Oracle vs MSSQL vs PostgreSQL

[harmeetcheema]

I am building an application that will process hourly information for all time of the year. So, it will be storing data for about 1K objects each hour into the database. At each weekend it will deliver output to another application. It will be accessed by one user only to model static data.

MSSql is user friendly and cheaper but less expensive than Oracle but less scalable than Oracle. Based on the information provided above, could anyone provide resources or advice which database I should use: MSSql, Oracle, or PostgreSql. I tried to search for resources but I have not seen any complete and unbiased resource.

Thank you very much,
Harmeet

 

[Noway2]

This isn't a whole lot of information to go on, except that it will be inserting about 1000 objects in an hour. Putting it an other way, there are 3600 seconds in an hour, so you will be doing an insert statement about every 3.6 seconds.

It also sounds like you are comfortable with MySQL, which I personally have found to be a quite fine program. Do you have reason to believe that it won't be capable of keeping up with the requirements? Keep in mind that the SQL syntax used and whether you give it multiple value sets to work with at once will greatly impact the efficiency.

Unfortunately, I don't have direct experience with either PostgreSQL, or Oracle, to give you specific advice, but in my opinion all three of the ones mentioned are good tools. All of them will support the basic SQL syntax but you will have differences to consider regarding the extensions.

 

[bitcrafter]

This may be too late to do you any good, but here goes. You mentioned MS/SQL, Oracle, and PostgreSQL, but not MySQL. Your posting transaction rate can easily be handled by any of these, as can your once-a-week queries. MS/SQL may be the weakest player in the group, and ties you to Microsoft server platforms which can be a real pain. You did not mention which platform you preferred to run, and that might be a significant factor. Oracle is very good, but very expensive and difficult to manage. MySQL and Postgres are both free for non-commercial use, and both easily capable of doing what you describe.

Both have very extensive user communities for support. Postgres is the most "Oracle-like", using table spaces and such, while MySQL can use very simple one-file-per-table for storage in a single directory per schema. Does not sound like your total data volume will be that large, so either would work just fine.

MySQL does have one key advantage in our experience -- the excellent support for replication means you do not have to shut down the primary server to do a backup in most cases. Just replicate the primary to a slave (usually just a few seconds behind the primary), then backup from the slave (which is functioning as a hot backup itself) at whatever point you wish.

Cannot even count the number of folks I have spoken to who have sorely regretted going with a Microsoft solution (MS/SQL) when they got down the road a bit, and Oracle is just too heavy unless you are an enterprise-class user.