Microsoft Office Tutorials and References
In Depth Information
Having a good understanding of the types of workbooks that will be loaded by the ECS increases the
effectiveness of capacity planning. In most cases, settling for a pretty good understanding or even being
able to rule out certain types of workbooks will be about all you can hope for. Unless you have very
narrow use for the ECS, or already possess the majority of the workbooks to be used by the ECS, almost all
of the Excel features supported by the ECS can be used for workbook authoring, provided, of course,
that you configure the server to support them.
The following sections describe a few of the features that you should consider during capacity planning.
Here are a few areas to consider for external data:
The size of the data — The size of the result set returned by a data query impacts the ECS process
size as well as the CPU of the application server. As rows and columns in the result set expand,
the amount of memory used to cache the data increases. The ECS also consumes more CPU as the
result set grows, because it takes more processing time and power to work with the larger
number of rows and columns of data. Generating the HTML to send to the EWA is also more
expensive as the EWA grid becomes more populated with data. Of course, the time necessary to stream
the response to the EWA is closely tied to the size of the package.
Data cache sharing — The number of unique credentials used for the data queries impacts the
amount of memory devoted to caching the query results. The memory consumed specifically
for caching of data query results increases in relation to the number of different credentials
used for querying the available data sources. This is because there is less sharing of the query
results that are cached by the ECS. (Chapter 5 provided a detailed discussion of external data.)
The refresh frequency of the data source — Knowing the frequency at which data sources return
updated results for the data sources that will be trusted by the ECS allows you to optimize the
ECS External Data Cache Lifetime setting for maximum efficiency. If, for example, the external
data source will have new data every 60 minutes, you should configure the External Data Cache
Lifetime to use ECS cached query results for approximately 60 minutes. Allowing the cache to
be updated more often isn’t going to provide the workbooks with any newer data, yet the ECS
is expending resources refreshing the data cache. The more often the results can be retrieved
directly from the ECS data cache, the better the performance will be.
The data cache timeout frequency — Each workbook author can set a refresh rate for the workbook,
which could have the workbook configured to refresh the data when the workbook is opened by
the ECS, or to refresh at a set frequency after it is opened. The ECS External Data Cache Lifetime
setting determines when the data source gets queried.
Evaluate workbooks for their calculation intensity. Consider the following during your capacity planning:
Calculating workbooks on the ECS can be a CPU-intensive operation. Pay special attention to
workbooks that have long calculation times. This will help you account for the increased time at
potentially high CPU when these workbooks are calculated. The ECS can continue to
successfully respond to other requests while the calculation operations are in process, but throughput
will most likely degrade during those periods.