🚀 go-pugleaf

> I need to design a retrieval system for ~15TB of data, of which ~5TB are
> retrieved with a high frequency (25 requests/second) and an access time
> of avg. 30 seconds (max. 120 seconds). Retrieval lacks any locality, so the
> 5TB are real random-access.

Since our feed has been sucking mud for weeks now I'm unable to follow
any online discussion (if there is any). So far I have received several
very helpful comments per email. I'll summarize in a few days. Here's
answers to the most commonly-asked questions:
	1) Cost: We know it's expensive. Budget is ~$15 million.
	2) Access: Read-only, with few enough exceptions to handle
		them separately.
	3) Overall system load: ~4.5 MBytes per second, sustained
	4) File organization: Flat UNIX file system
	5) File size (~165 KB average)

I would appreciate it (if there is any online discussion about this topic)
if somebody would be willing to email the discussion to me. But please ask
me first, I don't want to drown in messages :-)

Thanks in advance,

---
Volker Herminghaus-Shirai (vhs@rhein-main.de)