Recently in SOA Category

I mentioned in a previous post called TheServerSide Symposium Las Vegas that I was on a SOA expert panel at the show. Well I recieved the link to the video the other day and thought I'd better post it. Click on "Launch Presentation" on SOA Panel from TheServerSide Symposium. The transcript is also available on the launched panel through a tab.

I won't give you my own opiniion of the panel. You can judge for yourself. Feedback welcome and appreciated. If you have trouble viewing it please let me know too.

I've been visiting several telecommunications companies over the last couple of years about the difficulty in getting to the convergence nirvana. Recently I've been discussing the development of the service network with colleagues of mine, including Tony Parker and Brian Whittaker. One of the exciting developments is the idea of opening up their networks for third party applications offered as services. It's the idea of providing your network as a marketplace for Software as a Service (SaaS).


One of the overriding themes Tony, Brian and I have been hearing lately is the idea that while the technology is more or less available for the rapid deployment of services on the network, consumers of those services will have difficulty trusting these services and the network. And service providers are wary of the impact that consumers or other service providers can have on their services. In order for the network provider to be successful the consumers of services must be confident in the quality of the services on the network. These consumers include the produces of new services based on composites (mashups) of other services.

I attended TheServerSide Symposium in Las Vegas last week. I participated in an export panel on SOA. The panel was video taped and so I hope to post a link to that panel when it becomes available. The panel changed from advertised with Venkat Subramaniam (Agile Techologies), Neal Ford (ThoughtWorks), and Mark Richards (IBM) joining late when Eugene Ciurana (WalMart) dropped out.

We discussed the role of various technologies including ESBs and registries/repositories, we talked about SOA testing, best practices and use cases etc. I thought it went very well.

I was interested in Ross Manson's Mule & Spring talk. What I found most interesting is how Spring and Mule are maturing and therefore having to deal with enterprise class problems. It was like going back in time 10 years to an OMG CORBA event with people asking about high availability and other enterprise features. All middleware and container technologies go through the same cycle - cool "Hello World" type demo, add extra powerful features, hit the complexity of enterprise computing. The biggest challenge is trying to keep the "simple" technology simple as you layer on the extra functionality. Distributed SOA infrastructure manages to get complex as you try to continue to scale it. High availability wasn't something on offer with Mule ... yet.

IONA's Celtix Enterprise integrates with Mule and also contains Spring as a container for POJO type services. William Tam (IONA) showed me a really cool mash-up demo using Celtix with Spring and Mule. It involved a company tracking trucks through a city ... and of course it used the Google Maps API. I have it installed on my Mac and hope to play with it again soon. I hope someone can host this - perhaps I will. I'll let you know.

IONA and Microsoft

| No Comments | No TrackBacks

Recently I've been working on IONA's partnership with Microsoft. That has been keeping me very busy. Trips to Redmond, Nashville and Copenhagen for various events, and lots of other activities.

IONA and Microsoft have a history. IONA has always recognized the value of Microsoft in the IT organizations of the worlds most successful corporations. Back when people were talking about COM Vs CORBA in the 90s, IONA recognized the opportunity to bridge both technologies. IONA understood that Microsoft dominated the desktop in large enterprises (OS/2 was still around but declining). IONA built COMet as a very successful COM/CORBA bridging technology. In my opinion that product helped both companies considerably. Today, as both companies roll out more SOA technologies and products, it is important for us to partner.

IONA continues to help bridge Microsoft based applicaitons and services with non-Microsoft based applications and services today with Artix and Artix Connect. IONA has been working to help integrate with Microsoft products such as Connected Services Framework (CSF) and Customer Care Framework (CCF) - see our Microsoft partner page for more details.

Of course there is also a lot of cooperative activity with Microsoft in the various WS-* standards bodies and working groups. And IONA has participated in interoperability events and plugfests to help ensure that our products interoperate. As new products from IONA and Microsoft come to market, IONA is making sure that we continue to interoperate and integrate with Microsoft.

Many of you know me as a Mac guy. Well I just got a new Dell to help me focus on some of my Microsoft activities. More on that later.

The Gravity of Hubs

| No Comments | No TrackBacks

Why is it that though the industry pundits and experts describe SOA based solutions as more lightweight and distributed, that the vendors are continuing to push out hubs (large stacks, containers, etc.) to their customers? Why is it that customers are still buying hubs?

Is it because the so-called experts are wrong and the customers understand the problem better? Perhaps an academically sound approach just isn't needed in the real world and hubs will do just fine, thank you very much. (I believe there is a class of user this just can't scale too).

Is it because buyers are just plain uneducated or ignorant? I've come across many of these over the years. And I've come across system integrators that build out poor designs tha make distributed architectures look bad.

Is it because customers don't care? They're so burned out buying infrastructure that they've given up the fight? Let's face it they've gone through RPC, DCOM, CORBA, J2EE, .Net etc. and now they're faced with ESBs, JBI, SCA, WCF .....

Is it because hubs can make good sense for small to medium sized businesses and people can't differentiate between a small/medium sized problem and a big problem? The solutions must be different to scale and take care of issues such as transactions and other enterprise features. But, perhaps, it's just seems easier with a hub.

Is it because of perceived benefits of a hub like management? This is of course only a perception problem there are many technologies out there that allow for central management of a distributed architecture.

Hubs seem to have gravity and people get attracted to them. Unfortunately hubs can turn into black holes. Sucking in all that surrounds it into a dense mass from which nothing can escape.

Of course many customers are getting sick of hub based approaches and are not buying. They have enough infrastructure and plenty of hubs. Now they want to leverage that existing asset while a the same time taking advantage of SOA.

WS-Policy WG

| No Comments | No TrackBacks

I've had the privilege of working with the WS-Policy Working Group (WG) at W3C over the last few months. I know, it may seem hard to believe that working on a standards body working group is a privilege, and often it does seem like a chore, but there are several reasons for why I feel this way.

Firstly, I'm gaining new experience; experience of standards body processes. It's always a privilege to learn something new. And secondly, these are a smart bunch of people. At times some of the debate seems trivial but very smart people are putting their minds together in order develop some standards that will make Web services more interoperable with more advanced and rich features in the future.

Companies like Microsoft, BEA, Sun Microsystems, IONA, SAP, Sonic Software, Nokia, IBM, Nortel, Adobe, webMethods, etc. invest lots of resources to these standards bodies (WS-Policy Participants) Some of the people are in several working groups and basically have built a career just working on standards,. And it is certainly not a cushy number. These people work hard on some very tedious material! It can do you head in!

I am an infant in this world. Though I have lots of enterprise computing experience and interoperability experience I feel like a complete novice. I'm fortunate to have landed with a very civil bunch who are gracious at bringing me up to speed.

Now there are many times that this sort of working group activity will do my head in. Bickering over the semantics of a word or the usage of a word or the absence of a word is not how I'd like to spend my day. But I've come to appreciate what can happen when ideas and standards are ambiguous. Chaos can ensue and perfectly good initiatives can die.

I'm hoping to pull post an article giving an overview of WS-Policy. Stay posted.

This week the WG had a face-to-face in Bellevue, Washington. I finally got to meet the people I've been talking to on conference calls every week for the last few months. We got to find out a little more about each other - not just our views on WS-Policy. Bellevue/Seattle was beautiful when I arrived but turned ugly from Wednesday. It was wet like Ireland. We did have a wonderful meal at the Seastar restaurant. I'd recommend it.

Recently I have been involved in some projects that introduce SOA as a means to reduce testing efforts in large organizations. Taking two or three weeks out of the testing phase of a project and lowering the overall software development lifecycle (SDLC) can save millions of dollars for large organizations.

IONA's Professional Services organization has been researching and implementing methods and practices that can leverage the benefits of SOA for integration testing. Many of IONA's larger customers have a plethora of middleware and platforms. Having a consistent and automated approach to testing across these various technologies has been very difficult until now.

Leveraging best practices, testing tools and the unique capabilities of Artix, IONA PS has developed Certification Kits that allow for independent testing of integration end-points by disparate or remote groups. And this can be achieved no matter what the underlying middleware: CORBA, MQ Series, Tuxedo, Tibco, J2EE, Web services, CICS etc. This is achieved by harnessing the unique capabilities of Artix as an Extensible ESB. Artix employees WSDL as the common underlying interface definition or contract between the endpoints. WSDL is employed no matter what the underlying technology.

At one customer, IONA PS were able to take three weeks out of the testing phase. Consider how many lifecycles each application or team has per year and how many teams there are. You can then imagine how the savings add up. And then consider the time to market advantages.

I was a skeptic myself until recently. I saw an impressive demonstration of the capabilities performed by my old buddy Ashwin Karpe. Really great stuff!

Who is IPBabble

William Henry IP Babble is the personal blog of William Henry.

William has 20 years experience in software development and distributed computing and holds a M.Sc from Dublin City University. He is currently working in the office of CTO at Red Hat on the MRG product. This weblog is not funded by Red Hat.

Posts are intended to express independent points of view, but understand that there is probably a bias based on the influence of working with standards based middleware for over a decade. (See disclaimer below)

December 2008

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30 31      

Disclaimer

The views expressed in this blog are solely the personal views of the author and DO NOT represent the views of his employer or any third party.

About this Archive

This page is an archive of recent entries in the SOA category.

REST is the previous category.

Sun is the next category.

Find recent content on the main index or look in the archives to find all content.