IMS tries to do 1 ping per second per device (spread across 4 worker threads- configurable; when a statistic for a device is gathered (every 8 seconds) the last 8 seconds of ping data is aggregated and placed against it in the form of connectivity and latency values; here are some example raw pings and their aggregated results:

• 8 failed pings
  • connectivity = 0%
  • latency = null
• 4 successful pings (25ms, 50ms, 50ms, 25ms), 4 failed pings
  • connectivity = 50%
  • latency = 37.5ms (average of the successful latencies)
• 6 successful pings (25ms, 25ms, 25ms, 200ms, 25ms, 25ms), 2 failed pings
  • connectivity = 75%
  • latency = 54.166ms