PostgreSQL or MSSQL with ArcGIS for Server

[MartinSirkovsky1]

Hey.

This might be a quite general question, but I am wondering if there is any recommendation on which DBMS to use with ArcGIS for Server.

I am deciding on using PostgreSQL or MSSQL. If we do not think about the cost benefits of PostgreSQL, are there any performance or functionality advantages of using one over another?

Thank you

Martin

[KevinDunlop]

There is also a gap in help on examples to for doing back end development.  There is SQL Server and Oracle but nothing else.

Example: Resolving domain codes to description values using SQL—Help | ArcGIS Desktop 

As for which is better.  We use both SQL Server and Postgres in our office.  I find almost everything database management is easier with SQL Server.  In addition, the Always On options that SQL Server provides is much better and simpler than the Postgres high availability options.  

[KyleGonterwitz]

I love this question!  So what is best for GIS, SQL Server or Postgresql (or Oracle)?  

Depends.  It's almost a question akin to what is your favorite flavor?  I think the answer depends a lot on the nature of your agency, your IT support, your security needs, and to some degree database administrator bias.  

My DBA's have traditionally supported Oracle and SQL server, they have invested significant training and developed expertise in those platforms. Recently a project required them to deal with Postgresql - this does not make them happy, but it might open a door for a more serious discussion about this in the future.  

I use PostgreSQL on my desktop for development and local processing and I love it for that, I use PGIII for admin - but I wouldn't run my enterprise on it at this point because I would rather rely on my DBA team than try to go my own way.   

So here are some more of my opinions -  Postgresql with PostGIS has more native spatial functionality than MS SQL server, and rivals Oracle functionality.  If you want to do a lot of tabular and spatial queries and use advanced native SQL functions like linear referencing, then PostgreSQL and PostGIS is great.  VACUUM is a nice tool with PostgreSQL. I think more SQL spatial expertise exists in the Oracle community, and the transition for the analyst having Oracle Spatial SQL experience to PostgresSQL more empowers that rare developer than going to Microsoft SQL server.  The geodatabse containers in PosgreSQL are a little more like SQL server in a database as opposed to Oracle, which is at the schema.  Updating a geodatabse in Oracle is a chore compared to SQL server or Postgresql geodatabase updates, but you didn't ask about Oracle so I digress...anyway that is about the same between MSSQL and Posgresql so for me each of those options score a point against oracle there.

The advantage to me to using MSSQL in my agency is the support and expertise I get from IT involved in hosting, backing up, configuring, and maintaining databases in a supported environment - not to mention from Microsoft and Esri. From what I've seen Esri has a great relationship with the PostGRESQL developer community, some of the Esri GDB team are contributors the the FOSS project, postgresql is enterprise ready, even Gartner says so, and postgresql implementations are an increasing trend.  However unless you pay someone for support (like we do microsoft and oracle) who do you have to email or call and call you back right away when you have a support issue?  There are companies that I'm sure will provide that service for a cost, just something to think about that can be important to an enterprise.  

[JacobBoyle412]

All great points Kyle Gonterwitz! 

The long and short of it is there is no "right" answer in this realm, many factors should and will contribute to a decision of this size, like Kyle laid out above. 

The long and short of it, in my opinion, is use whatever database Esri supports that you and your support staff are most comfortable with.