Hello to my IT colleagues,
For sometime I wonder if ESB is good or bad. From SOA point of view is it a friend or a foe?
Well, I wrote about it and also designed some architecture that bypass the need for ESB.
I claimed that in the SOA world, the internet is my ESB and I do not need more that access to the desired web service (or REST service).
From my current position, I understand why ESB is popular.
It is simply comfortable for the IT organization. Normally, there is a specialized team that develop the integration requirements with the ESB.
SOA is nice but we keep getting all these different services, so the the integration work is still overloaded with XML transformations, connection strings, protocols and routing decisions.
So, despite the fact that ESB is a bottleneck in runtime and implementation time. Even if we try to adopt Jim Webber's ideas, we will be back to the ESB implementation.
I think it driven by the same reason we still carry electrical socket adapters when we travel abroad: The national standard for socket is already implemented, it will take a lot of investment to implement a new standard and who will define the new standard?
Try to ask your self the same questions about services in your IT environment and maybe someday you will not need ESB.
Please let me know what is the ESB usage in your organization.
Yours,
Noam
For sometime I wonder if ESB is good or bad. From SOA point of view is it a friend or a foe?
Well, I wrote about it and also designed some architecture that bypass the need for ESB.
I claimed that in the SOA world, the internet is my ESB and I do not need more that access to the desired web service (or REST service).
From my current position, I understand why ESB is popular.
It is simply comfortable for the IT organization. Normally, there is a specialized team that develop the integration requirements with the ESB.
SOA is nice but we keep getting all these different services, so the the integration work is still overloaded with XML transformations, connection strings, protocols and routing decisions.
So, despite the fact that ESB is a bottleneck in runtime and implementation time. Even if we try to adopt Jim Webber's ideas, we will be back to the ESB implementation.
I think it driven by the same reason we still carry electrical socket adapters when we travel abroad: The national standard for socket is already implemented, it will take a lot of investment to implement a new standard and who will define the new standard?
Try to ask your self the same questions about services in your IT environment and maybe someday you will not need ESB.
Please let me know what is the ESB usage in your organization.
Yours,
Noam
ESB is good only as SO pattern.
ESB product is good for very small organisations or start-ups that have no IT infrastructure. For mid- or large-size organisations with already existing IT infrastructure it is FOE.
ESB does not provide any additional or own business values over other IT technical means. Moreover, ESB as a product creates additional problems in the management of Service Contracts between service consumers and service (service provider).