전산학과 이재승 The Physiology of the GRID I. Foster, C. Kesselman, J. Nick, and S. Tuecke Open Grid Service Infrastructure WG, Global Grid Forum, June 22, 2002

2 Table of Contents 1. Introduction 2. The Need for Grid Technologies 3. Background 4. An Open Grid Services Architecture 5. Application Example 6. The OGSA Service Model 7. Network Protocol Bindings 8. Higher-Level Services 9. Related Work 10. Conclusion 11. Discussion

3 1. Introduction  Problem  In e-business and e-science, we need to integrate services across distributed, heterogeneous, dynamic “virtual organizations” formed from the disparate resources within a single enterprise and/or from external resource sharing and service provider relationships.  This integration can be technically challenging because of the need to achieve various QoS when running on top of different native platforms.

4 1. Introduction  Solution: Open Grid Services Architecture  defines a uniform exposed service semantics (the Grid service)  defines standard mechanisms for creating, naming, and discovering transient Grid service instances  provides location transparency and multiple protocol bindings for service instances  supports integration with underlying native platform facilities  defines mechanisms for creating and composing distributed systems including lifetime management, change management, and notification

5 1. Introduction  Focus  on the services that respond to protocol messages  how Grid technologies can be aligned with Web services technologies (We call this alignment an Open Grid Services Architecture (OGSA)) - WSDL interfaces for a Grid service - service instance supporting reliable and secure invocation - lifetime management, notification, policy management, credential management, and virtualization - discovery of Grid service instances and creation of transient Grid service instances.  on commercial applications rather than the scientific and technical applications Requirements: - seamless integration with existing resources and applications - tools for workload, resource, security, network QoS, and availability management.

6 2. The Need for Grid Technologies  The Evolution of Enterprise Computing  Rise of the Internet and e-business  intelligent network  IT services decomposition is also occurring inside enterprise IT facilities. - different programming models, commodity servers - need to distribute and cache content closer to the edge of the network  Need for highly robust IT infrastructure  IT systems integrators take on the burden of re-integrating distributed compute resources with respect to overall QoS.

7 2. The Need for Grid Technologies  Service Providers, B2B Computing  emergence of SPs demands: - dynamic resource allocation in accordance with service-level agreement policies - consistent response times and high levels of availability  B2B Collaboration demands: - security, auditability, availability, service level agreements, and complex transaction processing flows

8 3. Background  The Globus Toolkit  community-based, open-architecture, open-source set of services and software libraries that support Grids and Grid applications.  security, information discovery, resource management, data management, communication, fault detection, and portability  GRAM protocol, gatekeeper service: secure, reliable, service creation and management  MDS-2: information discovery, data modeling, local registry  GSI: single sign on, delegation, and credential mapping.  essential elements of a service-oriented architecture

9 3. Background  The Globus Toolkit (cont)

10 3. Background  Web Services  focus on simple,Internet-based standards to address heterogeneous distributed computing  SOAP: provides a means of messaging between a service provider and a service requestor  WSDL: Service interfaces are defined abstractly in terms of message structures and sequences of simple message exchanges  WS-Inspection: comprises a simple XML language and related conventions for locating service descriptions published by a service provider.  Web Services - advantages  dynamic discovery and composition of services in heterogeneous environments  Web services can exploit numerous tools and extant services

11 4. An Open Grid Services Architecture  Service Orientation and Virtualization  OGSA focus on services: computational resources, storage resources, networks, programs, databases, and the like are all represented as services  critical requirement is interoperability  service-oriented view allows us to address the need for standard interface definition mechanisms, local/remote transparency, adaptation to local OS services, and uniform service semantics  Virtualization allows for consistent resource access across multiple heterogeneous platforms with local or remote location transparency   OGSA supports local and remote transparency with respect to service location and invocation.  OGSA provides for multiple protocol bindings

12 4. An Open Grid Services Architecture  Service Semantics: The Grid Service  a Web service that provides a set of well-defined interfaces and that follows specific conventions  The interfaces address discovery, dynamic service creation, lifetime management, notification, and manageability  The conventions address naming and upgradeability  The interfaces and conventions are concerned with behaviors related to the management of transient service instances

13 4. An Open Grid Services Architecture  The Grid Service – Upgradeability Conventions and Transport Protocol  Services within a complex distributed system must be independently upgradeable - versioning, compatibility must be managed  A service’s description indicates the protocol bindings that can be used to communicate with the service. - Reliable Service Invocation - Authentication

14 4. An Open Grid Services Architecture  The Grid Service – Standard Interfaces  Discovery  Dynamic Service Creation  Lifetime management  Destroy  Notification  Other Interfaces: authorization, policy management, concurrency control …. (in the near future)

15 4. An Open Grid Services Architecture

16 4. An Open Grid Services Architecture  The Role of Hosting Environments  OGSA answers the `what-question', not the `how-question'  OGSA does not address issues of implementation programming model, programming language, implementation tools, or execution environment.  OGSA specifies interactions between services in a manner independent of any hosting environment

17 4. An Open Grid Services Architecture  Using OGSA Mechanisms to Build VO Structures Simple Hosting EnvironmentVirtual Hosting EnvironmentCollective Operations

18 5. Application Example

19 6. The OGSA Service Model  Creating Transient Services: Factories  Factory interface creates new Grid service instances.  The Factory interface’s CreateService operation creates a requested Grid service and returns the GSH and initial GSR for the new service instance.  Service Lifetime Management  Soft state approach: Grid service instances are created with a specified lifetime.  advantages: - A client knows, or can determine, when a Grid service instance will terminate - A hosting environment is guaranteed that resource consumption is bounded, even in the face of system failures outside of its control.

20 6. The OGSA Service Model  Managing Handles and References  The result of a factory request: a GSH and a GSR  GSH (Grid Service Handle): is guaranteed to reference the created Grid service instance in perpetuity  GSR (Grid Service Reference): is created with a finite lifetime and may change during the service’s lifetime. (WSDL document describes how to communicate with the Grid Service)  how does one establish communication with a Grid service given only its GSH?  define a handle-to-reference mapper interface (Handle Map)  Each Grid service instance should be registered with at least on handleMap, called home handleMap, which is included in the GSH  Identification of the home handleMap is established by requiring that all home handleMap services be identified by a URL

21 6. The OGSA Service Model  Service Data and Service Discovery  A set of service data is associated with each Grid service instance.  GridService interface defines a standard WSDL operation, FindServiceData, for querying and retrieving service data.  One application of the GridService interface’s FindServiceData operation is service discovery  Registry interface is used to register a GSH and the GridService interface’s FindServiceData operation is used to retrieve information about registered GSHs.  Notification  OGSA notification framework allows clients to register interest in being notified of particular messages (the NotificationSource interface)  OGSA supports asynchronous, one-way delivery of such notifications (NotificationSink).

22 6. The OGSA Service Model  Change Management  Any changes made to the definition of a Grid service must be reflected through new interface names  This feature allows clients that require Grid Services with particular properties to discover compatible services.

23 7. Network Protocol Bindings  Four Primary Requirements  Reliable transport: support for reliable service invocation ex) HTTP-R  Authentication and delegation: support for communication of proxy credentials to remote sites ex) TLS extended with proxy credential support  Ubiquity: It must be possible for any arbitrary pair of services to interact  GSR Format: ex) WSDL document

24 8. Higher-Level Services  Distributed data management services  Workflow services  Auditing Services  Instrumentation and monitoring services  Problem determination services for distributed computing  Security protocol mapping services

25  some current Globus Toolkit protocols  a potential refactoring to exploit OGSA mechanisms 8. Higher-Level Services

26 9. Related Work  Globus Toolkit v2.0  factory (GRAM gatekeeper), registry (GRAM reporter, MDS-2), secure remote invocation with delegation (GSI), reliable remote invocation (GRAM) is form Globus Toolkit v2.0  differences: how these difference mechanisms are integrated no uniform interface definition language  Distributed Object System  OGSA exploits ideas developed previously in systems such as Eden, Argus,CORBA, SOS, Spring, Globe, Mentat, and Legion  In contrast to CORBA, OGSA like Web services addresses directly issues of secure interoperability and provides a richer interface definition language

27 9. Related Work  Applications of Web services mechanisms to Grid  De Roure et al. propose a “Semantic Grid,” by analogy to the Semantic Web, and propose a range of higher-level services  Work on service-oriented interfaces to numerical software in NetSolve and Ninf  Etc  Sun Microsystems’ JXTA system addresses several important issues encountered in Grids, including discovery of, and membership in, virtual organizations groups.”

Conclusion  OGSA supports the creation, termination, management, and invocation of stateful, transient services as named, managed entities with dynamic, managed lifetime.  Within OGSA, everything is represented as a Grid service, that is, a service that conforms to a set of conventions (expressed using WSDL) for such purposes as lifetime management, discovery of characteristics, notification, and so on.  The merits of the service-oriented model: - All components of the environment are virtualized  By providing a core set of consistent interfaces, we facilitate the construction of services that can be treated in a uniform way - Virtualization also enables mapping of multiple logical resource instances onto the same physical resource, composition of services regardless of implementation  By integrating support for transient, stateful service instances with existing Web services technologies, OGSA extends significantly the power of the Web services framework, while requiring only minor extensions to existing technologies.

Discussion  Performance, end-to-end QoS  Development Environment, Hosting Environment, Implementation Issues  Deployment Issues  commercial service 에 바로 적용 가능한가 ?  과연 dominant 한 computing 환경으로 자리 잡을 수 있을까 ?