Home
Analytics
High volume letter generation
adwight
Hi,
I've been working on a project that is using BIRT to generate letters. Think mail merge in Word. This might be a bit of a square peg in a round hole, but we are happy so far with the ability to design the layout of a letter, bind this to an XML data source and emit a PDF for printing. The project also involved writing a "webform" emitter that generates a webform to collect data from a user to populate the associated letter. The report design is used to generate a webform, the webform is filled out and submitted, a simple servlet uses velocity to create the XML data source and the BIRT report engine takes the XML and report design to generate the PDF or PS file.
The same principle could be used with a JDBC data source to do bulk mailouts.
My question is if we use the BIRT reporting engine to generate letters in bulk, has anyone collected metrics on the volume of reports that can be generated in an hour? I know this is also a hardware question, but could BIRT scale to process 100,000+ designs per hour?
Cheers.
Find more posts tagged with
Comments
mwilliams
Meaning if you have a list of 100,000 that you want to create reports for with a generic letter that you're just changing the name and address and such? If so, it would depend on the amount of data that BIRT has to process and your hardware for sure. I think this would be a tough metric to put on paper since it would be very much based on the amount of data to be processed (size of report) and the speed of your hardware. Is the bulk of the text going to be in the report template? Or will it be brought in through the dataSet? This would also factor in. Hopefully this helps. I don't know if I have a GREAT answer for you. Sorry!
adwight
Yeah, hard to get an answer with so many variables to consider. The letters are between 1 and 3 pages long, mostly static text with name, address and some content based on the XML data source. I guess the best way to get some idea would be to set up a job to run a few thousand through, even if its the same data source and report design, and collect some metrics.
Thanks.