Macromedia live cycle 7.2 Manual

Page of 123
     117
F
Enhancing Server Performance
This appendix provides general tips that you can use to improve server performance when using LiveCycle 
products.
Optimizing inline documents and impact on JVM memory
If you are typically processing documents of a relatively small size, you can improve the performance 
associated with the document transfer speed and storage space by implementing the following LiveCycle 
product configurations:
Increase the maximum inline size for LiveCycle products so that it is larger than the size of most 
documents.
For processing larger files, specify storage directories that are located on a high-speed disk system or a 
RAM disk.
The default maximum inline size and the storage directories (the Adobe LiveCycle products temporary file 
directory and the Global storage directory) are properties of the Data Manager Module. You can configure 
the Data Manager Module using Configuration Manager. (See 
.)
Note:
The default maximum inline size is 65536 bytes.
Document size and maximum inline size
When a document that is sent for processing by LiveCycle products is less than or equal to the maximum 
inline size, the document is stored on the server inline and the document is serialized as an Adobe 
Document object. Storing documents inline can have significant performance benefits. However, if you are 
using LiveCycle Workflow, the content may also be stored in the database for tracking purposes, and so 
increasing the maximum inline size might affect the database size.
A document that is larger than the maximum inline size is stored on the local file system (in the storage 
directories specified using Configuration Manager), and the Adobe Document object that is transferred to 
and from the server is only a pointer to that file. 
When document content is inlined (that is, less than the maximum inline size), the content is stored in the 
database (as part of the Document's serialization payload). So, increasing the maximum inline size might 
affect the database size.
JVM maximum heap size
An increase in the maximum inline size requires more memory for storing the serialized documents and so 
generally also requires an increase in the JVM maximum heap size. 
A heavily-loaded system that is processing a large number of documents can rapidly saturate the JVM 
heap memory. To avoid an OutOfMemoryError, the JVM maximum heap size must be increased by an 
amount corresponding to the size of the inline documents multiplied by the number of documents that 
are typically executed at any given time. 
JVM maximum heap size increase = (inline documents size) x (average number of documents processed)