I'd like to discuss this to get a sense of how you guys interpret relationships between data. I'm particularly curious about the following relationships specifically (please forgive me if I'm misusing terms; I'm trying to keep the terms consistent with what I recall you mentioning), but the more schema the merrier here:
Realms -> Clients
Institutions -> Realms -- particularly in how I seem to recall this being at a yearly level, but I get the sense you have some clients with multiple "cycle runs" a year.
Applicants -> Realms
Applications -> Realms
Applications -> Applicants
Integration needs
This is specifically regarding client's security requirements of data transmission and storage compliance.
This diagram is a fairly straightforward load balanced web tier with a clustered database tier scenario. Since our usage expectations can be modeled and predicted there is little need for a distinct hardware load balancer resident between web and database tiers. I think we can generally trust the SQL Server controller node(s) to sufficiently route db traffic to the correct partitioned cluster. Since elasticity is not a primary concern, session coherence will be managed with sticky sesssions on the load balancers. The prevailing requirement on the web, database (and likely domain controller) servers is that it'll be running Windows. The mail relays may be Linux-based running sendmail or postfix.
All connections to public-facing infrastructure -- whether they be over HTTPS or LDAPS -- are restricted to Scholarship America public IP addresses. This can be achieved either at each constituent endpoint (web and backup domain controller) or directly via the VPN box. Our eapp environment cur