Retrieve Server Time in milliseconds

Guide for Server Time Synchronization (Using EWMA)

To ensure accurate synchronization with the server time, follow these steps:

1. Call the API endpoint /device/server-time multiple times and get server time in ms(epoch_ms). A higher number of requests improves accuracy but increases synchronization time. At least 10 times could be a good start.
2. For each request, note the time when the request is sent (sent_at) and when the response is received (received_at).
3. Calculate RTD (Round Trip Delay): RTD = received_at - sent_at.
4. Adjust for network delay by adding half of RTD to the received server time: server_time_estimated = epoch_ms + RTD / 2
5. Calculate the time difference between the server_time_estimated and the client’s current time (client_ts): time_diff = server_time_estimated - client_ts.
6. Instead of a simple average, use Exponentially Weighted Moving Average (EWMA) to smooth out variations:

  alpha = 0.3 # Smoothing factor, 0.1 to 0.3 (recommended for balancing stability and responsiveness).
  ewma = alpha * time_diff + (1 - alpha) * ewma

7. Use the EWMA-adjusted offset to compute the synchronized server time when sending data:

  synchronized_server_time_ms = client_ts + ewma

8. Use the synchronized server time to send data to the device in the corresponding API endpoints.