1 universal worker service guy rixon to gws-wg at ivoa interoperability meeting, kyoto, may 2005

11
1 Universal Worker Universal Worker Service Service Guy Rixon to GWS-WG at Guy Rixon to GWS-WG at IVOA interoperability IVOA interoperability meeting, Kyoto, May 2005 meeting, Kyoto, May 2005

Upload: jasmine-king

Post on 27-Mar-2015

217 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

1

Universal Worker Universal Worker ServiceService

Guy Rixon to GWS-WG at IVOA Guy Rixon to GWS-WG at IVOA interoperability meeting, interoperability meeting,

Kyoto, May 2005Kyoto, May 2005

Page 2: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 2

DefinitionDefinition

UWS defines a WSDL contract and UWS defines a WSDL contract and service semantics for job-based service semantics for job-based services. ‘Job-based’ means that:services. ‘Job-based’ means that: Service executes work asynchronously, Service executes work asynchronously,

(‘in batch’)(‘in batch’) Service takes instructions in a Job-Service takes instructions in a Job-

Description Language (JDL) document, Description Language (JDL) document, not like an RPC.not like an RPC.

Service is stateful and provides a way to Service is stateful and provides a way to interact with individual jobs.interact with individual jobs.

Page 3: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 3

Why UWS?Why UWS?

More code reuse, because standard More code reuse, because standard WSDL contractWSDL contract

Good for workflows, because job-step Good for workflows, because job-step defined by a document in a job-defined by a document in a job-description languagedescription language

Good for long-running activities, Good for long-running activities, because job-steps run asynchronouslybecause job-steps run asynchronously

Good for connecting to grids, because Good for connecting to grids, because semantics are inherently grid-likesemantics are inherently grid-like

Page 4: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 4

UWS is abstracted from CEA v1UWS is abstracted from CEA v1

1.1. Job-description language (JDL)Job-description language (JDL)2.2. WSDL contractWSDL contract3.3. Standard semantics for job-based serviceStandard semantics for job-based service4.4. IVOA resource schema for job-based servicesIVOA resource schema for job-based services5.5. IVOA resource schema for applicationsIVOA resource schema for applications6.6. Application-description schemata (used in Application-description schemata (used in

JDL, resource schemata and local JDL, resource schemata and local configuration)configuration)UWS is the WSDL, the semantics and maybe UWS is the WSDL, the semantics and maybe the service-resource schema. Part of CEA v2, the service-resource schema. Part of CEA v2, maybemaybe. . It needs a JDL but need not use the It needs a JDL but need not use the one from CEA v1.one from CEA v1.

Page 5: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 5

Why change CEA v1?Why change CEA v1?

Easier to implement: refactored to Easier to implement: refactored to use WS-RFuse WS-RF

More features in WSDL contract, e.g. More features in WSDL contract, e.g. run-time estimation, run-time estimation, checkpointing/restartingcheckpointing/restarting

More flexible: spec’d to work with More flexible: spec’d to work with JDLs other than the CEA one.JDLs other than the CEA one.

Page 6: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 6

Why not use a grid standard?Why not use a grid standard?

Good question!Good question! There aren’t any, yet:There aren’t any, yet:

OGSA neither finished nor implemented. OGSA neither finished nor implemented. Others are Others are de factode facto, not , not de jurede jure and not multi- and not multi-

vendor vendor Grid stuff doesn’t cover all of CEA:Grid stuff doesn’t cover all of CEA:

Application registration.Application registration. Application description for UI.Application description for UI. Application libraries with mirror servers.Application libraries with mirror servers.

……and we’d like to go on using CEA, please. and we’d like to go on using CEA, please.

Page 7: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 7

UWS issues: shall we proceed?UWS issues: shall we proceed?

Do we want Do we want anyany standard for async. standard for async. Activities?Activities?

Would we rather bless CEA v1 Would we rather bless CEA v1 without mucking it around?without mucking it around?

Do we want some other way of doing Do we want some other way of doing this stuff?this stuff?

If we go ahead with UWS, who wants If we go ahead with UWS, who wants to work on the prototypes and spec?to work on the prototypes and spec?

Page 8: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 8

UWS issues: operation setUWS issues: operation set

Is the command set in the v0.1 spec. Is the command set in the v0.1 spec. sufficient?sufficient?

Is it sane?Is it sane? Is it too much? (Should we make Is it too much? (Should we make

more of it optional?)more of it optional?)

Page 9: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 9

UWS issues: pluggable JDLUWS issues: pluggable JDL

Is it sensible to have a WSDL that Is it sensible to have a WSDL that allows any JDL?allows any JDL?

Is it better to have one WSDL per Is it better to have one WSDL per JDL, but to keep to a common JDL, but to keep to a common pattern?pattern?

If the latter, should we refactor into a If the latter, should we refactor into a common port-type plus one port-type common port-type plus one port-type per JDL.per JDL.

Page 10: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 10

UWS issues: registrationUWS issues: registration

Is the CEA application/service split OK, or Is the CEA application/service split OK, or do we need a different arrangement? (Q.v. do we need a different arrangement? (Q.v. forthcoming data-model team in Registry-forthcoming data-model team in Registry-WG)WG)

How do we say “this service is a UWS”?How do we say “this service is a UWS”? Do we have a resource type for UWS?Do we have a resource type for UWS? Is this different from ceaApplicationType?Is this different from ceaApplicationType?

How do we say which JDL a UWS is using?How do we say which JDL a UWS is using? Do we have one resource type per JDL?Do we have one resource type per JDL? Q.v. one-WSDL-contract-per-JDL idea.Q.v. one-WSDL-contract-per-JDL idea.

Page 11: 1 Universal Worker Service Guy Rixon to GWS-WG at IVOA interoperability meeting, Kyoto, May 2005

IVOA interop meeting, Kyoto, May 2005 Universal Worker Service, slide 11

UWS issues: data gridUWS issues: data grid

CEA depends on AstroGrid MySpace.CEA depends on AstroGrid MySpace. Should UWS depend on the data grid?Should UWS depend on the data grid?

On the VOSpace layer?On the VOSpace layer? On the VOStore layer only?On the VOStore layer only?

NB: CEA v2 NB: CEA v2 mustmust handle the data grid. handle the data grid. Is data-grid handling pluggable along Is data-grid handling pluggable along

with the JDL?with the JDL?