The South East Asia SOA Weblog

Some would have us believe that consolidation is good for the IT industry. I find that kind of hard to stomach as it really means less choice for consumers at the end of the day. Interestingly there has been very little backlash over Oracle’s intent to purchase Sun. There is so much at stake with this acquisition for the IT industry and business not just Open Source. Sun currently controls several key technologies and Open Source products one of which is MySQL but others such as OpenSSO are also valuable contributions to the community. There is quite a long list… foremost on that list is Java. The industry needs a trustworthy custodian for Java. I really hope that we are getting that with Oracle (assuming the deal will be approved).I was amazed at how fast the US approved the deal. There probably should have been a lot more scrutiny on this as it does represent a large shift in the power structure around Java and other technologies. Sun have been a benevolent and kind force driving Java to where it is today.
So what will be left of Java and MySQL when the Sun goes down?

In my day I have used many and varied data integration products and approaches. Obviously all of the solutions have various strengths and weaknesses. We have recently started using Talend Open Studio for data integration(www.talend.com). The depth and breadth of this suite is incredible for open source. Talend it brings a whole new dimension to the solution space and rivals any commercial product suite. As with most data integration tools, Talend provides the ability to map the process from a diagram perspective down to the specific mappings and various adapters to accomplish ETL tasks. The design of the tool has been well thought through and resulted in the most flexible and powerful data integration tool set i have used. Talend also has a great capability for building jobs and deploying them as packages that can be run as autonomous Java or PERL and/or deployed as Axis Web Services (can be consumed by any ESB of course) or into the jBoss ESB.

Here is a shot of the debugging panel

In short, there is way to much to describe here. Check it out for yourself or consult with us for advice in deployment.

This is just to let readers of the blog know that IASA (International Association of Software Architects) Indonesia Chapter now has a mailing list operational on Yahoo Groups. As per my earlier post there is also a Facebook Group (intended for the social side). IASA Indonesia would like to use the Yahoo Group iasa_indo@yahoogroups.com for all mailing list communications. All of you practicing architects in Indonesia please feel free to join our chapter and enjoy the benefits.

The solutions now offered in Open Source for implementing SOA have begun to provide extensive value but customers and systems integrators should be cautious and think strategically when assessing the value of these solutions. Projects such as WS02 and semi-open source (non-GPL, non-BSD, non-Apache) solutions such as Intalio BPMS are now comparable and even superior to many or even all their closed source counterparts. Considering some of the open source licenses offered by vendors it is necessary to take a cautious and well informed approach to open source offerings. Many of these solutions are contrived contributions that do not have a strong 3rd party developer community and are principally developed by one vendor with the dubious goal of proliferating their software and gaining the mind-share of the broader developer community. In my opinion the strategic focus of customers seeking open source solutions should be on those open source vendors licensing completely open and community driven software licenses along with support and subscription based licensing. In general customers should be cautions of vendors offering so called open source or community editions and a so called closed source enterprise edition. Watch this space for updates.

Just letting any readers of the blog locally in Indonesia know that we have just set up a Facebook group for IASA (International Association of Software Architects) in Indonesia. The group name is “IASA - Indonesia Chapter” and it is a closed group for professional practicing IT architects in Indonesia. The IASA Indonesia web site will also be revised soon but presently can be found here http://www.iasahome.org/web/indonesia. If readers of the blog are interested in IASA membership then please contact either david.forden@iasahome.org or henry.chandra@iasahome.org. We are both practicing professional IT architects involved in the IASA.

Noticed the rampant increase in the quality and enterprise penetration of open source software in recent years? Interestingly there are now many open source solutions that have evolved into worthy and even superior adversaries to enterprise software products. Facing reality in the level of support (at a 20+% premium of license cost) that most enterprise vendors offer, it is my experience that customers may as well consider open source alternatives where they make sense. What does this all have to do with Cloud Computing you might ask?

Well now that you have asked… in this industry it is always all about the next big thing and the next big thing is usually always about vendors grabbing markets and money from each other and/or the community at large (with two notable exceptions… GPL and LGPL). The powers that be in vendor land must be getting more than just a little concerned about the manner in which open source has started to transform enterprise computing. A response is needed and it is needed now!

What better means to bring it all back home than to bring the development platform back home component by component. Back to the oldest strategy in the IT vendor play book. Once again the pendulum swings. Let’s just call it a cloud and host it at our place. One question for you is whether those vendors have earned the level of trust that SAAS and ultimately cloud computing demands?

One of the big questions on the enterprise software vendor agenda a few years back was how to embrace and leverage open open standards. That same question was closely followed by the …open source version. Open standards (and more recently open source) have simply been embraced by vendors as a response to the backlash of the customer community against proprietary technology. This is one major similarity in the evolution of open standards and open source… by the community for the community. Cloud computing is dissimilar.

Responses from the vendor community to open standards and open source have ranged from the intentionally aggressive agenda toward open standards such as that exposed by the infamous Halloween memos ("embrace, extend, extinguish"), to softer approaches such as those leveraging the cathedral and the bazaar (thanks Eric http://www.catb.org/~esr/) which have manifested in the plethora of community and open source licenses with “enterprise” counterparts intended to increase the adoption of the product, foster its development by feeding the closed source tree from the open source tree and lock the developer community into the technology through skills investment.

Meanwhile amongst all of the vendor chaos real open source (IMHO GPL and LGPL) has evolved to take on major challenges and delivered beyond expectations. The questions one needs to ask when using enterprise open source are becoming somewhat analogous to the questions one should also ask when making a decision about using a word processor… how much of that functionality do you really ever use? Where is the break even point for justifying a commercial product over something that is freely available that has 80% of the commercial products features? The old 80/20 rule working for you again.

Interestingly, even open source SOA products are now maturing rapidly and delivering some amazingly feature packed solutions. One of the big hanging questions I now have is that once the soup-to-nuts comparisons begin to delve into features that only the rockets scientists can really appreciate then how does one justify and differentiate enterprise products over the open source counterparts?

I am sure that the vendor community can see this coming and under the guise of “SOA is falling into the too hard basket” they are being forced to dream up the next big thing. It is a bit like the relationship between suppliers, intermediaries and customers where the balance of power continuously swings back and forth. The inertia has been with open source, SOA and Web 2.0 for quite a time now so it is time for the inertia to shift.

Having worked for a handful of vendors in my time I feel I can faithfully predict where this is going. Once the pendulum has swung left then it is a matter of gravity and inertia to make it swing right again. In the IT industry that inertia seems to be building in shorter cycles… maybe boredom and the IT press has something to do with it. Cloud computing in the hands of vendors seems like a risky proposition to me at this juncture.

So in summary… cloud computing… beware the cloud… whose cloud? Go on… just plug into the cloud, it’s easy. For developers you get the advantage of having vendors centralize your resources, stifle your innovation and profit from it. For customers… you get the unmitigated privilege of being supported and controlled by those trustworthy vendors to whom you have been paying 20+% of license cost for that fantastic support over the last 20 years… now they are telling you to just just plug into the cloud.

Go ahead, make my day. They will figure out how to make you pay.

Unless you have had your eyes closed for 3 or 4 years you will probably have noticed a number of very negative articles written in the online press and blogs etc. regarding SOA being DOA (dead on arrival) along with a general bagging of the vendor community for selling the IT industry a lemon with SOA. A good example is a recent blog article by Anne Thomas Maines of Burton Group on January 5 2009 http://apsblog.burtongroup.com/2009/01/soa-is-dead-long-live-services.html. There are many others… just google “SOA DOA".

I find this mentality hard to rationalize given the potential benefits of SOA when done right and the fact that IT transformation needs to be a committed path for any organization and not just a short project. It is a shame that people are often openly negative about technologies so late in the hype cycle (…when SOA is approaching the plateau of productivity).

Many of these types of articles proclaim that SOA is dead on arrival due to the complexity of the underlying technologies and the high barriers to adoption etc.. Granted, there are some reasonably complex technologies underlying SOA but these technologies have been formulated in a layered manner so as to ease the burden on organizations and allow for prodigious reuse of the underlying infrastructure. They all have a very specific purpose. Gone are the simple day of client-server (which was hard enough for the community to digest at the time). Gone are the early days of the internet boom where HTML and Javascript were the IT industries bread and butter. We now have a vast array of layered point technologies at our disposal for developing unbelievable applications.

Many new and valuable technologies which have been loosely grouped under the banner of SOA are extremely valuable in their own right and solve many of the historical problems that IT has been confronted with in both enterprise and B2B integration along with implementing business process management.

One thing that many of the articles I have read fail to recognize is that “SOA” technologies are now embedded in most of the new enterprise applications offered by vendors and the open source community both simplifying integration of these applications and providing an extremely open and extensible enterprise IT platform. Now surely that is a significant benefit!!

In my 27 or so years of IT experience I have found that most new technologies have their place and some value for a time. It is just a matter of understanding that value at the right time in the cycle and putting it to work for you the right way. For example even CORBA is still at work for you (under the hood of J2EE as the transport mechanism for EJB. CORBA is one technology that many refer to as dead but it is the one sure fire way of integrating EJB transport across multiple vendor stacks (as opposed to trying to implement T3 bridges etc. to cope with disparate transport implementations).

Next thing we will see people in the community going negative on Web 2.0, RIA and Cloud Computing (which by the way has no fixed standards yet). If they think SOA was difficult then they should try some of these technologies on for size. New technologies need hard core architects and developers not hard core IT journalists to get successful…

I guess it is no secret that there is a global financial downturn presently and that many companies have slashed their IT budgets and projects. Having been in the industry for over 25 years, it never ceases to amaze me how companies miss the opportunity to upgrade and extend their processes and systems during the quiet times. In my opinion, business people in general have not yet fully adopted the concept of using IT as a competitive weapon. In traditional warfare either side will always use any break in the battle to fortify their position, improve their supply lines and prepare for the next battle.

Given the overarching goals of SOA in respect to enhancing competitive capabilities and agility it seems like the perfect reason to run SOA projects during the quiet time when the business is less distracted and can make more resources available to help the project deliver more value.