (104).

  3. Policy definition information which is set by an arbitrary third party (operator 114) and includes evaluation conditions such as reliability and customer satisfaction for each server resource (101).

  Then, the navigation processing unit 111 uses the locator processing unit 112 to define policy definition information (108, 108a) in which operating conditions (CPU performance conditions, memory capacity conditions, communication speed conditions, etc.) of services (SaaS) prepared in advance are described. ), The policy definition information (108, 108a) and the information on each server resource (101) stored in the reputation database (113) are compared, and a list of server resources that can satisfy the policy definition is obtained. After obtaining the server resource list, the server resource most suitable for the predetermined selection condition is selected from the list.

  If the selected server resource is different from the server resource currently operating SaaS, the server resource is specified as the relocation destination, and the information of the SaaS to be relocated and the information of the server resource specified as the relocation destination are In addition, the relocation processing unit 110 is notified.

  The relocation processing unit 110 copies the service (SaaS) from the currently active server resource that is the relocation source to the relocation destination along with the server information, and after the copy is completed, the server on the relocation destination server resource Start the server from the information and delete the server information from the relocation source. Through the series of operations described above, the service (SaaS) can be arranged and operated on an appropriate server resource.

  Thus, according to the cloud service rearrangement system of this example, information such as performance and price required for the service (SaaS) and operation status information of the server resources that operate each service (SaaS) are used. Thus, it is possible to easily perform an efficient optimal distributed arrangement among SaaS and server resources.

  In addition, this invention is not limited to the example demonstrated using FIGS. 1-8, In the range which does not deviate from the summary, various changes are possible. For example, in this example, IaaS / HaAS has been described as an example of server resources. However, any server resource may be used, and SaaS that is realized using cloud computing technology and used for operating SaaS. The present invention can be applied to all control functions in a data center on-demand resource service that provides possible resources.

  Also, regarding the computer configuration example in this example, a computer configuration without a keyboard or optical disk drive may be used. In this example, an optical disk is used as a recording medium. However, an FD (Flexible Disk) or the like may be used as a recording medium. As for the program installation, the program may be downloaded and installed via a network via a communication device.

  101: Server resource 102: Server meter 102a: Server meter information 103: Server profiler 103a: Server profiler (system reputation information) 104: Operator (cloud (server resource) provider) 105: Optimal distributed arrangement System 106: SaaS (cloud service) 107: operator (SaaS developer (provider)) 108: policy definition 108a: policy definition information 109: SaaS meter 109a: SaaS meter information (monitoring information) 110: Relocation processing unit, 111: Navigation processing unit, 112: Locator processing unit, 113: Reputation database, 113a: Server evaluation information, 114: Operator (evaluator), 114a: Human reputation information.

Claims (10)

A cloud service relocation system that relocates a cloud service provided by cloud computing to a server resource that executes processing of the cloud service,
Server meter means for monitoring the processing operation of each server resource and collecting load information of the server resource;
Policy definition that stores policy definition information that includes at least the function conditions and price conditions, environmental performance conditions and evaluation conditions, and the priority of each condition necessary for providing the cloud service set in advance for each cloud service Information storage means;
By comparing the load information of each server resource collected by the server meter means with the condition information included in the policy definition information stored by the policy definition information storage means, all server resources satisfying the conditions are extracted. Then, from all the extracted server resources, one server resource most suitable for the condition with the highest priority in the policy definition information that is predetermined for the cloud service and stored in the policy definition information storage unit is specified. Locator means to
A cloud service rearrangement system comprising: a relocation unit that rearranges the cloud service on a server resource specified by the locator unit. A cloud service relocation system that relocates a cloud service provided by cloud computing to a server resource that executes processing of the cloud service,
Server profile storage means for storing profile information that is set in advance for each of the server resources and includes at least function information and offer price information of the server resource;
Server meter means for monitoring the processing operation of each server resource and collecting load information of the server resource;
Policy definition that stores policy definition information that includes at least the function conditions and price conditions, environmental performance conditions and evaluation conditions, and the priority of each condition necessary for providing the cloud service set in advance for each cloud service Information storage means;
Each condition information set for the relocation target cloud service stored in the policy definition information storage means and included in the policy definition information is collated with information in the profile information of each server resource stored in the server profile storage means. And extracting all server resources that satisfy each condition information included in the policy definition information,
For each extracted server resource, the load information of each server resource collected by the server meter means is compared with each condition information included in the policy definition information stored by the policy definition information storage means. Extract all server resources that satisfy the conditions,
A locator that identifies one server resource most suitable for the condition with the highest priority in the policy definition information that is predetermined for the cloud service and stored in the policy definition information storage unit from all the extracted server resources Means,
A cloud service rearrangement system comprising: a relocation unit that rearranges the cloud service on a server resource specified by the locator unit. A cloud service rearrangement system according to claim 1 or claim 2,
Service meter means for monitoring the operational status of the cloud service and collecting operational load information including at least the CPU usage and memory usage, the disk usage and the communication volume of each cloud service,
When the value of the operational load information of the cloud service to be rearranged collected by the service meter means exceeds a predetermined threshold for the cloud service,
A cloud service rearrangement system characterized by executing a processing operation by the locator means. In the cloud service relocation system according to claim 1 or 2, at a predetermined timing,
A cloud service rearrangement system characterized by executing a processing operation by the locator means. The cloud service rearrangement system according to any one of claims 1 to 4,
The above policy definition information
As the server resource functional conditions, including server resource CPU performance, memory capacity, disk capacity and communication speed,
As the price condition of the server resource, the CPU resource unit price, the memory unit price, the disk unit price, and the network unit price of the server resource,
As an environmental condition of the server resource, including the power consumption of the server resource,
As the evaluation condition of the server resource, including customer satisfaction and provider evaluation and reliability for the server resource predetermined for each server resource in advance,
The locator means is
Based on a predetermined threshold corresponding to the priority of the priority information in the policy definition information, it is determined whether or not the value set for each condition is satisfied, and all the conditions are satisfied. Select each server resource to
From each selected server resource,
A cloud service rearrangement system that identifies one server resource that most satisfies the highest priority condition from the priority information. The cloud service relocation system according to claim 5,
The locator means is
A cloud service rearrangement system that identifies a server resource that satisfies a preset condition when there are a plurality of highest priority conditions in the policy definition information. A cloud service rearrangement system according to any one of claims 1 to 6,
If the server resource identified by the locator means is providing the cloud service,
A cloud service relocation system characterized in that the relocation processing by the relocation means is not performed. A cloud service rearrangement system according to any one of claims 1 to 7,
The relocation means is
The server image on the server resource being processed of the cloud service to be rearranged is deployed on the server resource specified by the locator means, and the cloud service is started on the server resource,
The session information of the cloud service is replicated to the activated server resource,
After the replication of the session information is completed, the access destination of the user of the cloud service is changed to the server resource specified by the locator means, and the server image in the server resource being processed is stopped and deleted. Cloud service relocation system.   The program for functioning a computer as each means in the cloud service rearrangement system in any one of Claims 1-8. A cloud service relocation method for a system in which a cloud service provided by cloud computing is arranged on a server resource that executes processing of the cloud service by a computer device,
The computer apparatus includes each means in the cloud service relocation system according to any one of claims 1 to 8, as means for executing programmed computer processing,
9. A cloud service rearrangement method, comprising: rearranging server resources for executing the processing of the cloud service by the processing according to claim 1 by each of the means provided.

Images