Hi All, As a person involved in both GRID and WEB Services community, I am keenly interested in the current discussions. I agree with Paul that the DII type functionality offered by WSIF if integrated with AXIS, will be EXTREMELY useful.
I will however like to point out that the client side invocation framework should take into account a number of broad issues. Here I will like to built upon Pauls example, of service oriented architecture as described in wsif for Apache.html - Pg.2 and try to illustrate in two generic problems that we face in the GRID community. 1.As pointed out by Paul, the Service Oriented Architecture has wider goals to address. Current implemntations of WEB SERVICES provide only Client / Server mode. The XML\SOAP RPC mechanism is suitable for such simple senarios, but fails to address the complex scenarios of composed services. A typical issue in composition is the propogation of state between the interacting services and the maintainance of the global application state. 2. The second issue we face is in trying to adapt the Web Services framework to a collaborative computing environments. I think that in addition if integrating WSIF with AXIS, we should also have a debate on the longer term issues of state and session management in Web Services. I will however like the readers of http://www.globus.org/research/papers/ogsa.pdf to read the following too http://www.extreme.indiana.edu/~gannon/OGSAanalysis3.pdf Regards, Vijay Dialani ----- Original Message ----- From: "Steve Graham" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Wednesday, April 24, 2002 1:09 PM Subject: Re: Proposed Contribution to Apache - WSIF > Indeed, Paul is correct here. The work we are doing in the Open Grid > Services Architecture [1] is quite looking forward to augment the Java WSIF > with other language implementations (eg C, C++). Many of our customers > have JVMs in their environment, many don't. Web services is at its heart > language neutral; we still have the hard engineering work ahead of us to > realize this. > > [1] http://www.globus.org/research/papers/ogsa.pdf > > ++++++++ > Steve Graham > [EMAIL PROTECTED] > (919)254-0615 (T/L 444) > Emerging Technologies > ++++++++ > > > > > "Paul Fremantle" > <[EMAIL PROTECTED] To: <[EMAIL PROTECTED]> > com> cc: > Subject: Re: Proposed Contribution to Apache - WSIF > 04/24/2002 05:38 > AM > Please respond to > axis-dev > > > > > > Steve > > I guess this isn't exactly your main point, but one aspect of WSIF we are > very interested in is producing a non-Java (C, C++, C*) implementation. In > particular, there has been some interest from Grid researchers in this. > > Paul > > > ----- Original Message ----- > From: "Steve Loughran" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Tuesday, April 23, 2002 9:52 PM > Subject: Re: Proposed Contribution to Apache - WSIF > > > > > > ----- Original Message ----- > > From: "Sam Ruby" <[EMAIL PROTECTED]> > > To: <[EMAIL PROTECTED]> > > Sent: Tuesday, April 23, 2002 5:54 AM > > Subject: Re: Proposed Contribution to Apache - WSIF > > > > > > > > > I'm thinking in terms of a separate PMC. > > > > > > These types of things always involve much more than anyone would > > reasonably > > > predict. The web site is the least of my concerns. The problem is > that > > > when you contemplate a change like this, people start revisiting and > > > questioning every assumption. Should Axis be the code base? Should > Axis > > > be split up? Should xml-security be included? How about xml-rpc? Are > > > there other code bases we should look to incorporate into the fold? > > > > Take a step back: should Java be the code base? It seems to me that C++, > > Perl and even C# should be in there as and when appropriate. I can easily > > imagine an assembly C# classes to provide interoperable implementations > of > > some of the more complex Java stuff, for example. > > > > What about REST based web services? There are indubitably things we can > do > > to improve that model of interaction; to help formalise it rather than > make > > it the generic 'call all post/get functionality a 'REST' service' > > marketology we have today. > > > > More importantly, what role can Apache play in actually setting a roadmap > > for web services? Currently it is implementing standards developed by > > mainstream bodies (soap1.2 & w3c), but there is a definate trend in the > web > > service world for the WS-I to announce standards (c.f last week' > > sauthentication announcement), under unclear IPR issues. > > > > I assume that the ibm.com folk (Sam included) are in-loop on the WS-I > stuff, > > but dont want Apache to end up at the bottom of the food chain for > > implementation of strategic stuff (i.e. be relegated to the client side > > library for commercial SOAP server products), and it is important for all > > the development community to be aware of what is going on. > > > > Apache could work with this group explicitly, but I have to start > worrying > > about any IPR issues; we dont want a repeat of the (hopefully now > resolved) > > Jakarta/JCP situation. > > > > > > > > What is the vision, mission, and goals of the new project? > > > > > > My intent at the moment is just to plan a seed so that it can possibly > > > germinate later. These are excellent things to pursue after an Axis > has > a > > > release. Meanwhile, I'd prefer that these considerations not derail > > > discussions as to whether or not there is general interest in WSIF as a > > > technology within Apache. > > > > -Steve > > > > > >
