Metadatarich Service Oriented Grid Portals Mehmet Nacar mnacarcs

Metadata-rich Service Oriented Grid Portals Mehmet Nacar mnacar@cs. indiana. edu Mehmet Nacar

Outline Research Objectives Background Grid portals Research issues Architecture Contributions 9/14/2021 Mehmet Nacar 2

Research Objectives I Improve ease of Grid portal development We will research problems in federating multiple portal instances n n n Integrating portals based on the same code (e. g. all Grid. Sphere) Integrating portals with different code bases Integrating portals with non-portal services w identity management w authorization 9/14/2021 Mehmet Nacar 3

Research Objectives II We will investigate problems turning portals into Grid Web services n n It is the right architecture especially for Grid services WSRP is the Web Service standard WSRP must be coupled to other Web service standards for addressing and reliability This will require an examination and implementation of SOAP handlers 9/14/2021 Mehmet Nacar 4

What is a portal? Web based application provides n Personalization, single sign on and content aggregation Web, science and educational portal n n n Yahoo, Google CIMA, VLAB and LEAD One. Start and Oncourse Portal vs. portal framework n Jetspeed, Grid. Sphere and Sakai frameworks Modern portal framework vs. traditional portals n n Jetspeed, Grid. Sphere and u. Portal CGI-based portals, Reciprocal. Net 9/14/2021 Mehmet Nacar 5

Yahoo portal 9/14/2021 Mehmet Nacar 6

Portal page 9/14/2021 Mehmet Nacar 7

Portlets What is a portlet? n Building blocks of portals Portlet examples n n Calendar, chat, file transfer, weather and news. We have developed Grid portlets for OGCE w Myproxy, job submission and Grid. Ftp. Portlet container n Provides lifecycle management of portlets n Defines window states and portlet modes n Equips packaging and deployment JSR 168 specification defines n n Portlet container Portlet API 9/14/2021 Mehmet Nacar 8

Portal framework aggregates portlet contents Portal services are portal capabilities Portal services are evaluated in different aspects: n Administrative services w Account management, portlet management n User management services w User profile, layout management There also related portlets n n Administrative portlets are account creation, portlet registry User portlets are file transfer and calendar Portal front page displays anonymous content with login portlet What services involves when a user logs into the portal n Login service Authorization Service Layout Service 9/14/2021 Mehmet Nacar 9

Problem statement Portals and portlets are standardized but this opens new problem areas, particularly for Grid portals n n Federating multiple portal instances built from JSR 168 compatible frameworks. We must decouple portal “capabilities” from the container and make them into interoperable Web services. w Examples include identity management, authorization and logging. n Portlets and portals themselves can become Web services that can be federated. w But we must show these can be combined with other Web service registration, discovery, reliability, and others. 9/14/2021 Mehmet Nacar 10

Grid portals What makes a portal Grid-enabled? n n n Portlets provide access to Grid services Portal server leverage Grid credentials to access Grid services Provide capabilities/portlets w Grid. FTP, job submission and job monitoring What tools and services are available to build Grid portals? n n Java COG: Grid client programming API My. Proxy: credential management service OGCE: collection of portlets and tools for building portlets GAMA and PURSe: integrated Grid account management We will examine two motivating examples n n VLAB: traditional job submission portal for computational chemistry CIMA: scientific instrument and data portal 9/14/2021 Mehmet Nacar 11

VLAB Portal 9/14/2021 Mehmet Nacar 12

VLAB: The Virtual Laboratory for Earth and Planetary Materials Primarily a traditional job submission, monitoring, and management portal. Collaborative Grid services and portals support computational material science. Component based Grid portlet development makes application development easier. VLAB Challenges: n n Grid Portlets must be easy to develop using component libraries. HTML <form> actions in Grid portals typically have several steps: w w 9/14/2021 Stage data files in and out of the desired remote host. Run one or more executables. Keep track of job progress Store all of the information as “job pedigree” for reproducibility. Mehmet Nacar 13

Research issue: Grid components Grid portlet applications require dynamic interaction and fine-grained components. n n n Portlets themselves need to be built out of components Grid services mostly use request/response paradigm Grid portlets use web forms heavily w Compared to wikis, blogs, RSS-driven news portals, which have a different problem of content management. n Grid widgets can provide components for: w Proxy credential w Grid. Ftp operations w Job submissions, multi-staged jobs n Using widgets as tag libraries help to encapsulate reusable Grid components 9/14/2021 Mehmet Nacar 14

Grid Tag Libraries Grid tag libraries use the Java COG Kit abstractions Grid tags simplify association of composite Grid actions n And increase reuse of code There associated custom JSF tag extensions we’ve developed: n <o: proxy/>, <o: task. Graph/>, <o: task. Add/> and <o: context. Store/> Taskgraph supports multi-staged jobs <f: verbatim> <o: task. Graph id="my. Graph" method="#{taskgraph. test}" binding="#{taskgraph. taskgr}" > <o: proxy id=“myproxy" method=“#{proxy. create}" /> <o: task id="task 1" method="#{task. create}" type="File. Transfer. Input" /> <o: task id="task 2" method="#{task. create}" type="Job. Submit" /> <o: task id="task 3" method="#{task. create}" type="File. Transfer. Output" /> <o: task. Add id="taskadd 1" name="task 1" depends="task 2" method="taskgraph. add" /> <o: task. Add id="taskadd 1" name="task 2" depends="task 3" method="taskgraph. add" /> </o: task. Graph> </f: verbatim> 9/14/2021 Mehmet Nacar 15

CIMA Crystallography portal CIMA picture snapshot 9/14/2021 Mehmet Nacar 16

CIMA (Common Instrument Middleware Architecture) Primarily a data portal to online instruments Crystallographers collect data in participating laboratories and collaborate on samples. Portlets have to access data with group privileges. 9/14/2021 Mehmet Nacar 17

Research issue: federating portal instances CIMA and LSST Dark Energy Survey portal presents this problem: n n Multiple CIMA portal deployment instances are associated with different labs (IUMSC, Purdue, University of Sydney etc) Need to solve problem of portal identity federation. w Recall identity management is key portal capability. n Other problems (such as federated authorization) flow from this. Portal frameworks like Grid. Sphere, u. Portal and Jetspeed only support local “capabilities”. This approach has drawbacks: n n n Portal services tightly coupled to portal container and cannot being reused. Portal frameworks have proprietary services. Authorization capabilities of the portals define access rights for portlets not for contents These “capabilities” should be full-fledged Web services n This architecture greatly simplifies the federation problem 9/14/2021 Mehmet Nacar 18

Federating multiple portals I Requirements n n n Browser-based login is required for portals Each site maintains its own portal server Single sign on across the portal servers Federation can be done providing: n n n Several portals have the same code base but run different instances Integrating portals with different code bases Integrating portals with non-portal services, which have identity and authorization that is external to the portal framework. 9/14/2021 Mehmet Nacar 19

Federating multiple portals II 9/14/2021 Mehmet Nacar 20

Identity service issues security tokes for users Identity service support interfaces including: n n n Issuing identity/security tokens Authenticating security tokens for redirection Interfaces for administrative tasks including: w Adding, removing, editing users 9/14/2021 Mehmet Nacar 21

Federating multiple portals III 9/14/2021 Mehmet Nacar 22

Integrating portals with services • Mediator is a Web service • Administration of Web services through portal 9/14/2021 Mehmet Nacar 23

How WSRP fits into this architecture? • Handles Grid job management problems • WSRP provides Grid Web services 9/14/2021 Mehmet Nacar 24

Implications Provides standard way for portals to use external services Coordination of portal services is handled outside of portal framework More services can be added dynamically Service interfaces are extensible n New features can be added to services Enables lightweight portal frameworks (minimum local capabilities) Migration of portlets and services between portal frameworks will be easier Administration of Web services through portal 9/14/2021 Mehmet Nacar 25

Does this architecture scale well? There will be some overhead n Local capabilities vs. services In long term portal servers will be overloaded n n The architecture turns out to be advantageous Adaptable to for new features Simplifies migration across portal frameworks Services can be replicated n Identity service could run on several locations 9/14/2021 Mehmet Nacar 26

Milestones Implementation of identity, logging and authorization services using WS technologies. Design of mediator service n n n Mediator needs to coordinate services This will require the investigation of SOAP handlers and their proper pipelining. Managing SOAP header processors rather than services. Integration of reliable and secure WSRP requires two handlers n WS-RM reliability handler. n WS-Security handler 9/14/2021 Mehmet Nacar 27

Contributions of this thesis Simplification of portlet development for Grid applications through reusable libraries. n n Portlets are “coarse-grained” components. Tag libraries for portlets are “fine-grained” components. Design, development and validation of service-oriented portal architecture n Extends the current field w Limited by close coupling of portal core services to specific frameworks n n Our approach provides a natural way to federate multiple portal instances and to integrate portal services with other Grid services. We will examine the extremes of portal scalability in this approach. This framework will be a demanding test of Web Service standards. n Requires integration of several Web services standards w Integration of several standards: registration, security, reliability and others. n SOAP header processing is done by intermediaries w Intermediary coordination is similar to but distinct from Web Service workflow standards such as BPEL. 9/14/2021 Mehmet Nacar 28