Scaling TEXTML Server
TEXTML Server can be scaled by increasing the resources on the machine or replicating the docbase.
Scale up
- Additional users: Increases the number of transactions, especially for heavy users.
- Increasing number of documents: Increases the size of the docbase and requires additional TEXTML Server indexing.
- New languages: Increases the size of the docbase and requires additional TEXTML Server indexing.
- Increasing number of Publish/Localize operations: Increases the size of the docbase and requires additional TEXTML Server indexing.
In all these cases, adding more RAM will increase performance. If memory is so large as to contain the whole docbase, more CPUs can be added to increase performance.
The disk space must also be large enough to hold the docbase.
Scale out
TEXTML Server can be partially scaled out by adding one or more replicated docbases, which can be used to improve read access speed for remote users.
In this solution, a replicated docbase is installed at secondary sites. Lite users at these sites access the replicated docbase to perform their read transactions (such as performing a search), so execution time is much faster than having to access the docbase at the primary site. Write operations (writing documents, locking, releasing, etc.) still go through the primary docbase, so execution time for these transactions remains the same.
Note that the performance of write operations for users at the secondary sites is slightly slower than for users at the primary site.