We have a smart client application and historically have always used the local dataportal (two tier). We have some potential engagements with large customers coming up. Are there any specific best practices or guidance on how to do a proper needs analysis to determine the best physical deployment scenario regarding two vs. three tier? I recognize all of the variables that could be considered, not the least of which being the nature of our application itself, the resources of the hardware on both client/server side, network infrastructure. With those considerations in mind, this task does seem difficult because we do not have all of these answers yet. What I'm looking for is guidance or a starting point on going about such an analysis.
Thanks,
Greg
On methodology aspect roughly know the ATAM Scenario-Base Software Architecture Evaluation Methods. Actually, it's a method to verify whether a choosen architecture fits the requirements. In your case you would define scenarios, which are different with large customer, and then "prove" on paper if a target architecture is adequate. The good thing about this, is that you describe non-functional requirements is less ambiguous, than terms like "scalable", "more flexible".
On the technical side, here are some ideas:
HTH
Copyright (c) Marimer LLC