Header Only - DO NOT REMOVE - Extreme Networks

Brief Overview of NTP and SNTP Operation

  • 5 November 2013
  • 0 replies

Userlevel 3
Article ID: 5855


Related to

Brief explanation of NTP and SNTP operation
Difference between NTP and SNTP
Leap Second processing



All time servers are compatible with NTP (Network Time Protocol) and SNTP (Simple Network Time Protocol). The packets of information sent to and received from the time server do not differ between these two modes of operation.

Whether NTP or SNTP is used is determined by the time sync client software. On PCs running a Windows OS, the time sync client is 'W32Time'. Windows 2003 uses the NTP algorithms whereas Windows 2000, XP and NT all use the SNTP algorithms. On other networked client devices, consult the product documentation for operating mode expectation.

The effective difference between these two operating modes is in the algorithms used for time synchronization and error checking. NTP calculates the drift of the client's clock and then adjusts the drift rate so that the time is always correct. SNTP jumps the time to the correct value at each specified server polling interval. The NTP algorithm provides greater accuracy but is much more complicated to use.

Time Zones and Daylight Savings Time

In all cases the time server conveys timestamps, in a simple 64-bit format, that represent the number of seconds since 12 AM Greenwich Mean Time on January 1 1970. (GMT is now known as Coordinated Universal Time, abbreviated UTC.) It is the responsibility of the NTP/SNTP client to then not only compute the current date and time in the UTC (+0) Time Zone, but also to apply any adjustments relative to other Time Zones, Daylight Savings Time, or other such local concepts as desired.

If the client supports Time Zones but not Daylight Savings Time, emulating Daylight Savings Time may be accomplished by manually adding one hour to the Time Zone offset at the start of DST and then subtracting that hour back out from the Time Zone offset at the end of DST.

Leap Second processing

When the need for a Leap Second arises, the time server warns clients well in advance on the day the leap second is to be inserted, by including the approrpiate leap indicator bits in all data packets sent by the server on that day. It is then up to the clients to take note of these bits and accordingly insert the leap second on their own when their local clock reaches 23:59:59 of that day. The server would itself also insert the leap second into its own local clock. On 00:00:00 and after, the server stops including the leap indicator bits.

The NTP/SNTP client embedded in our network hardware ignores the leap indicator bits. So, after insertion of the leap second into the server's clock, the client's clock will lag by one second until it is resychronized with the time server. Other than that, use of leap seconds should not cause any problems.

For more about NTP, see RFC1305.
For information about SNTP, see RFC2030.

0 replies

Be the first to reply!