The first question I'd want to answer is what kind of load will your client's DMS see?
Storage capacity will be easy to manage by allocating the right number of drives, but CPU/RAM capacity will all depend on how many read/write queries per second you have to manage. IMHO the question is impossible to answer until you have those usage metrics. That's why a cloud based solution with vertical and horizontal scalability can be a winner in such a situation.
I personally currently have a production ~1TB MongoDB running on a 4 node replica set made of EC2 Large instances, 7.5GB RAM, 4 computing cores. Easy to scale those vertically by allocating XLarge or XXLarge instances, and easy to scale horizontally by adding extra read-replica-nodes or sharding.
On Thursday, October 4, 2012 6:32:35 AM UTC-6, kensuser wrote:
Hi,We have developed Document Management System to manage documents using mongodb as database.
one of of our potential customer has approximately 2TB(2000 GB) of data(documents, mainly pdf's and scanned images) that should be managed by our application.
My question is, to use this data, what should be the best Hardware specifications(in terms of CPU core, RAM, Storage Disk and other things) that we should recommend for that customer so that we can get good performance.
Any help/links in this regard is appreciated.
Thanks in advance,
kensuser