Gervas Douglas (gmail) wrote: > Now we that we have this wonderful, flat SOA vision, is it perhaps not a > little too easy to look back and point out how badly things were done in > the past?
Hind sight is 20/20 right? > Despite the subsequent criticism it later copped (all too rife 5 years ago), > EAI was at least a serious attempt to break down the silo walls. Its > brittle rigidity, expense and other shortcomings were eventually > exposed, but what was the alternative at the time? Likewise with ESBs: > they may not be perfect and it is easy to chuck mud and other brown stuff > at them, but what else are you proposing in their place? Some of you, > like Anne, think they are a temporary phenomenon. Isn't all technology? Technology that doesn't add significant value is temporary. Look at the persistence of the C programming language. It's support, as a standard, on multiple platforms, with a language support portability layer (the ugly preprocessor that can be horribly abused). The portability has been important, but the simple memory management problems in internet applications is becomming a larger detractor, driving software systems toward GC, instead of explicit memory managment. The upswing of XML has largely been along the same lines with multiple vendor support enabling its use in development environments. It's not all that pretty, and its not as effecient as native wire protocols though. So we'll see if more effective solutions drive the industry towards other solutions. You know which is my choice for Java based applications. > OK, there is perhaps a general excitement about SOA because people feel > that they have finally stumbled across a paradigm that can provide a > strategic solution to such important issues as integration, process > agility etc. As this Group's arguments demonstrate SOA is still in an > early phase of development - by the time we get there it will probably > be largely superseded! The problem is, there is NO "silver bullet" that comes in a box. For any particular application, there will be customization, enhancements and management of the creation, performance and lifecycle of that application. There are certain types of applications that are best solved with particular technologies. Some feel that inter-business data exchange can be easily facilitated by such standardized languages as SQL, IDL and now the next IDL, XML. XML, is so malable, that it works better than the never completely standardized interfaces to SQL. And, because IDL/CORBA have been around for a while, data only exchanges with XML don't feel particularly bad. Being dependent on an APP server feels the same as being dependent on someone's ORB. We'll see how things evolve over the next 5 years. My prediction is that the speed and performance needs of systems will overwhelm the flexibility and portability that XML is perceived to provide. This will drive SOA systems to be simplified with native transports for many if not most parts of systems. Only the really distant, diverse extraprise communications will use XML. These XML interfaces will be laced together with huge amounts of software to reimplement all the type security and data quality that native language support provides to provide a certain QOS expected by the parties involved. The result of using the layered implementations of the XML based WS-* toolset will be that these outside interfaces will be an order of magnitude more running software than the language platform's native portion of the application. In turn, this is more to go wrong. The layers and layers of add on software that has a large scale layered implementation instead of an integrated native implementation will be a sizable impact on the software systems maintainability and performance. If the value is truely there, the systems architecture and cost might be appealing, tolerable and supportable. Only time and experience will tell... Gregg Wonderly ------------------------ Yahoo! Groups Sponsor --------------------~--> Most low income households are not online. Help bridge the digital divide today! http://us.click.yahoo.com/I258zB/QnQLAA/TtwFAA/NhFolB/TM --------------------------------------------------------------------~-> Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/service-orientated-architecture/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
