No single best practice can defend against determined attackers, and no combination of these steps is a magical elixir. While most products come with a solid set of general rules out of the box, getting the most out any tool requires that someone familiar with the behavior the application is protecting is in charge of building out custom rule sets.
Even then, Walton warns, "highly crafted targeted attacks might still be able to be effective before blocking rules and alerts are triggered."

• Linux & Network Tools:
  
• iptables/ufw - Firewalls
  
• Fail2ban - IPS/IDS
  
• SELinux - Securing Enterprise Linux
  
• Remove unauthorized users
  
• Login with security keys (keep in lock box)
  
• Train personnel and perform psychological evaluations
  
• Move away from IPv4 to IPv6 (enable IPSec AES256 ESP/AH VPN connection between two sites using MPLS rd#:# connections)
  
• Implement TRILL for R-Bridge configuration where IPv6 runs atop of it
  
• Implement NGFW where they connect to a ML site to process numerous attacks
  
• Reduce the size of your attack vector (email traffic goes through a proxy, and Web traffic goes through NGFW with embedded intelligence and process at line speed - ASICS)
  
• Create a baseline OS that is locked down, accepted by all
  
• Implement Kubernetes for web applications and lock down those applications using Carbon-Black/Blue-Vector, Sophos or any machine learning capability
  
• Monitor the systems using Extrahop (monitors performance, availability and application performance using an assortment of metrics)
  
• Provide training to all personnel, mentor personnel with senior/advanced security personnel
  
• Implement simulation tests with all personnel to help gather metrics of how the team and individuals are performing (depends on the level of comfort)