Two months ago, I bought this book simply because I could not find another book on application server. At first glance, it was a bit superficial and fell short of my expectation since I wanted to find a book on a detailed explanation of all currently available application server technologies. On further reading, I realised that the author wrote the subject from a perspective that I found both refreshing and useful.
As the name of the book suggests, the author provides you a fair treatment to the under-the-hood knowledge of BUILDING an application server. He gives a high-level description of the considerations that you have to be aware of if you are gonna construct one of your own.
Let draw an analogy. Of course no one would attempt to build their own DBMS nowadays. However, a fair understanding of the internal workings of DBMS would certainly help you quite a lot in development or tuning.
For example, armed with this knowlege, you are in a better position to win your technical manager over to the purchase of an object-relation mapping tool as the intricacies and difficulties of constructing the persistence layer in-house certainly cannot make a business case.
As the application server market is still at its inception, many companies are still at the "proof of concept" state of their corporate technology adoption process. This handy reference clears the mystique in that it provides a vendor-neutral description of what functionality a decent application server should possess. Say, I have never come across a book mentioning time service apart from the "core servies" like concurrency, security, persistence and so on.
On top of this, the author shares with you the hard-won design experience with application server in context.
If you think that the author treated the subject in too low a level, he really did to an extent. And this is exactly why I think it is thought-provoking.