Since the earliest days of our history, we have had a very close relationship with time. Religion, philosophy, astrology, causality, agriculture, navigation, transportation, communication, our legal system, humor — these are just a few areas where time plays a significant role, and there is more information about this later.
Computers and a wide variety of other devices need to know what time it is, too. The trick is that computers aren’t intelligent, and in general they keep pretty bad time:
The hardware clock is updated once per second and cannot display fractions of a second. Its timing uncertainty is determined by the quality of the crystal oscillator it uses as its time base. These crystals cost less than $1 in single quantities and offer only marginal timekeeping performance. They are sensitive to temperature and other factors, and their frequency uncertainty is not likely to be better than 1 x 10-5 (about 1 second per day). In actual operation, most hardware clocks gain or lose about 5 to 15 seconds per day, with 10 seconds per day being typical. Although the hardware clock usually outperforms the software clock, its performance pales in comparison to even a low-cost wristwatch. See this NIST document for some background.
Professor David L. Mills has been working on this problem for over 30 years’ time, and the software he and our group of volunteers have produced does a remarkably good job at keeping clocks synchronized. The software and the “rules” it follows are known as the Network Time Protocol (NTP). I’m Harlan Stenn, and I started working with NTP around 1990. By 1996 I had become “The Janitor of NTP”. While my title has changed a few times since then, Network Time remains a passion of mine. Along the way some new approaches to keeping accurate time on computers have presented themselves and it became obvious to me and others that NTP needed to grow and evolve to work with these new approaches.
Having correct time on most computers is essential. Time is used for many different types of security and audit-trail purposes and also for access control and proper “sequencing” of events.
A few years ago we realized that our workload was ever-increasing and that it was outstripping our ability to stay on top of things as a simple ad-hoc group of volunteers. In an attempt to develop a funding source for Network Time I started the NTP Forum at Internet Systems Consortium (ISC). But NTP was an “unfunded mandate” at ISC and ISC had other priorities. It became clear that Network Time needed its own vehicle to move forward.
By June of 2011 I had created Network Time Foundation, with the specific purpose to provide direct services and support to improve the state of accurate computer network timekeeping in the general community. In 2014 we were approved for 501(c) (3) Non-Profit status with the US Government.
Network Time Foundation is all about advancing and supporting these efforts. The scope of this work ranges from helping to craft the precise description of how to exchange time (standardizing protocols and behaviors), writing high-quality software and documentation, understanding and accommodating how things behave in a wide variety of real-world situations, and helping people and organizations trust that they have reliable and accurate timestamps. We do our level best to make sure that Network Time is running, ready, and available to everyone who needs it.