Which database is generally recognized as the big boy on the block for enterprises?

I have experience with various versions of SQL Server and Oracle and my general sense is that, deserving or not, Oracle probably has a better reputation for being the preferred database although I sense MS has been closing ground for some time and sometimes even claiming that it outperforms Oracle in situations x,y, and whatever, a close cousin to z.

A friend of mine who works for the gov't has told me t开发者_如何学运维hat "they can't use either of those databases because they aren't "robust enough" - or something to that effect and that they had to use IBM's DB2 database.

I'm would expect this to be a very difficult question to answer in such broad terms, but could someone just give me an idea as to which db products are generally regarded as being powerful enough (or whatever word you want to apply) for large size high volume enterprises?

If you want to throw in your perspective of how reality differs from general public perception, I'd be interested, too.

My gut tells me that either of these three products could probably be used o successfully implement the largest of enterprises if you know how to design and implement it, but again, I am looking for a little bit of what one product might have over the other vs. public perceptions, deserved or undeserved.

..and if you cancel this question, help me understand why subjective is a tag.:-)

---

Your gut is largely correct and a good argument can be made for any of the three. Keep in mind that what is used in any given Government, Corporate or other enterprise environment is largely a function of who got to the decision maker first or who the decision maker trusts. That is, the person setting policy may not be a domain expert at all but nonetheless has the power to dictate policy because of their position in the hierarchy. If a given government department will only use DB2, I'd suggest that you take that "evidence" that DB2 is the "best" or "most robust" database with a very large grain of salt.

---

I have worked many contracts and in the past 10 years I have seen Oracle and PostgreSQL more than anything.

IBM's DB2 is usually used in large enterprise data farms. For scale I have had Oracle systems handle Petabytes of data.

I have had to build systems specifically for database migration and upward scaling. I can tell you sometimes it is a matter of being locked in to a current Database System by application and operations. Meaning it is more cost effective to scale up the system/database in place than; say migrate your Oracle DBs to DB2.

---

The SQL engine is just one part of the equation, RAM, Disk IO subsystem (SAN), CPUs, fibre etc etc all have to be included.

All of the different engines have different things to offer like RAC, compression of data and backups, partitioning, index/materialized views etc etc etc

Design and proper indexing is also very important...a crappy design on any of these systems will be a crappy system in general

---

I think our company probably qualifies as one of the "big boys". We are one of the largest solar manufacturing companies in the world. We are exclusively a SQL Server 2005 and 2008 shop. Within the last year we evaluated Oracle as a possible replacement and rejected it. SQL Server manages everything from plant floor operations to ERP. We couldn't be happier with it.