Routing Policies

Simple routing

Untitled.png

  • No health check, all send to client for their choice

Failover routing

Untitled 1.png

  • if a health check pass, use the primary record as normal
  • if a health check fails, direct to secondary record for a failure page

Multi value routing

Untitled 2.png

  • return all healthy record and send to client for their choice

Weighted routing

Untitled 3.png

  • Specified the record weight, total weight is calculated (40+40+20=100)
  • the frequency of records returns as the specified weight
  • unhealth = skipped, but the weight won’t change
  • good for load balancing or testing new software versions

Latency-Based routing

Untitled 4.png

  • send the lowest latency to users
  • if one’s health check fail, move to 2nd lowest latency

Geolocation routing

Untitled 5.png

  • IP check verifies the location of the user and return the matched location
  • check state > country > continent > ‘no answer’
  • can be used for regional restrictions and geographical load balancing
  • NOT THE CLOSEST RECORDS

Geo-proximity routing

Untitled 6.png

  • you defined regional rules
  • direct the records to the closest recourse
  • introduce “+”/ “-” bias for certain counties that change the calculation, so more or less resources were routed to those countries