Our gear
We'll have 2 physical servers running VMWare (each with 2 VMs running Ubuntu), and HP's "Lights out Management" - if not for these non-hardened tools, I'd use a simple IP tables firewall. That's life though :/
Usage
We intend to serve about 6,000,000 requests per month (with a typical blog-like traffic over time distribution), about 60% being static images and text-based files (JavaScript, CSS, etc.). All of our text-based files will be GZipped by Nginx. About 25% of the traffic will be SSL traffic.
Things we'd like to have
- So, we'll need the ability to VPN into our network to access both servers on a network that is segregated from the network through which :80 and :443 traffic will pass (VPN in to access VMWare, Lights Out, and for SSH).
- Pretty Great Rate Limiting... OR - I'd like to be able to configure some nifty rules about when to block traffic and for how long, to prevent bad API usage by noob developers and other related things (e.g., I'd like to be able to use
IF X REQUESTS IN Y SECONDS: BLOCK FOR Z MINUTESor something to that effect). - I'd like to be able to set up the firewall to serve server A until server A stops working, then failover to server B.
A) Are there other considerations I should be making, and B) do you have any specific recommendations on product lines, models, etc?