Transform->Load vs Transform on the fly

[jeffs42885]

Hello al!

We currently use Xenos to transform PCL/MTA/AFP to PDF, and then store it OnDemand.

In talking with certain vendors other then xenos, they mentioned that alot of their clients store as AFP, and convert to PDF on the fly.

Can anyone provide feedback (positive/negatives) as do doing either of these methods

[Justin Derrick]

Historically, AFP compressed far better than PDF.  That may change in the newer versions of CMOD, which can 'decompose' PDFs (aka remove bulky resources like images and fonts), then 'recompose' them on retrieval -- in exactly the same way CMOD does for AFP data. 

The requisite for this is the PDF indexer, which is a separately licensed product or 'bolt-on-feature', which makes the cost of the separate software another variable...

Good luck!

-JD.

[jeffs42885]

Thanks JD. I'm just trying to figure out which is better..transform then load, or load then transform,

Does it make sense that transform then load as PDF would cut back on overhead from having to transform on the fly?

[Justin Derrick]

In order to know for sure, you'll have to try it yourself...  But from what I've seen, the cost of some cheap x86 hardware (or an extra CPU or two) to do some transforms is far less than the cost of twice as much enterprise-class disk running 24x7x365.

-JD.

[frasert]

My vote would be to store as AFP and convert to PDF on the retrievals.  Given that most reports are never retrieved anyway, why waste server resources converting to PDF?