Friday, November 10, 2006

Globus GT 4.0.3 Not WSRF compliant!

Downloaded Globus GT4.0.3 to evaluate it for the NanoCMOS project. Unfortunately it uses really old versions of the WSRF and WS-Addressing specifications. This means that GT 4.0.3. is not WSRF compliant. It also means that services developed with GT 4.0.3 are not WSRF compliant either!

I asked when it would become compliant, but have got no response yet.

I cannot recommend that we use GT4 until we know what its timelines are. Once it does become compliant there is a strong chance older versions will not interoperate, making existing code and services obselete. Also many other WS-* tools are using the correct version of WS-Addressing, so they won't play nicely with GT4.

3 Comments:

Blogger Dave Berry said...

Mark,

Which other tools might we be using that we would need to interoperate with? As long as we're using a consistent set of tools, we should be OK.

I'm more concerned about the WS-A compliance than WSRF compliance, but this is from a very superficial understanding of the situation.

Dave.

8:52 AM  
Blogger Mark Mc Keown said...

Dave, my concern with GT4 not being WS-Addressing/WSRF compliant is that it will change in the future and we will not have backward compatibility with services/clients. If we knew the timetable we could evaluate and plan for the risks for the NanoCMOS project. I do not think GT4, old version of WSRF, will work with GT4 correct version of WSRF.

We could stick with current version of Globus throughout the course of the project but I think that would be unsatisfactory.

With regard to WS-Addressing I expect AXIS and .NET clients to use the proper correct version of the standard - making problems for interacting with GT4. For example Taverna does not support WS-Addressing at the minute, when they do add support they will only want to use the correct version of the standard.

If for example we knew that GT4 would be standards compliant in 6 months then the risks for NanoCMOS would be neglible. The relevant specifications have been standards for a long time now - I think going on a year.

Also you must be careful when talking about current GT4 services, you cannot say they are WSRF - that would be incorrect.

cheers
mark

8:05 AM  
Blogger Ian Foster said...

Mark,

You raise a few different issues:

1) The fact that different tooling support different versions of WS-Addressing is certainly an obstacle for interoperability. As far as I know, neither .NET nor Axis 1.2 support the final WS-Addressing spec. Neither does GT4.0.3, as you point out. I personally think that tooling should be more lenient in what versions of specs like WS-A they accept, but that is a different issue.

2) When will GT support the final WS-A, WSRF, and WS-N? We haven't been pushing on that, for a few reasons: a) people haven't been asking, b) other tooling is not ready, and c) we wanted to work out how to address backward compatibility. However, we recognize that it is time to set a date, and we will do so in the next few weeks. I will be surprised if it is not within the 6 months.

3) We didn't respond to your emails asking about this. My apologies! However, sticking it in your blog seemed to work (-:

Regards -- Ian.

11:13 AM  

Post a Comment

<< Home