• Run DNS process on server < 20% CPU
• Maximize cache hit ratio by managing the memory cache value
• Configure recursive Contexts to be at 10-15% of total available RC's during sustained operation
• Distribute as much as possible
• Keep servers as close as possible to subscribers
• Use multiple operating systems and hardware types if your operations and deployment methods allow.
  • This is usually hard and could become cost prohibitive
    • Resources that understand multiple OS's
    • Multiple operating procedures for support
    • Varying deployment models
    • Expensive to operate
  • Performance metrics could vary based on the OS/hardware type

Monitoring: It is very important to keep track of every available system and software metric to keep your environment running at 99.999% uptime.

• CPU utilization
• MEM utilization
• DISK utilization
• I/O Subsystem stats
• Interface statistics
• Caching Server Process
• Recursive context statistics
• Queries per second
• Top clients devices generating DNS queries
• Top domains accessed

Patches and upgrades:

• Validate patches/upgrades in a lab environment
• Initially deploy patches/upgrades to a single server or a site and run for a period your operations team are comfortable with
• Update any methods and procedures (M&Ps) for your operations team if new features and or functions have been added.