The first house I ever bought was built in 1936. It had style, it had character, and it had really narrow hallways and tight corners. The sofa we had bought - the one that went perfectly with all the style and character - wouldn't fit in the house. Apparently folks in 1936 had smaller furniture. Eventually I learned out how to take apart a window and was able to get the sofa into the house, but in the process, Pandora snuck out.
Behind the window frame was rot. I looked into some of the other windows and found rot and bugs. My initial reaction was that I'll just need to replace the windows, but (of course) it wasn't to be that easy. The windows, you see, had been built from scratch with the house. In fact, everything about the house was custom built. The parts of the house were all very tightly coupled with one another making replacement with componentized, functional equivalents too expensive. Though it had been well taken care of, the bottom line was that it was an old house and needed a lot of attention in order to maintain its primary function: being a living space. Ultimately we decided that the best alternative was to move into something newer that didn't need so much attention.
Many of my clients find that they're in a similar place with their software portfolio. As recently as ten years ago, they had to build systems infrastructure that wasn't then commercially available. They needed scalability, clustering, fault tolerance, high performance and integration to legacy systems and they needed to be able to add new business features at a break-neck pace. Design compromises were made, corners were cut, and the platform was not kept up-to-date. Though most have achieved great success, they now find their systems to be bloated, brittle, too highly coupled, and staggeringly expensive to maintain.
Most businesses don't have available the equivalent of buying a new house; they can't throw everything away and start from scratch. Instead, many are choosing the Herculean task of incremental renovation, of moving toward a service-oriented architecture in which their systems and their business process operate as an enterprise, not just a collection of departments. However as they turn over the rock of SOA, most are finding a dizzying array of new plumbing and buzzwords. The question is how to achieve the benefits of SOA without again becoming plumbers and in a way that is truly flexible. The answer is to focus on delivering custom business services in a standard way on a proven, commercial platform. Thats the way James Fenner sees it. It's all most the same way i see it in projects where old software has to be webalized (not to mismatch with webalizer). Beside the two possibilities , modernizing the system and build the complete system from scratch, there is a way i respectively the companys i work for have done it, we build a system alongside the existing and include the old system over wrappers, and bit by bit we implement the wrapped functions in the new systems. Over the period of a year or two, depending on the size of the old system, it is completley renewed and the old system could be abandoned.
Behind the window frame was rot. I looked into some of the other windows and found rot and bugs. My initial reaction was that I'll just need to replace the windows, but (of course) it wasn't to be that easy. The windows, you see, had been built from scratch with the house. In fact, everything about the house was custom built. The parts of the house were all very tightly coupled with one another making replacement with componentized, functional equivalents too expensive. Though it had been well taken care of, the bottom line was that it was an old house and needed a lot of attention in order to maintain its primary function: being a living space. Ultimately we decided that the best alternative was to move into something newer that didn't need so much attention.
Many of my clients find that they're in a similar place with their software portfolio. As recently as ten years ago, they had to build systems infrastructure that wasn't then commercially available. They needed scalability, clustering, fault tolerance, high performance and integration to legacy systems and they needed to be able to add new business features at a break-neck pace. Design compromises were made, corners were cut, and the platform was not kept up-to-date. Though most have achieved great success, they now find their systems to be bloated, brittle, too highly coupled, and staggeringly expensive to maintain.
Most businesses don't have available the equivalent of buying a new house; they can't throw everything away and start from scratch. Instead, many are choosing the Herculean task of incremental renovation, of moving toward a service-oriented architecture in which their systems and their business process operate as an enterprise, not just a collection of departments. However as they turn over the rock of SOA, most are finding a dizzying array of new plumbing and buzzwords. The question is how to achieve the benefits of SOA without again becoming plumbers and in a way that is truly flexible. The answer is to focus on delivering custom business services in a standard way on a proven, commercial platform. Thats the way James Fenner sees it. It's all most the same way i see it in projects where old software has to be webalized (not to mismatch with webalizer). Beside the two possibilities , modernizing the system and build the complete system from scratch, there is a way i respectively the companys i work for have done it, we build a system alongside the existing and include the old system over wrappers, and bit by bit we implement the wrapped functions in the new systems. Over the period of a year or two, depending on the size of the old system, it is completley renewed and the old system could be abandoned.