Implementing ArcSDE cost effectively using PostgreSQL?

Question

I wish to implement an ArcSDE environment without breaking the bank. I would be wanting SQL databases.

If I use PostgreSQL what are the limitations and or the requirements to make it work?

We have 15 concurrent ArcGIS for Desktop Advanced licenses with up to 20 different users.

• PolyGeo: Welcome to GIS SE! As a new user be sure to take the Tour where you will see that we request just one question be included in each question asked here. I recommend that you edit your question to focus it on whichever is the most important to you. I will be surprised if the limitations of ArcSDE for SQL Server Express have not been covered in a previous Q&A here.
• Vince: Please note that "SQL" is a language common to many database products. The Microsoft database product's name is "SQL-Server", with the least capable version named "SQL-Server Express". Fifteen concurrent users would exceed even a Workgroup license of ArcGIS Server by 50%. If you don’t want to pay for an enterprise class database, you'll need to consider PostgreSQL (though the most expensive part of a databse is the salary of the administrator, irrespective of database product).
• Reza: You need to use 'ArcSDE For SQLServer' version of ArcSDE.Before that you have to have SQL Server installed on your database server.We experienced ArcSDE 9.2 with SQLServer 2005 and 2012,without any problem.If you already have your database and need to use it as a geo-database,you need to register your layers with ArcSDE using sdelayer command(from command prompt) but if you want to construct a new database,simply use ArcCatalog and make a new connection to your sde database.

Answer 1

Postgres works fairly well as an ArcSDE GDB, but as the number of datasets increases, so does the connection time. This is not an issue with other enterprise databases.

For example, we have an Oracle SDE instance with 5000 datasets which takes about 5s to connect. We also have a postgres SDE instance containing the same data and it takes about a minute.

Esri are aware of the issue and are not prepared to fix it, so I would say only use postgres if you don't have many datasets.

• Radar: In terms of the connection time - is that just when you initially make the connection via ArcMap or ArcCatalog?
• jon_two: Initially. Once we're connected it's ok. Esri Inc have looked into this and basically ArcCatalog does a lot of querying the database to find out exactly what is in there. I believe Oracle, SQLServer and others generate tables to hold this information but PostGres does not.
• Farid Cheraghi: @Jon_Two, have you tested this behavior yourself? do have any test results? I use postgres with 500 dataset and it works perfectly.
• jon_two: @FaridCher I don't have any figures, but for small numbers of datasets like 500, connection time was ok. As you add more data though, the connection gets slower.