Recently we have been talking about RESTful SOA principles. This topic is very interesting, so we are going to write about our experiments more in the future.

Who needs SOA? Why some people say that this is the only way to build good software for today's and tomorrow's needs? Why so many companies have been investing so much money to implement SOA solutions and SOA governance?

The answer is: this is because SOA should dramatically lower the costs of software development. Thomas Erl in one of his books talks about SOA from the business point of view. This is what I think after reading it:

• SOA-based solutions cost 1.5x more to develop at first
• Then SOA-based solutions cost 6x less to grow and reuse web services
• It is much cheaper to create new software using the collection of web services
• SOA-based IT infrastructure is more flexible

This is the theory of SOA and real numbers depend on situation, but they seem to be true.

As for everything else, harmony exists and SOA has some problems too:

• SOA-based software is slower and may require optimization workarounds
• The first stage of SOA-based software development is more expensive
• Organizations have to align themselves to SOA and implement SOA governance to avoid SOA failure

Idiles people experienced that too. One of the biggest problems in our SOA development was the speed of those web applications. We had to put more time and money to make optimizations and fix problems that do not exist in monolithic apps. When you have loosely-coupled web services, you cannot do joins and other operations on the database layer, so some logic is put into the code and it becomes much slower. Now we try to avoid this developing RESTful web apps that are small, well-structured and implements a set of logic for use in other web applications.

Practice gives you lessons and makes you better each time you fail. We hope this blog is will be helpful for you.