You must limit the number of profiling queries to the number of CPU cores of the database or fewer than the number of cores. In relational data source profiles, the Profiling Service Module runs a profile on every column in a separate query to the relational source. The optimal run of enterprise discovery depends on the size of the relational source.

If you have lesser resources than the recommended guidelines, the time that the Profiling Service Module takes to finish the enterprise discovery job increases.

The following table indicates the recommended number of cores based on the number of database tables in a typical enterprise discovery job:

Number of Tables	CPU Cores	Memory	Concurrent Jobs
Less than 100	4	8 GB	3
Between 100 and 500	8	16 GB	5
Between 500 and 1000	16	32 GB	10
Between 1000 and 2000	32	64 GB	20
More than 2000	>=64	>=128 GB	40

You can estimate the Data Integration Service resources for enterprise discovery in two ways. The first approach is to decide the level of concurrency for the number of tables in the enterprise discovery job. Then, you can use the worksheets for the specific mix of profile jobs.

The second approach is to use the general recommendations in the following table:

The following table lists the recommended number of cores based on the number of database tables in a typical enterprise discovery job:

Number of Tables	CPU Cores	Memory	Temporary Disk / Spindles	Concurrent Jobs
Less than 200	4	8 GB	20 GB / 1	3
Between 200 and 1000	8	32 GB	80 GB / 2	5
Between 1000 and 2000	16	64 GB	160 GB / 3	10
More than 2000	>=32	>=128 GB	>= 320 GB / 4	20

When you plan for a deployment that has both single-table discovery and enterprise discovery jobs, choose the maximum configuration between the two use cases.