©ferenc vajda 1 open grid services architecture ferenc vajda [email protected] computer and...
TRANSCRIPT
©Ferenc Vajda
1
Open Grid Services Architecture
Ferenc [email protected]
Computer and Automation Research Institute
Hungarian Academy of Sciences
2
©Ferenc Vajda
GRID
Past: Globus
Present: OGSA
Future: Semantic Grid
“Past is history, future is mystery”
©Ferenc Vajda
3
Roots of OGSA
Globus Toolkit: GTPL (Globus Toolkit Public License)GT3: Platform Globus (Commercially supported)
“The Anatomy of the Grid”
“The Physiology of the Grid” papers by Ian Foster et al.
Grid Forum GGF (Global Grid Forum)
OGSA Working Group
©Ferenc Vajda
4
Main issues of “The Anatomy of the Grid”
-SDK (Software Development Kit)
• Virtual Organization (VO)
• Nature of Grid Architecture
-Interoperability
-Protocol
-Services
-API (Application Programming Interface)
©Ferenc Vajda
5
-Applications
Main issues of “The Anatomy of the Grid” 2.
• Architecture Description
-Fabric: interfaces to local control
-Connectivity: communicating easily and securely
-Resource: sharing single resources
-Collective: coordinating multiple resources
©Ferenc Vajda
6
The layered Grid architecture
“The Anatomy of the Grid”
©Ferenc Vajda
Main issues of “The Physiology of the Grid”
-Business-to-Business (B2B) Computing
-Web Serviceso SOAP
o WSDL
o UDDI
o WSFL
o WS-Inspection
• Grid technologies
-Enterprise Computing
-Service Providers (SPs)
• Background
-Globus Toolkit
©Ferenc Vajda
8
Building an Open Grid
©Ferenc Vajda
9
OpenStandards
Building an Open Grid
©Ferenc Vajda
10
OpenStandards
OpenSource
Building an Open Grid
©Ferenc Vajda
11
OpenStandards
OpenSource
OpenInfrastructure
Building an Open Grid
©Ferenc Vajda
12
OpenStandards
OpenSource
OpenInfrastructure
OpenGrid
Credit to Ian Foster
Building an Open Grid
©Ferenc Vajda
13
Incr
ease
d fu
nctio
nalit
y,st
anda
rdiz
atio
n
Time
Customsolutions
Open GridServices Arch
GGF: OGSI, …(+ OASIS, W3C)
Multiple implementations,including Globus Toolkit
Web services
Globus Toolkit
Defacto standardsGGF: GridFTP, GSI
X.509,LDAP,FTP, …
App-specificServices
Grid and Open Standards
Credit to Ian Foster
©Ferenc Vajda
14
Open Grid Services Architecture
-Language interoperability
From Web Services
Standard interface definition mechanisms
-Interface and implementation
(multiple protocol bindings)
-Local/remote transparency
-Service semantics
©Ferenc Vajda
15
Open Grid Services Architecture 2.
• Further services: resource management, authorization, etc.
From Grids
• Lifecycle management
• Reliability and security models
• Discovery
©Ferenc Vajda
16
Open Grid Services Architecture 3.
• Objective: to integrate services across
- distributed
- heterogeneous
- dynamic virtual organizations
• Solution:
-standard mechanisms for creation, naming, discovery
-location transparency, multiple protocol binding
-integration with hosting environment
©Ferenc Vajda
17
OGSI
TransportProtocolHosting EnvironmentHosting Environment
Host. Env. & Protocol Bindings
OGSA Platform services: registry,authorization, monitoring, data
access, etc., etc.
More specialized &domain-specific
services
Models for resources &
other entitiesO
therm
odels
Environment-specificprofiles
Domain-specificprofiles
OGSAPlatform
GWD-R (draft-ggf-ogsa-platform-3) Editors:Open Grid Services Architecture Platform I. Foster, Argonne & U.Chicagohttp://www.ggf.org/ogsa-wg D. Gannon, Indiana U.
OGSA Platform
©Ferenc Vajda
18
Principal elements of OGSA Platform
• Open Grid Services Infrastructure (OGSI)
• OGSA Platform Interfaces
• OGSA Platform Models
©Ferenc Vajda
19
OGSA Platform Profiles
• Sets of domain-specific services
Environment, domain-specific supplements:
• Protocol bindings
• Hosting environment bindings
©Ferenc Vajda
20
Open Grid Services Infrastructure
•Service group
Key Features:
• Grid Service description and instances
• Service state, metadata and introspection
• Naming and name resolution
• Fault model
• Lifecycle
©Ferenc Vajda
21
OGSI
OGSI = Grid Technologies + Web Services
In OGSI everything is represented as a Grid service.
Service: a network-enabled entity that provides some capabilities through the exchange of messages.
©Ferenc Vajda
22
OGSI
Most important components:
•Grid Service
• Factory
• Registry
• HandleMap
©Ferenc Vajda
23
The Grid Service
-Grid service instance
• A WSDL-defined service that conforms to a set of conventions relating to its interface and behaviors.
•Description composed of two parts:
-Grid service description
-Describes a client’s interaction with service instances:syntax and semantics (PortType)
-Can be used by any number of Grid service instances
©Ferenc Vajda
24
The Grid Service 2.
-Has one or more Grid Service References (GSRs)
-Grid service instance
-Embodies a state
-Has one or more unique Grid Service
Handles (GSHs)
©Ferenc Vajda
25
The Grid Service 3.
•GSR (Grid Service Reference): abstraction for e.g. protocol binding, network address, etc.
•typed (characterized by the offered capability)
•interfaces (set of operations)
• invocation (by sequences of message exchange)
•created/destroyed (dynamically)
•GSH (Grid Service Handle): globally unique name
©Ferenc Vajda
26
Factories: Creating Transient Services
Factory = Factory interface + implemented service
CreateService operation: -creates a requested Grid service
-returns a GSH + an initial GSR
©Ferenc Vajda
27
Handles and References
-returns a valid GSR
• GSH: reference for service instance “forever”
• GSR: changes during lifetime
HandleMap: handle-to-reference mapper
-takes a GSH
©Ferenc Vajda
28
Registry and Notification
Semantics: protocol binding property
e.g. SOAP/HTTP protocol
Registry: registry interface + service data elements
(info about GSH)
Notification
Subscribe operation
NotificationSource interface
NotificationSink: message (“keep alive”)
“push” model (FindServiceData “pull” model)
©Ferenc Vajda
29
Client runtime architecture
Credit to S. Tuecke et al. (Grid Service Specification)
©Ferenc Vajda
30
GSH Resolving
Credit to S. Tuecke et al. (Grid Service Specification)
©Ferenc Vajda
31
Open Grid Service Infrastructure
Implementation
Servicedata
element
Other standard interfaces:factory,
notification,collections
Hosting environment/runtime(“C”, J2EE, .NET, …)
Servicedata
element
Servicedata
element
GridService(required)
Dataaccess
Lifetime management• Explicit destruction• Soft-state lifetime
Introspection:• What port types?• What policy?• What state?
Client
Grid ServiceHandle
Grid ServiceReference
handleresolution
Credit to Ian Foster
©Ferenc Vajda
32
1a. Request to Registry for sources of data about “x”
1b. Registry responds with
Factory handle2a. Request to Factory for access to database
2b. Factory creates GridDataService to manage access
2c. Factory returns handle of GDS to client
3a. Client queries GDS with XPath, SQL, etc
3b. GDS interacts with database
3c. Results of query returned to client as XML
SOAP/HTTP
service creation
API interactions
Registry
Factory
Grid Data Service
Client
XML / Relational database
OGSA-DAI (Data Access and Integration
©Ferenc Vajda
33
Security Challenges
Integrate
Extensible architecture
Using existing services
Implementation agnostic
Interoperate
Secure interoperability
Publishing QoS
Federation
Trust
Trust relationship
Trust establishment
Presumed trust
Assertions
©Ferenc Vajda
34
Grid Security Requirements
• Integrity
• Authentication• Delegation• Single sign-on
• Credential lifespan and renewal
• Authorization
• Privacy
• Confidentiality
©Ferenc Vajda
35
• Securing infrastructure
Grid Security Requirements 2.
• Policy exchange
• Secure logging
• Assurance
• Manageability
• Firewall traversal
©Ferenc Vajda
36
Components of Grid Security Model
©Ferenc Vajda
37
Security Architecture Building Blocks
©Ferenc Vajda
38
Security as Service
• Privacy service
• Authentication service
• Identity mapping service
• Authorization service
• I/O policy service
• Credential conversion service
• Audit service
• Profile service
©Ferenc Vajda
39
Semantic Grid
Ferenc Vajda
Computer and Automation Research Institute
Hungarian Academy of Sciences
©Ferenc Vajda
40
Data/Information/Knowledge
Data: observed facts
Information: organized and related facts with attributed properties
Knowledge: “sum of what is known”: concepts, objects with characteristics, principles, laws, know-how, etc.
Semantics: a term used for meaning, interpretation, knowledge through reasoning
©Ferenc Vajda
41
Different Evaluations of the Grid
1. Grid generations
• To link supercomputer centers
(e.g. I-way)
• Toolkit- and middleware-based
(e.g. Globus)
• Service-oriented (OGSA)
©Ferenc Vajda
42
Different Evaluations of the Grid 2.
2. Based on the technologies used
• Protocol-based
• Service-based
• Semantic Web based
3.Based on application requirements
• Data/computational Grid
• Information Grid
• Knowledge Grid
©Ferenc Vajda
43
Problems Related to Semantic Web
• Knowledge Evaluation
• Knowledge Representation
• Ontologies
• Agents
©Ferenc Vajda
44
Resource Description Framework (RDF)
-Set of triplets: subject, property,object
• Metadata: structured data about data
• Resource identification: Universal Resource Identifier (URI)
• Most common type of URI: Uniform Resource Locator (URL)
• Qualified URI: URI + fragment identifier
• Concepts:
-Graph model
©Ferenc Vajda
45
RDF 2.
Subject ObjectProperty
-Data types: based on XML Schema
-Vocabulary: URI-based (Both nodes and arcs)
©Ferenc Vajda
46
RDF 3.
©Ferenc Vajda
47
What is an Ontology?
Greek: ontos = being, logos = science
• world view regarding a domain
• shared understanding
• definitions, inter-relationship
• conceptualization
©Ferenc Vajda
48
What does an Ontology look like?
• vocabulary of terms
• specification of their meaning (i.e. definitions)
- highly informal (natural language)
- semi-informal (restricted, structured form of natural language)
- semi-formal (artificial, formally defined language)
- rigorously formal (formal semantics, proofs, completeness)
©Ferenc Vajda
49
Use of Ontologies
• communication (between people and organizations)
• system engineering (specifications, reusable components)
• inter-operability (between systems)
©Ferenc Vajda
50
Ontologies
• Web Ontology Language (OWL)
• Ontology: defines the terms used to describe and represent an area of knowledge
-taxonomy: object classification + relationship among them (properties and inheritance of properties)
-inference rules
• DAML (DARPA = Defense Advanced Project Agency
Agent Markup Language)
©Ferenc Vajda
51
Agents
Agent: Capability to understand and integrate diverse information resources (based on domain ontologies)
©Ferenc Vajda
52
Agents 2.
©Ferenc Vajda
53
Semantic Web Layers
Credit to Berners-Lee (XML2000 address)
©Ferenc Vajda
54
Semantic Grid
©Ferenc Vajda
55
Semantic Grid
Basis:
• Metadata enabled
Goal:
Grid + Semantic Web
• Ontologically principled
New e-Science infrastructure
©Ferenc Vajda
56
Services
e.g. -semantic database integration
-semantic workflow description
• Base services
-data/computational services (network access, resource allocation and scheduling, data shipping, etc.)
-information services (queryprocessing, event notification, instrumentation management,
etc.)
• Semantic services
©Ferenc Vajda
57
Services 2.
-application
• Knowledge services
-acquisition
-modeling
-publishing, use and maintenance
-resource management
©Ferenc Vajda
58
Knowledge Grid Architecture
Credit to Carole Goble et al.
©Ferenc Vajda
59
Roles of Ontologies
Credit to Carole Goble et al.
©Ferenc Vajda
60
The term ‘procedure’ used by one tool is translated into the term ‘method ‘ used by the other via the ontology, whose term for the same underlying concept is ‘process’. procedure
viewer
translator
Ontology
method
library
give me the procedure for…
translator
here is the
METHOD for…
procedure = ???
procedure =
process
give me the
process for…
here is
the process for…METHOD =
process
??? = process
Roles of Ontologies (Example)
Credit to Rokhlenko Oleg
©Ferenc Vajda
61
Knowledge Services
Credit to Carole Goble et al.
©Ferenc Vajda
62
Typical Applications
• Service discovery
• Knowledge annotation
• Workflow composition
• Data interpretation
• Collaborative science
©Ferenc Vajda
63
Grid Service Discovery
Simple discovery
• attribute-base
• name lookup
• type matching
Semantic discovery
• matchmaking
• based on ontology description
©Ferenc Vajda
64
Brokering vs. Matchmaking
©Ferenc Vajda
65
Grid Service Discovery Framework
Ontology based description used by• service provider
• service requester
• service matchmaker
• service registry database
Matchmaking process
• comparison: request to registry
• decision: based on filters
• information
©Ferenc Vajda
66
Service Description
“What the service does”: service profile
“How it works”: ServiceModel
“How it is used”: ServiceGrounding
Description by RDF(S): Resource Description Framework Schema
Service profile
• description (human readable)
• functionalities
• functional attributes
©Ferenc Vajda
67
Service Description 2.
Credit to DAML-S White Paper
©Ferenc Vajda
68
Filtering
Independent filtering is based on
• context matching
• syntactic matching
- comparison of profiles
- similarity matching
- signature matching
• semantic matching
©Ferenc Vajda
69
myGrid project
©Ferenc Vajda
70
Role of Ontologies in myGrid