Ok, I think I owe everyone in this thread an apology. I did leave one important fact out, albeit unintentionally. The company I work for handles call routing. This means they have to process up to 20 million records a day.
It occured to me to mention this now because of IT4EVR's benchmark link.
The company doesn't deal with websites, per se. However, from the look of the benchmark, I'm beginning to wonder if maybe Oracle isn't really the way to go.
Keep in mind that throughput for the server will increase significantly in the comming years, maybe even sooner.
Also, I need to find out if the home office isn't already using Oracle or if they are the ones contemplating it. This just occured to me from Walid's remark. The home office is responsible for handling the bulk of the record traffic. I'm betting it was them that put the bug in my bosses ear about Oracle.
However, I work at a satellite office. We provide certain services to the home office. Just because they switch to Oracle doesn't mean we have to also, other than just for the sake of having homogeneous data.
Overall, what I'm seeing in the Oracle vs Ms SQL debate is possible Performance vs Manageability.
However, it also looks like Ms SQL 2005 is making gains in the throughput arena.
Bottomline, as a satellite office, our main goal is to provide tools and services to the home office, i.e. rapid application development. My feel from the this thread is that Oracle doesn't really support rapid application development, at least not in the way Ms SQL does.
Yes, I do realize that .NET, (more specifically ADO.NET) makes different data sources a non-issue, so I could conceivably still develop a VB.NET application that pulls data from an Oracle DB as easily as an SQL DB.
However, I'm as of yet uncertain how much back-end processing we will be required to do.
As an application developer, I prefer to write back-end API's to do all the data processing. This is one of the reasons I'm so interested in OLAP Cubes. This would preclude me having to write some messy, complicated SQL Stored Proc that would return a set of values to the client app.
Also, many Off-the-Shelf applications already support OLAP, such as Excel's Pivot Tables and Office Pivot Table... not to mention Access.
Bottomline, however, goes to SQLSister. Yes, before any decision is made, a thorough cost-benefit Analysis must be done. I could attempt to do it, but I think I would be out of my league here. However, I could write a report of everything I would personally have to change in order to facilitate such a move... and, just from the information in this thread alone, it would be a very, very lengthy report.
- mongril
p.s.(Walid, please don't apologize. There wan't anything offensive in your remark. In fact, it provided valuable food for thought that I'm still contemplating. Thank you.)