Hi all,

I'm pretty new to the LLBLGen, and I need to know the main differences between SelfServicing and Adapter - which one has better performance under what conditions? I couldn't find clear answer for my question in the documentation.....

orenpeled wrote:

Hi all,

I'm pretty new to the LLBLGen, and I need to know the main differences between SelfServicing and Adapter - which one has better performance under what conditions? I couldn't find clear answer for my question in the documentation.....

The core difference between them is in how you look at things. This is described in "Concepts - Template sets - When to use which template group?"

Performance is the same, sometimes one is a little bit faster (0.5% or so) than the other, but neither of them is slow compared the other.

Connection pooling, for example, is managed in the same way in both Adapter and SelfServicing? Can I control opening/closing/disposing connections in both Adapter and SelfServicing in same way?

orenpeled wrote:

Connection pooling, for example, is managed in the same way in both Adapter and SelfServicing? Can I control opening/closing/disposing connections in both Adapter and SelfServicing in same way?

Getting an open connection from the pool is a feature of the ado.net provider used, so that's the same for both.

Controlling if a connection is kept open or closed is not available on selfservicing but is on adapter (DataAccessAdapter.OpenConnection() / set KeepConnectionOpen to true (this can be done in teh constructor), -> do actions, -> close connection).

In this thread you've mentioned that performance is generally the same in both Adapter and SelfServicing After reading the concepts behind each one of them, I said to myself: hey, if DataAccessAdapter needs to "consult" another object in order to perform database-actions, it must be more expensive that SelfServicing, considering the fact that with SelfServicing each entity implements its own database-actions! Can you please clear this point for me?

orenpeled wrote:

In this thread you've mentioned that performance is generally the same in both Adapter and SelfServicing After reading the concepts behind each one of them, I said to myself: hey, if DataAccessAdapter needs to "consult" another object in order to perform database-actions, it must be more expensive that SelfServicing, considering the fact that with SelfServicing each entity implements its own database-actions! Can you please clear this point for me?

true, but that's just 1 call to a routine which picks it from a hashtable, or if it's not there, it creates the objects, which is very straightforward. Profiling the two for the same action doesn't reveal a big difference, most of the time they're on par.

OK, Thank you very much!