KR100799775B1 - Mobile grid gateway replication system and method of wireless grid network - Google Patents

Abstract

The present invention relates to a wireless grid network. To this end, the present invention, unlike the conventional grid computing using a wired communication network, serves as an interface of a mobile user in a wireless grid network, the characteristics of the requested application service And the mobile grid client that transmits the requirements by determining the characteristics of the resource, performs user authentication and authorization authentication while communicating with the mobile grid client, and the required performance of the resource and the level of grid service requirements. A wireless grid gateway that includes a mobile grid gateway that converges in the form of service level objectives (SLOs) and service level agreements (SLAs) to generate available resource quorums, and to plan and execute tasks for application services through resources acquired within the available resource quorum. Mobile Grid Gates in Grid Networks By providing a replication system, it can improve overall performance by ensuring the reliability and availability of the wireless grid network.

Mobile Grid Gateway Replication System, Grid Computing, Replica, Policy Quorum Based Management System

Description

Mobile Grid Gateway Replication System and Method for Wireless Grid Networks {MOBILE GRID GATEWAY REPLICATION SYSTEM AND METHOD OF WIRELESS GRID NETWORK}

1 is a block diagram of a mobile grid gateway replication system of a wireless grid network according to the present invention;

2 is a flowchart illustrating an operation of a mobile grid gateway replication system of a wireless grid network according to the present invention;

3 is a flowchart illustrating a SLO, SLA, and budget based policy negotiation process in a mobile grid gateway replication system according to the present invention;

4 is a flowchart illustrating a mobile grid gateway replication process in a wireless grid network in accordance with the present invention.

<Description of the symbols for the main parts of the drawings>

102: Mobile Grid Client 1021: Gateway Connection Manager

1022: Application Service Registration Module 1023: First Policy Manager

10231: SLO Decision Module 10232: SLA Agreement Module

10233: service budget adjustment module 1024: grid resource request module

1025: user UI module 1026: data temporary storage

104: Mobile Grid Gateway 1041: Client Interface

10411: Authentication Manager 10412: Authority Manager

10413: SLA fastening module 1042: internal linkage component

10421: Manager Monitoring UI 10422: CONTEXT Manager

10423: logic mapping module 1043: gateway middleware

10431: authentication manager 10432: service manager

10433: Second Policy Manager 10434: Replica Manager

104341: Replica Catalog 104342: Replication Module

1044: Grid Service Manager 10441: Access Control Manager

10442: grid resource management module 10443: data management module

10444 Process Control Module 10445 Process Scheduler

10446: Billing Manager

The present invention relates to a wireless grid network, and more particularly, to a mobile grid gateway replication system and method of a wireless grid network suitable for performing mobile grid gateway replication in a wireless grid network. will be.

As is well known, traditional grid computing techniques have been used to meet the complexity of performing various simulations and applications based on the robust distributed computing and high performance parallel processing techniques typically required by scientific or engineering systems. .

In general, grid computing connects disparate computing resources that are geographically dispersed into one large virtual organization (VO) to provide users with a computing environment that provides users with a computing environment that is like using a single supercomputer. Means a computing paradigm.

In general, heterogeneous computing resources include not only physical resource nodes such as desktop computers, workstations, servers, but also software, processor cycles, storage spaces, memory, databases, and operating systems. It is a concept that encompasses all logical resources such as OS.

In addition, the general configuration environment of the grid solves the heterogeneity of grid computing resources, satisfies the performance of various user or application services, selects and provides appropriate computing resources, and allocates application services to resources. The middleware that can do it is the core.

In other words, it has three stages of configuration, such as 'user environment, middleware, grid computing resources'. However, in view of scalability of infrastructure scale, the above-mentioned brief grid computing configuration is limited to the configuration of one individual grid environment for a specific organization or enterprise, so from a national or global computing infrastructure perspective, Organization of grid computing infrastructure is possible.

In addition, the grid computing infrastructure is not a rigid structure consisting of two computing hosts that perform fixed business logic like the traditional client / server computing model, but a highly flexible computing infrastructure that allows organization and decommissioning of resource composition at any time. It can be seen as a next-generation computing infrastructure in that it can support an organization and ensure an environment in which a single computing host can serve both a resource user and a provider.

Therefore, the grid middleware system should be able to reflect the selection of resources available as grid computing resources and the respective performance of each of the selected resources and reflect them in the request of an application service. To this end, it is essential to design and implement the performance evaluator's internal system capable of quantitatively expressing the status of monitors and resources capable of observing resource status at all times or on demand based on the designed algorithm.

Basically, as the grid middleware system is a key goal to actively ensure flexible linkage between user (mobile grid client) requirements and grid computing resources, the pre-design and implementation of a robust interface is important.

Conventional grid computing middleware is for monitoring, controlling, and distributing the demand for resources within a widely distributed computer networked environment.

Among the conventional distributed computing middleware technology, "P2P-based parallel distributed processing system and method" disclosed in Korean Patent Application No. 2004-0084831 is that any computer that allows resource sharing in grid computing technology requests or uses resources of another computer. The present invention relates to a peer-to-peer parallel distributed processing system and method for providing a resource to a computer so that the computer using the resource can function as a supercomputer.

Such a parallel distributed processing system implements a resource sharing middleware installed in a computer based on a predetermined criterion to perform a resource sharing function with another computer. In the parallel distributed processing system, the computer with the resource sharing middleware installed grabs and stores the resource information including the IP address, CPU performance, RAM capacity, and network speed. If there is a request for use, if the computer's self-reservation resource provision conditions are met, the method permits the use of resources and performs a requested task from the other computer, and then provides the computer with a work result.

Since this parallel distributed processing system is not based on the service level decision based on the characteristics of the application service, there is a risk of random mapping in the proper resource selection procedure. Compared to the quorum-based resource management scheme that enables this, the efficiency can be greatly reduced and the computational complexity can be very high, and it does not reflect the requirements of the user, the subject of the application service execution request. It is difficult to establish a business profit model.

In particular, when hundreds or thousands of resources are available, the existing resource planning system based on linear programming (LP) cannot select a resource that is most suitable for the user. . Since such grid computing environment requires different parameters of resources depending on the characteristics of the application, there is a need to flexibly adapt to changes and considerations.

In addition, the parallel distributed processing system is limited to the object of grid computing to the wired network, it is difficult to reflect the rapidly developing requirements of users of wireless devices, and to support resource sharing through simple resource monitoring and selection procedures only. Although the method cannot guarantee the grid computing infrastructure power and the reliability of the user's application service execution from problems such as system error and network defect, there is a problem in that it does not mention an appropriate method.

Therefore, among the grid computing middleware technologies, the resource management technology monitors the status of resources that change in time and space in consideration of the complexity and uncertainty of the grid computing environment, and actively discovers and requires the resources that users need in the current state. There is a need for resource management and scheduling techniques that can be implemented and ensured reliability until the task is completed.

To date, grid computing has been experimentally implemented on wired network environments, for systems that can provide high-speed networks and some large computing resources. However, in order to build a practical grid environment and to induce realistic user participation, there is an urgent need to change to a structure capable of eliciting popular support such as current P2P or the Internet.

In addition, in order to induce the participation of companies which are service providers such as existing networks, computers, and contents, it is necessary to plan an appropriate billing system according to the use of grid services.

As a result, the high reliability and availability of mobile grid gateway-based computing resources to actively include not only wired networks but also rapidly developing wireless network-based users in the area of grid computing infrastructure to satisfy two key points. The development of a grid middleware system that guarantees guarantees and proper reflection of user requirements is greatly required.

Accordingly, the present invention is to solve the above problems of the prior art, the mobile grid gateway replication of a wireless grid network that can build and provide a mobile grid gateway system based on a replication technique that can ensure reliability for a wireless grid network. Its purpose is to provide a system and method.

In addition, another object of the present invention is to provide a mobile grid gateway replication system and method of a wireless grid network that can support the offline operation (off-line operation) of the mobile user by designing and building a mobile grid gateway system.

In order to achieve the above object, the present invention provides a mobile grid gateway replication system in a wireless grid network, which performs a role of an interface of a mobile user, inputs characteristics and environments of requested application services, and determines characteristics of resources. A mobile grid client that transmits a requirement, performs user authentication and authorization authentication while communicating with the mobile grid client, and performs a service level objective (SLO) and a service level (SLO) on the required performance of the resource and the required level of the grid service. Mobile grid gateway replication system of a wireless grid network including a mobile grid gateway that converges in the form of an agreement) to generate an available resource quorum, and to plan and execute tasks for the application service through the resources obtained in the available resource quorum. It provides.

In addition, in order to achieve the above object, the present invention provides a method for replicating a mobile grid gateway of a wireless grid network, comprising: a process of granting user authentication and resource usage rights by a mobile grid gateway according to a session connection request of a wireless device user; A second process of performing a registration for performing an application service requested by a user and returning information on its validity period, and service level objectives and agreements for entering into a policy-based service contract between the user and the grid service provider; A third process of configuring a mobile user side policy through budget adjustment, a fourth process of requesting selection of grid computing resources based on the user side policy descriptor, receiving the returned process, and execution of an application service registered in the mobile grid gateway; Above selected grid A fifth process of allocating sub-tasks to computing resources, a sixth process of performing the assigned sub-tasks and monitoring such performance, and returning the result of the performed sub-tasks to the user, and charging for this A mobile grid gateway replication method of a wireless grid network including a seventh process of notifying the details is provided.

In order to achieve the above object, the present invention provides a method for replicating a mobile grid gateway in a wireless grid network. The present invention provides a method for identifying application service characteristics in a mobile grid gateway and generating a policy of SLO and SLA for selecting a gateway candidate group. Requesting selection of a gateway, requesting selection of a gateway candidate group using the generated policy, returning location and status information of a gateway resource group according to the gateway candidate group, and selecting the selected gateway Record corresponding information of the candidate group, calculate the optimal number of replicas, select a master / secondary gateway, check whether the operation state of the selected master gateway is normal according to a predetermined period, Information A mobile grid gateway replication method of a wireless grid network including synchronizing is performed.

Finally, in order to achieve the above object, the present invention provides a method for replicating a mobile grid gateway in a wireless grid network, wherein the mobile grid gateway includes a binary form including an index of each individual data for a data set and an execution log for performing an application service. And delivering a file to a temporary storage of data on a mobile device, and performing offline operation support for compensating for low network quality according to a wireless environment in the temporary storage of data and preventing a delay in execution time of the application service. A mobile grid gateway replication method of a wireless grid network is provided.

The above and other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Unlike the conventional method of performing grid computing using a wired communication network, the core technical aspect of the present invention grants user authentication and resource use authority, performs registration for performing a requesting application service, and provides service level objectives and agreements. In addition, after configuring the mobile user-side policy through service budget adjustment, requesting selection of grid computing resources, and receiving it, sub-tasks are allocated to selected grid computing resources to be performed and monitored. By returning the results of the sub-tasks and notifying the billing details thereof, it is easy to achieve the purpose of the present invention through such technical means.

As the core technology of the mobile grid gateway replication system according to the present invention, the first is the policy configuration of mobile users using service level objectives (SLOs), service level agreements (SLAs), and service budget decisions.

Second, it is a Policy Based Management Scheme that derives an optimal policy that can satisfy common requirements through consultation with the grid computing resource provider's policy based on the configured policy.

Third, a robust design of a replication technique is required as a technique for ensuring reliable network connectivity between the wireless device and the gateway. Specifically, there is a need for providing an algorithm capable of selecting a replication candidate resource and managing a list, determining the number of replica generations, execution cycles, and timing.

Fourthly, in order to overcome the disconnection of the process execution due to the frequent mobility of mobile resources, the definition of user cache structure to support offline operation and the devising of utilization techniques are proposed. Required.

Fifth, grid computing resource selection technology using policy matters concluded between user and service provider through mobile grid gateway should be provided. Fundamental contents of the present invention are the preparation of a detailed execution procedure and implementation of a procedure related to resource determination and resource grouping, resource reconfiguration, and decommissioning based on the quorum technique, which is one of the areas of distributed computing technology. to be.

Sixth, through the implementation of the Agent-Based Resource Monitoring System, which defines the most recent grid resource state, manages the state history, and quantitatively estimates and calculates the optimal resource state from it. Process Scheduling can be supported.

Seventh, the Billing Manager can charge appropriate fees, taking into account the final policy concluded and the quantitative figures of the resource capabilities provided to the user, for example, the computing power of the resources, the storage capacity provided, the class of service and the time. Practical implementation is required. It is essential to establish a differential and flexible billing decision-making system based on the service provider's service quality and the degree of guarantee of the requirements, rather than the one-way billing for users' service and grid computing resource usage.

1 is a block diagram of a mobile grid gateway replication system of a wireless grid network according to the present invention, which includes a mobile grid client 102, a mobile grid gateway 104, a database 106 for storing billing information, and agent-based resource monitoring. Server 108 and resource status monitoring agent 110. Here, the mobile grid client 102 may include a gateway connection manager 1021, an application service registration module 1022, a first policy manager 1023, a grid resource request module 1024, a user UI module 1025, and data temporary data. And a repository 1026, the mobile grid gateway 104 includes a client interface 1041, an internal linkage component 1042, a gateway middleware 1043, and a grid service manager 1044.

Referring to FIG. 1, the gateway connection manager 1021 of the mobile grid client 102 communicates through the client interface 1041 of the mobile grid gateway 104, and serves as an interface with a mobile user.

In addition, the application service registration module 1022 of the mobile grid client 102 provides a function of supporting information input and registration for an application or service that requires a mobile user to perform using the grid computing infrastructure.

In addition, the first policy manager 1023 delivers the user-side policy matters determined through the SLO and SLA techniques to the service budget adjustment module 10233 to create an appropriate estimated cost, and the policy matters thus prepared are transferred to the mobile grid gateway 104. Delivered to SLA fastening module 10413. This first policy manager 1023 includes an SLO determination module 10231, an SLA agreement module 10232, and a service budget adjustment module 10233.

Next, the grid resource request module 1024 returns the quorum resource set return of the mobile grid gateway 104 grid service manager 1044 having the capability to satisfy the final policy details reflecting the requirements of the pre-registered application service. It has a function required for the grid resource management module 10442.

In addition, the user UI module 1025 plays a role of visualizing an internal operation process of the mobile grid client 102 and a connection process with the mobile grid gateway 104.

In addition, the data temporary storage 1026 temporarily stores the index and log information of the data set in the form of a binary file to perform an application service supporting offline operation. For this purpose, the CONTEXT manager 10522 of the internal linkage component 1042 may be used. The execution process of the application service is checked for each time period or for each defined service step, and the details of each intermediate execution process are transmitted to the data transmission module 10443 of the grid service manager 1044 through the logic mapping module 10423. The data transmission module 10443 transmits a binary file to the gateway connection manager 1021, and as a result, is stored and managed for each user and application service in the data temporary storage 1026. Here, the data temporary storage 1026 is notified of the deletion of the relevant data at the end of the application service from the CONTEXT manager 10422 in order to prevent unnecessary waste of data storage space in the mobile grid client 102.

Meanwhile, the client interface 1041 includes an authentication manager 10411, an authority manager 10412, and an SLA execution module 10413, and basically functions such as accepting and authenticating a connection request of a mobile grid client, authentication, authorization, policy conclusion, and the like. In charge of.

Here, the authentication manager 10411 plays a role in charge of basic communication with the gateway connection manager 1021, and the authority manager 10412 has appropriate authority for each mobile user session connected through the authentication manager 10411. And the execution scope, the SLA execution module 10413 converges the user-side policy details transmitted from the first policy manager 1023, and the service provider side to the second policy manager 10433 of the gateway middleware 1043. After requesting the policy details, it is desirable to determine the policy that can yield the best cost performance by comparing them.

Next, the internal linkage component 1042 includes a manager monitoring UI 10421, a CONTEXT manager 10422, and a logic mapping module 10223, primarily for internal logic performance management.

Here, the manager monitoring UI 10421 means a graphical management interface provided for facilitating management of each individual mobile grid gateway 104 at the service provider side, and the CONTEXT manager 10522 is a work of performing internal logic. Manages Work-Flow, tracks, and delivers the degree of execution of logic and the occurrence of specific events to other managers, components, or module components through logic mapping module 10423 at appropriate points in time. The logic mapping module 10423 performs a function of notifying a specific event recognized by the CONTEXT manager 10423 to other modules inside.

In addition, the gateway middleware 1043 includes an authentication manager 10431, a service manager 10432, a second second policy manager 10433, and a replica manager 10434, and the upper component group in the mobile grid gateway 104. And a component group in charge of the middleware layer relaying a role between the sub grid service manager 1044.

Here, the authentication manager 10431 serves to link the user request connected through the client interface 1041 to the grid service account, and performs a role of requesting the access control manager 10441 to issue an account, and the service manager ( 10432 manages information about the application service requested to be registered through the application service registration module 1022, and the second policy manager 10433 provides this when the SLA execution module 10413 requests the service provider side policy details. In order to dynamically generate the service provider side policy history, the resource status is grasped in real time through the grid resource management module 10442.

Meanwhile, the replica manager 10434 includes a replica catalog 10435 and a replication module 10436, and serves to ensure reliability in a wireless grid environment. Here, the replica catalog 10435 is responsible for managing location and status information of the group of mobile grid gateways 104 generated through the replication module 10436, and problems occur in the master mobile grid gateway 104. In this case, an algorithm for selecting an optimal gateway to replace the algorithm is performed, and the replication module 10436 calculates the optimal number of gateways capable of performing the role as the secondary mobile grid gateway 104. It performs related information exchange and data transmission for information synchronization between gateways.

Grid service manager 1044 then access control manager 10441, grid resource management module 10442, data transfer module 10443, process control module 10444, process scheduler 10445, and billing manager 10446. It includes the application service allocation to the grid resource node, selection of the quorum resource set, monitoring the execution status of the process, and charging.

Here, the access control manager 10441 provides the grid service account requested by the authentication manager 10431, controls and manages matters related to accessing the grid computing infrastructure of the account, and the grid resource management module 10442 Receive real-time resource status information from the external agent-based resource monitoring server 108 to manage the optimal resource group through the quorum scheme, and the data transmission module 10443 stores a data set necessary to perform an application service. The method transmits the function to the selected resource node group through step 10442.

In addition, the process control module 10444 uses the resource status monitoring agent 110 to monitor whether there is an abnormality in the sub-tasks being executed in each resource node, and the process scheduler 10445 uses the sub-tasks of the sub-tasks to be performed in each resource node. For allocation, a budget and deadline scheduling technique is used, and the billing manager 10446 stores a billing database for storing billing information for each user and application service to charge a fee for performing the user's application service. And store an application service termination notification from the CONTEXT manager 10522, and forward the final billing details to the mobile grid client 102.

On the other hand, it is preferable that the external server includes a database 106 for storing billing information, an agent-based resource monitoring server 108 and a resource status monitoring agent 110 for the light weight and realistic implementation of the mobile grid gateway 104. .

Here, although not shown in FIG. 1, the mobile grid client 102, the mobile grid gateway 104, the database 106 for storing billing information, the agent-based resource monitoring server 108, and the resource status monitoring agent 110 are provided. Of course, each can communicate with each other as needed.

Meanwhile, in the present invention, the main techniques for configuring the mobile grid gateway replication system include a policy-based service level goal, service level agreement, service budget determination method, service level goal and agreement technology method, agreement procedure between user and service provider, and optimal policy. Derivation method, mobile grid gateway replication method, mobile device offline operation method, mobile device temporary storage operation method, agent based resource status monitoring method, resource status quantification method, quorum based resource grouping and reconfiguration method reflecting policy and resource status The main design techniques are the process scheduling scheme, the history management according to the process execution and the scheme of billing system.

Here, replication, which is a key feature of the present invention to ensure reliable linkage between the mobile network and the existing grid computing infrastructure, is designed to prevent possible network connection loss, poor bandwidth, and failure to perform work due to frequent movement of mobile devices. Used as the primary technique.

In order to use such a replication technique, first, a procedure for selecting a candidate resource set that may be considered as a replica needs to be performed. This is performed by the grid resource management module 10442 of the grid service manager 1044. It can be done through. As described above, the resource selection maps the status information of each individual resource item to the final policy determined between the user and the service provider and the available resource list obtained from the agent-based resource monitoring server 108 for each item. Use a satisfying method. In other words, the selection of candidate resource sets to be used for replication is also performed using the grid resource management module 10442. In general, unlike the selection of grid computing resources that perform sub-tasks of application services, it is possible to perform a role as a gateway. Replica selection is carried out by ensuring that the predefined minimum functional requirements are met. By using the same processing structure that unifies both the selection of replication and grid computing resources, it is desirable to avoid the complexity and unnecessary configuration of the system configuration due to redundant functions.

In the present invention, when the agent-based resource status monitoring and quantitative relative evaluation of resource performance are basically possible, the present invention has a structural feature capable of operating all functions of the upper system. Specifically, a separate monitoring agent is mounted for each grid resource, and a central resource state monitoring server collects monitoring data collected from each agent and manages it as a separate table. This table is based on the existing relational database (RDBMS) system, but the storage format can be converted into an XML document or an XML database.

In addition, in order to ensure smooth function implementation and operation of the mobile grid gateway system, it is preferable to include a separate agent-based resource status monitoring server 108 connected to the outside and a resource status monitoring agent 110 mounted on each resource. Here, each monitoring agent mounted on a grid resource is basically an end-to-end network quality of service between other grid computing resources that are in a network connection, sharing data, or participating in collaboration. Measures delay, bandwidth, and round trip time, which are important elements of (QOS: Quality of Service). In addition, as a system element, it monitors the processor capacity, storage capacity, memory size, and available system resources of the resource.

This resource status monitoring agent not only passively observes resource status but also actively detects system errors, problems, and status of running sub-tasks of grid resources and transmits the information to the mobile grid gateway 104. It serves to convey. In addition, the quorum is observed when the performance of the resource is observed in a quantitative state below that, in addition to the occurrence of common errors and exceptions, as well as the system resource performance items defined in the final policy statement, which are the result of agreements concluded between users and service providers. In addition to demoting or downgrading the rank of the grid resource in the resource group, new grid resources can be selected to satisfy the system performance requirements of the policy to ensure the transfer of authority to perform subtasks and the continuity of execution. It is preferable.

In the present invention, the above-described procedure is defined as grid resource reconfiguration. In addition, grid resource reconfiguration is a representative way to improve the overall performance by ensuring the reliability and availability of the wireless grid environment, along with replication techniques.

In addition, in order to support offline operation, in the mobile grid gateway, a data set may exist on the gateway or in a database to perform an application service requested by the mobile device, and the gateway may provide an index and a performance log for each individual data. ) Is transmitted to the data temporary storage 1026 of the mobile device at a predetermined cycle or each step of the service when performing an application service. The data temporary storage 1026 of the wireless device is preferably used as an excellent offline operation technique that can compensate for the low network quality in the wireless environment and prevent the delay of the execution time of the application service.

Next, the mobile grid gateway replication system of the wireless grid network having the configuration as described above performs authentication of the mobile user, generates available resource quorum according to the SLO and SLA policy details for the requested application service, and generates The process of planning and executing a task through the resources acquired in the available resource quorum, and setting up billing when the task is completed, is notified of the mobile user.

2 is a flowchart illustrating an operation of a mobile grid gateway replication system of a wireless grid network according to the present invention.

Referring to FIG. 2, the client interface 1041 of the mobile grid gateway 104 performs user authentication and resource usage authorization from the mobile grid client 102 through a proxy authentication scheme of a client / server structure ( Step 202).

Then, through the client interface 1041 of the mobile grid gateway 104, the requested performance for the resources of the authenticated mobile user, the required level of the grid service, and the requirements for the task are received (step 204). Here, register the requested application service from the mobile user through the application service registration module 1022 of the mobile grid client 102 and the authorization manager 10412 of the mobile grid gateway 104, and obtain a grid account for it. .

In addition, the mobile grid gateway 104 also provides SLOs for the performance requirements of the mobile users provided from the first policy manager 1023 of the mobile grid client 102, the requirements level of the grid service, and the requirements for operations. After collecting the SLA-based policy details, the second policy manager 10433 generates available resource quorum according to the user's request level, application service characteristics, and optimization policy through the policy quorum (step 206).

Next, the grid service manager 1044 of the mobile grid gateway 104 analyzes the sub-tasks of the mobile user to plan and execute the user's tasks through the resources obtained in the available resource quorum, and the real-time of the running tasks. Support monitoring (step 208).

At this time, the mobile grid gateway 104 checks whether the performance of the running job is lower than the user's requirement level (step 210).

As a result of the check in the step 210, if the running job is lower than the user's requirement level, the process scheduler 10445 of the grid service manager 1044 automatically reassigns the job in preparation for deterioration and failure of the resource state. Perform a resource reconfiguration function (step 212).

Then, the mobile grid gateway 104 checks whether a defect in the wireless network or gateway occurs (step 214).

As a result of the check in step 214, if a defect in the wireless network or gateway occurs, the client interface 1041 of the mobile grid gateway 104 resets the master gateway (step 216).

Next, the mobile grid gateway 104 checks whether the running task is complete (step 218).

When the check in the step 218 is completed, the running task is recorded, and after the data temporary storage 1026 is cleaned up (step 220), the user is notified that the work is completed, and the grid service manager The billing manager 10446 of 1044 sets a bill to be charged to the user and notifies the corresponding mobile user (step 222).

Next, in the mobile grid gateway replication system of the wireless grid network, the SLO and SLA are determined according to the application service characteristics, and the user-side policy details through budgeting are compared with the service provider-side policy details generated according to the grid resource POOL characteristics. The final policy details are determined, the final policy details are performed according to the selected optimal quorum resource set, and the process of estimating the expected charges for this is explained.

3 is a flowchart illustrating a SLO, SLA, and budget based policy negotiation process in a mobile grid gateway replication system according to the present invention.

Referring to FIG. 3, the application service registration module 1022 of the mobile grid client 102 identifies the application service characteristics, and the first policy manager 1023 provides a reference for SLO and SLA setting (step 302).

Accordingly, SLO setting is performed through the SLO determination module 10231 of the first policy manager 1023 and SLA determination is performed through the SLA agreement module 10232 of the first policy manager 1023 (step 304, 306).

Then, the mobile grid client 102 checks whether the validity of the user policy details based on the SLO and SLA is satisfied (step 308).

As a result of the check in step 308, if the validity of the user policy details is satisfied, an appropriate user-side budgeting procedure is performed through the service budget adjustment module 10233 of the first policy manager 1023 (step 310). In this case, if the validity of the user policy details is not satisfied, the process is repeated from step 304.

At this time, the mobile grid client 102 checks whether the estimated budget is within the budget range assumed by the user (step 312). As a result of the check in the step 312, if the estimated budget is not within the budget range, the process is performed again from the step 304, thereby performing a cost-adaptable service level determination procedure that reduces the estimated budget by lowering the quality of service. .

At the same time, the service provider determines the grid resource POOL characteristics using the grid resource management module 10442, the agent-based resource monitoring server 108, and the resource status monitoring agent 110 (step 314). At this time, the status information of each item divided into network quality of service (QoS) and computing QoS is used as a criterion for analyzing resource POOL characteristics.

The mobile grid gateway 104 then checks the status of the individual resources and creates a basic service provider policy (steps 316 and 318).

At this time, the second policy manager 10433 of the gate middleware 1043 checks whether the service provision validity is satisfied (step 320).

As a result of the check in step 320, if the service provisioning validity is satisfied, the second policy manager 10433 of the gate middleware 1043 classifies the analyzed grid resource POOL property into each grid service class (step 322). Here, if the service provision validity is not satisfied, the process is performed again from step 316.

Accordingly, the second policy manager 10433 of the gate middleware 1043 determines the final policy details by comparing the defined user side policy details and service provider side policy details (step 324).

The process scheduling scheme is determined according to the final policy details, and an optimal quorum resource set is selected (step 326). At this time, the sub-task is allocated to the quorum resource selected through the process scheduler 10445 for concrete execution of the final policy details.

Meanwhile, the mobile grid gateway 104 checks whether the policy satisfaction resource set is valid while performing the final policy details (step 328), and estimates the expected billing if the policy satisfaction resource set is valid (step 330).

Next, the mobile grid gateway replication system of the wireless grid network identifies the application service characteristics, selects the candidate candidate groups according to the basic policy levels of the SLO and SLA, and selects the master gateway and the secondary gateway among them. Explain the process.

4 is a flowchart illustrating a mobile grid gateway replication process in a wireless grid network according to the present invention.

Referring to FIG. 4, the mobile grid gateway 104 performs application service characteristic grasp using the application service registration module 1022 to perform a replication scheme (step 402).

At this time, a basic policy level of SLO and SLA for selecting a gateway candidate group is generated, and a request is made to the replica catalog 104341 of the replica manager 10434 (steps 404 and 406).

Then, the replica catalog 104341 requests the grid resource management module 10442 to select the gateway candidate group through the quorum technique using the generated policy (step 408).

In addition, location and status information of the gateway resource group is returned according to the gateway candidate group (step 410).

Corresponding information is recorded in the replica catalog 104341 generated through this, and the replica optimal number is calculated and the master / secondary gateway is selected using the replication module 104342 (steps 412, 414, 416).

Thereafter, it is periodically checked whether the operation state of the master gateway is normal (step 418), and if it is normal, information synchronization between the gateways is performed (step 420). If an abnormality is found in the master gateway operation state, the process is performed again from step 412 to select one of the candidate replicas selected as the auxiliary gateways as the master mobile grid gateway 104.

On the other hand, in the case of supporting offline operation in the mobile grid gateway replication system of the wireless grid network, the mobile grid gateway 104 temporarily stores a binary file including each individual data index and performance log of a data set for performing an application service. The data temporary storage 1026 may be used as an excellent offline operation technique that compensates for network quality of a wireless environment and prevents delay of execution time of the application service.

As such, the mobile grid gateway replication system and method of a wireless grid network is a flexible system capable of discovering a set of resources that meets the user's SLO, SLA and budget according to the time and space of the resource in the mobile grid. Give a way. In particular, the mobile grid gateway replication system and method of the wireless grid network according to the present invention can be manufactured to optimize the quality of service (QoS) by continuously maintaining a certain level of resources required by the user.

Next, an offline operation process of the mobile grid gateway replication system of the wireless grid network will be described.

5 is a flowchart illustrating an offline operation process supported by a mobile grid gateway replication system of a wireless grid network according to the present invention.

Referring to FIG. 5, the connection between the gateway connection manager 1021 of the mobile grid client 102 and the client interface 1041 of the grid gateway 104 is performed (step 502).

Then, SLA, SLO, and budget related policy matters necessary for registration and execution of the application service are performed, and the application service is performed on the allocated grid virtual resource according to the policy quorum-based grid resource management technique (step 504). ).

In addition, in order to generate index information on data required during application service execution, the data temporary storage 1026 converges main data from the mobile grid gateway 104, generates a log of the currently completed performance, and then executes the data. An index table for the data set to be created is created (step 506).

Next, the service manager 10432 of the gateway middleware 1043 checks whether the application service execution from the CONTEXT manager 10522 has been completed (step 508).

As a result of the check in the step 508, if the application service execution is in progress, the process of step 506 is carried out, and continues until such application service execution is completed, and as a result of the check in the step 508, the application When the service is completed, the connection loss between the mobile grid client 102 and the mobile grid gateway 104 is checked (step 510).

As a result of the check in the step 510, if the unstable termination of the application service due to the connection loss is confirmed, the first offline operation is activated to perform some data processing (step 512).

The application service is stored in the data temporary storage 1026 in the mobile grid client 102 as a second offline operation after the connection request is transferred to the predetermined secondary gateway through the replication technique and approved (step 514). The previous log file related to the execution is forwarded to the secondary gateway (step 516). At this time, the forwarded log file tracks the existence and location of data required to perform an application service and is used to confirm data validity.

Thereafter, to determine the suitability of the data, if there is a problem in the suitability of the data, the secondary gateway sends a data synchronization request message to the master gateway, and then the log file sent by the mobile grid client 102 This process is repeated as a reference to check data validity (step 518). When this process is completed, the process from step 506 is performed again.

On the other hand, if the connection loss does not occur as a result of the check in step 510, it is determined that the application service is normally terminated to clean up the data temporary storage 1026 of the mobile grid client 102 and at the same time the job completion notification of the mobile grid Receive from gateway 104 (step 520).

Therefore, the offline operation of the mobile grid gateway replication system of the wireless grid network may compensate for the delay of application service execution due to the connection loss between the mobile grid client and the mobile grid gateway.

In the foregoing description, the present invention has been described with reference to preferred embodiments, but the present invention is not necessarily limited thereto. Those skilled in the art will appreciate that the present invention may be modified without departing from the spirit of the present invention. It will be readily appreciated that branch substitutions, modifications and variations are possible.

As described above, the present invention, unlike the conventional grid computing using a wired communication network, grants user authentication and resource use rights, performs registration for performing the requested application service, and provides service level objectives and agreements. In addition, after configuring the mobile user-side policy through service budget adjustment, requesting selection of grid computing resources, and receiving it, sub-tasks are allocated to selected grid computing resources to be performed and monitored. By returning the results of the sub-tasks and notifying the billing details thereof, the mobile grid gateway replication system can be constructed based on the reliability guarantee replication scheme for the wireless grid network.

In addition, by building a mobile grid gateway replication system, it is possible to discover a set of resources that satisfy the user's SLO, SLA, and budget-based policy details according to the status of resources that change in time and space in the mobile grid environment. QoS can be optimized by maintaining a certain level of resources.

At the same time, the grid computing infrastructure affected by the loss of connection due to frequent movement of wireless devices in the mobile network environment, the inability to guarantee data transmission due to low bandwidth, and the inferior capacity of the wireless device's resource provisioning capability can be applied to the method of replicating the mobile grid gateway and offline. By protecting through operation, it is possible to prevent a decrease in reliability due to the integration of wired and wireless environments, and to guarantee a certain level of performance defined in the policy, thereby improving performance of application service performance.

In addition, by building a mobile grid gateway replication system based on policy quorum, after allocating application service tasks to resources, you can check the progress through continuous monitoring and reconfigure resources to flexibly cope with resource deterioration and failure. Users can be assured of a reliable job execution environment in a complex resource distribution environment.

Claims (22)

A mobile grid gateway replication system in a wireless grid network, A mobile grid client that acts as an interface for the mobile user, inputs the characteristics and environment of the requested application service, determines the characteristics of the resources, and transmits the requirements; Performs user authentication and authorization authentication while communicating with the mobile grid client and converges available resource quorum by converging the required performance of the resource and the required level of the grid service in the form of service level objective (SLO) and service level agreement (SLA). A mobile grid gateway that generates and plans and executes work for the application service through resources obtained within the available resource quorum Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 1, The system, A billing information storage database for storing usage history for each user and application service to charge a fee for performing the application service; An agent-based resource monitoring server that monitors real-time resource status information and returns information about it; Resource status monitoring agent that monitors the status of subtasks running on each resource node Mobile grid gateway replication system of the wireless grid network further comprising. The method of claim 1, The mobile grid client, A gateway connection manager that communicates through a client interface and acts as an interface, An application service registration module that supports entry and registration of information about an application or service required to be performed; A first policy manager for delivering a user side policy history determined by the SLO or the SLA, and preparing and delivering a user side estimated cost; A grid resource request module requesting the return of a quorum set that satisfies the policy details reflecting the requirements of previously registered application services; A user UI module for performing an association between the mobile grid client and the mobile grid gateway; Temporary storage of data temporarily storing the index and log information of datasets running application services in the form of binary files to support offline operation. Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 3, wherein The first policy manager, An SLO determination module for setting the SLO; An SLA agreement module for determining the SLA according to the set SLO; Service budget adjustment module to budget according to the SLO and the SLA Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 1, The mobile grid gateway, A client interface for accepting, authenticating, authorizing, and policy policing of the mobile grid client; An internal linkage component that performs internal logic, Gateway middleware that relays the role between the parent component group and the child grid service manager. Grid service manager that performs application service allocation to grid resource nodes, selection of quorum resource sets, monitoring the performance of the process, or charging Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 5, wherein The client interface, An authentication manager that communicates with the gateway connection manager; An authorization manager for granting session-specific permissions and specifying a range of execution of the mobile users connected through the authentication manager; SLA conclusion module that collects user-side policy details, requests service provider-side policy details, and determines policy for calculating optimal performance-to-cost Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 5, wherein The internal linkage component, Manager monitoring UI which is a management interface provided for management of each mobile grid gateway, A CONTEXT manager that manages and tracks the workflow of performing internal logic to convey the degree of execution of the logic and whether specific events occur to other managers, components, or module components; Logic mapping module for notifying the other module of the specific event recognized through the CONTEXT manager Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 5, wherein The gateway middleware, The authentication manager that is linked to the grid service account according to the user's request and requests to issue the account, A service manager that manages information about application services for which registration is requested; A second policy manager for providing a service provider-side policy history, and real-time resource status for dynamic generation of the service provider-side policy history; Replica Manager Ensures Reliability in Wireless Grid Environments Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 8, The replica manager, A replica catalog that manages the location and status of the created mobile grid gateway group, and selects a gateway that replaces it when a problem occurs in the master mobile grid gateway; Replication module that calculates the number of gateways that can act as the secondary mobile grid gateway and performs information exchange and data transmission for information synchronization between gateways Mobile grid gateway replication system of a wireless grid network comprising a. The method of claim 5, wherein The grid service manager, An access control manager that provides the requested grid service account and controls and manages access to that account's grid computing infrastructure; Grid resource management module for receiving the real-time resource status information from the external agent-based resource monitoring server to configure the optimal resource group based on the quorum; A data transmission module for delivering a data set necessary for performing the application service to a selected resource node group; A process control module for monitoring an abnormality of the sub-tasks being performed in each resource node by using a resource status monitoring agent; A process scheduler for allocating subtasks running on each resource node using budget and deadline scheduling techniques; Billing manager for storing the usage history for each user and application service in the database for storing the billing information to charge the charges for performing the application service, and delivers the optimal billing history to the mobile grid client in accordance with the application service notification Mobile grid gateway replication system of a wireless grid network comprising a. A mobile grid gateway replication method of a wireless grid network, 1 step of the mobile grid gateway granting user authentication and resource usage rights according to the session connection request of the wireless device user; Performing a registration for performing an application service requested by the user, and returning information on the validity period; A third process of configuring a mobile user side policy by adjusting a service level target and agreement and a service budget for concluding a policy-based service contract between the user and a grid service provider; Requesting selection of grid computing resources based on a user-side policy descriptor and receiving the fourth step; A fifth process of allocating sub-tasks to the selected grid computing resources for execution of an application service registered with the mobile grid gateway; A sixth process of performing the assigned sub-tasks and monitoring such performance; A seventh process of returning the result of the performed sub-tasks to the user and notifying the billing details thereof; Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The first process, A first step 1-1 of broadcasting a notification message notifying the wireless device of the wireless network participation to a neighboring gateway group; A first and second process of selecting a gateway set by measuring a gateway signal strength of a response sent from the gateway group; A first to third process of selecting one master gateway and at least two secondary gateways according to gateway proximity, signal strength, and workload degree among the selected gateway set; 1-4 process of receiving an acceptance message from the selected gateway Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The second process, A step 2-1 of describing a request for execution of the application service for registering the application service processing job; Step 2-2 of registering the executable file and the related data set of the application service to the mobile grid gateway, Step 2-3 where an effective grid account accessible to the grid service is allocated for each user or application service Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The third process, A 3-1 process of describing an SLO for determining a work target according to performing the application service; A 3-2 process of describing an SLA for determining an application service level according to performing the application service; 3-3 process of determining a service budget based on the service characteristics according to the SLOs and SLAs described above Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 14, The SLO and the SLA are described in the form of an XML document. The method of claim 11, The fourth process, Step 4-1 of including the adjusted service level target, agreement, and service budget in a grid resource selection request; Step 4-2 of transmitting a grid resource request descriptor to the SLA conclusion module of the mobile grid gateway, A 4-3 process of requesting an individual policy for each service provider of an available resource group through the mobile grid gateway; 4-4 process of comparing each element item described in the user policy and the service provider policy; 4-5 process of notifying the determined policy through the policy manager and delivering it to the grid resource manager; Processes 4-6 of receiving a resource quorum set generated from the mobile grid gateway; Steps 4-7 for inquiring whether to use the returned resource quorum set Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The fifth process, Step 5-1 of accessing the grid service manager using the assigned grid account; Step 5-2 of allocating the divided subtasks of the registered application service to the selected resource quorum set Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The sixth process, A 6-1 process of performing individual state monitoring on the sub-tasks which is allocated to each grid resource; Step 6-2 of notifying the mobile grid gateway when a task error occurs due to a logical error of a process or a physical defect of a resource in the performed subtask; Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 11, The seventh process, Step 7-1 of deriving a result for the performed sub-tasks; Step 7-2 of returning a data set on the derived result to the user; Process 7-3 comparing the expected charges against actual charges resulting from the execution of the application service based on user policy, service provider policy, and returned grid computing resource group, and notifying and returning related information Mobile grid gateway replication method of a wireless grid network comprising a. A mobile grid gateway replication method of a wireless grid network, Process of requesting selection of available gateways by identifying application service characteristics in mobile grid gateway, creating SLO and SLA-based policies for selecting gateway candidate groups, Requesting selection of a gateway candidate group using the generated policy, and returning location and status information on the gateway resource group according to the gateway candidate group;  Recording corresponding information of the selected gateway candidate group, calculating an optimal number of replicas, selecting a master / secondary gateway, Checking whether the operation state of the selected master gateway is normal according to a preset period and performing information synchronization between the gateways if the operation is normal. Mobile grid gateway replication method of a wireless grid network comprising a. A mobile grid gateway replication method of a wireless grid network, The mobile grid gateway delivers a binary file including an index of each individual data of the data set and a performance log to the mobile device data temporary storage to perform an application service. Compensating for low network quality according to a wireless environment in the temporary data storage and performing offline operation support to prevent delay of execution time of the application service. Mobile grid gateway replication method of a wireless grid network comprising a. The method of claim 21, The offline operation, After the connection between the gateway connection manager and the grid gateway, the SLA, SLO, and budget related policy matters necessary for registering and executing the application service are set, and the application service is performed on the assigned grid virtual resource. Generating index information of the application service, generating a log of completed execution contents, and creating an index table for a data set to be performed afterwards; Checking whether a service manager has completed execution of the application service from a context manager;  Checking the connection loss between the mobile grid client and the mobile grid gateway when the application service is completed; Performing a first offline operation when an unstable termination of an application service due to the connection loss is confirmed; A process of performing a secondary offline operation after receiving a connection request through a replication technique to a predetermined secondary gateway and receiving an approval; If the connection loss does not occur, it is determined that the application service is normally terminated to clean up data temporary storage and receive a job completion notification from the mobile grid gateway. Mobile grid gateway replication method of a wireless grid network comprising a.

Images