1. Mike Jackson EPCC michaelj@epcc.ed.ac.uk@epcc.ed.ac.uk OGSA-DAI Architecture + Extensibility OGSA-DAI Tutorial GGF17, Tokyo

2. 11 May 2006http://www.ogsadai.org.uk/2 Outline “Let us out” –Exposing data to clients – the server’s perspective “Let us in” –Getting to the data – the client’s perspective “More, more more…” –Extending OGSA-DAI

3. 11 May 2006http://www.ogsadai.org.uk/3 Data resources Relational –MySQL –Microsoft SQL Server –Oracle –IBM DB2 –PostGres –HSQL XML –eXist –Xindice File system –SwissPROT –OMIM –Text –Binary Relational XMLDB Files

4. 11 May 2006http://www.ogsadai.org.uk/4 Data resource accessors Interfaces between data resources and OGSA-DAI Relational –JDBC drivers –java.sql.* XML –XMLDB API and compliant drivers –org.xmldb.api.* File system –Java file and directory utilities –java.io.* OGSA-DAI extensibility point Data Resource Accessor Relational XMLDB Data Resource Accessor Files

5. 11 May 2006http://www.ogsadai.org.uk/5 Data service resources Data Resource Accessor Relational XMLDB Data Resource Accessor Data Service Resource Files Data Service Resource SQLOne XMLOne FilesOne

6. 11 May 2006http://www.ogsadai.org.uk/6 Data service resources OGSA-DAI’s core functionality Manages –Access to a data resource via a data resource accessor –Execution of data-related activities –Data caching and streaming of data to and from clients –Creation, access and termination of sessions Exposes data service resource properties –Information about a data resource –Information about supported activities –Information about current requests

7. 11 May 2006http://www.ogsadai.org.uk/7 Requests and responses Data Resource Accessor Data Service Resource SQLOne Relational SQL Query Results Perform Document Response Document ResultSet

8. 11 May 2006http://www.ogsadai.org.uk/8 Requests and responses Request –A connected collection of activities that the data resource should execute –XML perform document submitted by a client Activity –An individual data-related operation –0 or more inputs and 0 or more outputs Response –Status of execution of a request possibly with result data –XML response document returned to a client OGSA-DAI engine –Parses requests, executes activities, builds responses

9. 11 May 2006http://www.ogsadai.org.uk/9 Activities Relational –SQL query, update, stored procedure, bulk load, extract logical and physical schema –Convert ResultSet to WebRowSet, ResultSet to CSV, ResultSet to bytes, relational database schema to XML –Project ResultSet or WebRowSet onto a column –Extract bytes from ResultSet XMLDB –Resource and collection management, XPath, XQuery, XUpdate, bulk load Files –List directory, create, read, write and update files –Index files, search indexed files

10. 11 May 2006http://www.ogsadai.org.uk/10 Activities Transformation and Compression –GZIP compression, ZIP archive –XSLT –Project CSV data onto a column –Distribute numerical data onto spaces –Create random sample of data Delivery –From and to URLs, files, GridFTP, remote data service resources, SOAP attachments –To servlets, SMTP, resource properties Factory –Create and destroy persistent and transient data service resources Relational multi-resources –Bag and resilient queries OGSA-DAI extensibility point

11. 11 May 2006http://www.ogsadai.org.uk/11 Data services Data Resource Accessor Relational XMLDB Data Resource Accessor Data Service Resource Files Data Service Resource SQLOne XMLOne FilesOne Data Service

12. 11 May 2006http://www.ogsadai.org.uk/12 Data services Web services Expose 0..N data service resources to the outside world Two flavours –OGSA-DAI WSRF services –Compliant with the Web Services Resource Framework –Implemented using Globus Toolkit (4.0+) –OGSA-DAI WSI services –Compliant with vanilla WSDL –Implemented using Apache Axis (1.2.1 or 1.2RC3) Service-level operations –GetVersion – which version of OGSA-DAI is running –ListResources – which data service resources does a service expose –Deploy / Undeploy –Expose / withdraw data service resources to / from the outside world

13. 11 May 2006http://www.ogsadai.org.uk/13 Data services Data service resource-specific operations –Perform –Receive requests for a data service resource and return responses –GetFully / GetNBlocks / GetBlock –Stream data from a data service resource –PutFully / PutBlock –Stream data to a data service resource –GetResourceProperty / GetMultipleResourceProperties / QueryResourceProperties –Expose information about the data service resource state

14. 11 May 2006http://www.ogsadai.org.uk/14 Clients Data Resource Accessor Data Service Data Service Resource Client Perform Document SQLOne Relational Perform Document SQL Query ResultSet SQL Query Results Response Document Authorization points

15. 11 May 2006http://www.ogsadai.org.uk/15 Identifying a data service resource Data Service Data Service Resource Client Request SQLOne http://host:port/services/axis/DataService/DAISQLOne http://host:port/services/wsrf/DataService Request

16. 11 May 2006http://www.ogsadai.org.uk/16 Clients and the client toolkit Clients interact with data services via SOAP over HTTP –Deduce service interface from service WSDL description –Construct SOAP request to invoke operation –Parse SOAP response from service –Resource identification scheme must be assumed from WSDL namespace OGSA-DAI client toolkit: –Construct and submit requests in Java not XML –Toolkit handles SOAP request construction and response parsing –Renders OGSA-DAI service types transparent –Java abstractions of –Data services –Data service resource IDs and session IDs –Requests and responses –Activities

17. 11 May 2006http://www.ogsadai.org.uk/17 Relational multi-resources Multiple Data Resource Accessor Multi Resource Data Service Resource MultiOne Data Service Data Service Relational SQL Query Multi Results SQL Query Results x M Data Service Resource Data Service Relational SQL Query Results SQL Query Results SQL Query Results

18. 11 May 2006http://www.ogsadai.org.uk/18 Factory resources Factory Data Resource Accessor Factory Resource FactoryOne Data Service Resource Create Resource InstanceOne Instance Data Resource Accessor Instance Resource InstanceOne

19. 11 May 2006http://www.ogsadai.org.uk/19 Extending OGSA-DAI Data resource accessors –Use your own data resources within OGSA-DAI –Expose local or remote data resources –Expose virtual resources created by aggregation or integration Resource management –Creation and destruction of persistent and transient data service resources –Another type of data resource accessor Activities –Use your own activities that can be executed by a data service resource –Can be resource specific e.g query or update –Or generic e.g. transformation, compression, delivery, resource management, monitoring Authorization –Resource access –Activity execution –In application-specific ways

20. 11 May 2006http://www.ogsadai.org.uk/20 Summary The OGSA-DAI layer cake –Data –Data resource accessors –Data service resources –Data services –Clients Extending OGSA-DAI –Data resource accessors –Activities –Authorization