Does NetIQ have standards or performance benchmarks for how long a migration should take?
NetIQ does not have any recommended time-frame to complete a migration of any size.
The time used for migrations depends largely on the network setup and configuration of the client site. That is, the computer involved, the total number of user and group accounts, bandwidth restrictions, time of day for the migration, options selected for the migration, etc.
In short, there are no fast rules on how long a migration should take. However, some examples could be raised for each user account taking a few seconds under ideal network connections such as:
- 100 Mb (or faster) LAN connectivity with low usage levels
- After hour migrations
- No roaming profiles or reACLing
One client did a test run on a 28 GB file server. First the client ran 'Robocopy' and it took almost 30 hours to move the data, then the client did a run with NetIQ's Server Consolidator and it only took two and a half hours to move the same 28 GBs of data.
Performance will vary based on server disk hardware, processor, RAM, number of entries in each file security descriptor. Performance can typically be measured in the number of security descriptors processed per second. With DMA, we?ve seen numbers between 200-400 file security descriptors processed per second. DMA also re-ACL?s file auditing and ownership settings.
Product Management has stated that account migration can take anywhere from 0.5 to 3 seconds per account. The estimated variance in these times is based on the differences between hardware and network components in each customer's environment. Unfortunately, we are not able to test our product's response rate in every customer's specific environment. Therefore, we can only be as specific as 0.5 to 3 seconds per account.
Overall, our suggestion is to encourage you to test migrations with your specific network environment's configurations, hardware, LAN links, migrations selections within DMA's GUI, etc. to discover exactly how long the migrations will take in your environment. As always, it's a recommended 'Best Practice' to test this in a Lab environment rather than in a Production environment. Using 'Test' mode will result in nearly identical processing timeframes.