Data source retrieval patterns
148 Perceive Getting Started
Data source retrieval patterns
In previous versions of the product, you might not see a full day of historical data
from a Visualizer data source because of Visualizer population patterns and Perceive
data retrieval pattern. For example, a Visualizer data source can contain nodes that
are populated at different times, the result being that the nodes populated earlier
might not have data for the most recent intervals. In that case, Perceive does not
retrieve data from an alternative data source with data for that node.
Now, Perceive retrieves data for the specified interval, rather than the interval of the
whole data source.
Placement of data sources
BMC Software recommends the following guidelines for placement of data sources to
achieve the best performance with Perceive.
Oracle and SQL Server
BMC Software recommends placing Oracle or SQL Server data sources on separate
servers, properly configured and tuned.
UDR data sources
BMC Software recommends using NFS, mapped disks, or similar methods of
connection to remote UDR data sources. Locally placed UDR data can also be used,
but you might experience occasional erratic performance. This will be addressed in
future releases.
Scaling data sources
Perceive functionality is predicated upon the assumption that charts are typically
generated through a request or a series of sequential requests from one user at a time,
rather than from a group of users submitting requests simultaneously. BMC Software
welcomes any feedback on any ways you might use chart functionality differently.
By following the recommendations in this section, you can expect good performance
for data sources with up to 50 computers and 30 days of data. For larger data sources,
performance deterioration might occur.
