I am looking for someone who can assist in the planning phases of a HPC build-out. I am planning on hosting a suite of web based tools that will be accessible by authenticated clients who will have their own data space tied to their account. Basically, I need assistance in identifying best business practices for hosting a cloud and analyzing different users data while not contaminating it from another client's data.
I'm assuming basic accounts will start with 5TB of disk space and will scale based on need. How many nodes etc will I need to support X number of clients as well as load balancing, front end servers etc.
My end goal would be to have a solid diagram of the HPC configuration as well as an idea of expected costs associated with various pricing/contract options available in AWS.
just to recap with some bullets...
- Overall architecture and design on paper of system outlining components, functions etc.
- Best storage solution for multiple users that's scalable
- Ability to ingest bulk upload and streaming data
- Method to store raw content and metadata output from analytics by customer
- UI servers which access stored customer data as well as HPC output
- Must be able to authenticate users and keep data separate
- Utilize machine learning which can enhance analytic across all data sets
- My associated costs by client to host such a solution. On-Demand vs. contract offerings