Welcome to Spunk!
It looks like you have servers in different network tiers which will all generate machine data. You will also need to decide if you want a standalone system or distributed system (This all depends on how many users are searching concurrently and how much data your indexing per day). Assuming your using a standalone system with the basic components you provided, you will need to get a beefy server with atleast 32-64GB of RAM and a good CPU, more cores the better as each real time search will use 1 core.
After setting up your standalone indexer, you will need to then install forwarders on your remote servers and point them at your indexer. You can also send data over the wire which will be good for the WAF, routers, LB's and firewalls.
So step one is for you to obtain a *nix based server with enough beef to run your indexer on. Once obtained said server, you need to install Splunk, configure firewall rules so data can send via 9997/514 to the indexer, create your indexes, then install the forwarders on your remote machines, then configure tcp/udp traffic to go to your indexer. Once this is complete, data will start flowing into your indexer in real time and Splunk will come alive!!
This will get you started on step 1
http://docs.splunk.com/Documentation/Splunk/6.5.1/Installation/Whatsinthismanual
