Well, not just point to point solutions...
Highly efficient and highly scalable models and implementations of those models, such as publish/subscribe model on which for example Tibco technology is based, have been around since 80s. (Wall Street was first to use it, as it was born there)
How about MSQues, MQSeries, SAP, Corba, COM/DCOM and JMS?
At the end of that list are many proprietary and in house developed technologies as well.
There is nothing out there that will "dramatically simplify" integration problem, because the problem is not the integration layer in terms of a technology solution!!
The problem is not in the definition but rather in mapping, transformation and administration of business rules across the universe of "encapsulated and redundant legacy services"!!
True, there are many enabling tools in that arena, making that process easier from a point of view of definition and modelling, but all of them basically solve one dimensional problem, which is the service layer.
Building new applications on top of that service layer and using web services will be a brezee!! That is absolutely correct!
The scene behind the services and integration of services with legacy is the area that I am talking about.
That used to be a problem, is a problem and will remain the problem.
Hypes about new technologies solving old problems are cyclical trends that push its sale, but more often then not they end up short of fulfilling all of their promisses . Not that I am pesimist who is not willing to participate in new things, it is just the way things are.
All what I said, in no way should be taken as an opinion that undermines the advantages of internet and new enabling technologies. I hope I got my thoughts thru clearly.
Market Data 
Markets 
AI
