Document resource requirements

[alanorth]

We are evaluating whether to use Gigwa2 and I noticed there is no mention of resource requirements or recommendations in the readme or website. Storage space obviously depends on your data, but is there a rule of thumb for the number of CPUs and memory? Thank you!

[GuilhemSempere]

This is difficult to assess since the system is really scalable and requirements indeed depend on the use made. I can give a few hints though (for production use):

• I'd recommend to setup Tomcat and MongoDB on different machines / VMs, and to allocate at least 8G of RAM to Tomcat
• The hardware running MongoDB must have fast disks (we're planning to change our current server and switch from SATA SSD to NVMe). Reading speed is more important than writing speed
• Choosing the amount of RAM for MongoDB really is the question that is not easy to answer. I would say this mostly depends on how big your largest dataset is going to be. Is this something you know at this point?

[alanorth]

@GuilhemSempere I appreciate your insight and totally understand that it's a tricky question. As for the datasets, we're probably going to be looking into at least the following genomes:

• Lablab purpureus: ~423 megabases
• Napier grass: ~2 gigabases
• Oats: ~12 gigabases

I would be happy to share my observations from working with these in Gigwa2 to form some kind of example for other users in the README.

[GuilhemSempere]

I was more asking in terms of #variants * #samples
I'll be glad to account for your feedback, thanks. Will you be using the cloud or a physical server on your own infrastructure? Cloud setup would be great to be able to test different configurations...