[TYPO3-dev] The extbase dilemma

[Bastian Waidelich]

Mathias Schreiber [wmdb] wrote:

Hi Mattes,
Excuse the late reply, I had to digest this thread first ;)

>> I don't agree, a lot of people (including me) are successfully using
>> Extbase in productive projects. Some even with huge data sets [...]

> [...] Could you be a bit more precise on the term "huge"?

You got me there. I wrote "large" first and changed it to "huge" for 
some reason..
I can't serve you with 300k+ record projects I'm afraid, but we got some 
interesting use cases anyways. For instance:

http://www.busynoggin.com/portfolio/tri-kes/
"Only" 12k+ products but with ERP synchronization and location based 
AJAX searches

And: http://www.networkteam.com/blog/post/rossmannde-relaunch.html (german)
~5k products in 1000 categories with 2500 properties each.

There are a few more, that I don't have the numbers for (Apparently some 
clients don't like to make those public).

Anyways, the perception i got from talking to the implementers is:
Yes, you will probably have performance and memory issues if you use 
Extbase with a very large dataset and can't rely on the TYPO3 caching 
framework for some reason. But it was quite possible to work around this 
restrictions with a few tweaks and tricks for them.
We'll now focus on collecting and documenting those tricks and - 
preferable - of getting those into the core of Extbase & Fluid.

>> [...] there are certainly parts that are not (e.g. FE editing is problematic
>> with lots of (complex) records).

> Is there a wiki that names these?

Not yet, but it's on the agenda for the meeting today

> How much % increase in RAM usage would you say is ok compared to
> using piBase

I think, that really depends on the use case. But in any case, it would 
be really usefulto have some *objective* and *comparable* test cases so 
we can at least make performance measurements publicly available.

Best
Bastian