Status of This Memo This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.

The National Institute of Standards and Technology (NIST), through a collaboration with Perseus, a global provider of telecommunication services, is providing accurate, traceable, and verifiable time synchronization to stock exchanges in the United States, Europe, and Asia. The paper describes why accurate time is necessary for fair and equitable financial markets and summarizes current and proposed future synchronization requirements in the financial sector. We discuss reference time sources and provide a technical overview of how NIST transfers time to data center hosted stock exchange. We also discuss how Perseus distributes NIST time to financial market customers and describes how the time is verified. Measurement data are presented, along with a discussion of measurement uncertainty.

To help get a better handle on thread scheduling, we take a look at how FreeBSD 5.2 handles it. A busy system makes thousands of scheduling decisions per second, so the speed with which scheduling decisions are made is critical to the performance of the system as a whole. This article-excerpted from the forthcoming book, The Design and Implementation of the FreeBSD Operating System-uses the example of the open source FreeBSD system to help us understand thread scheduling. The original FreeBSD scheduler was designed in the 1980s for large uniprocessor systems. Although it continues to work well in that environment today, the new ULE scheduler was designed specifically to optimize multiprocessor and multithread environments. 1 This article first studies the original FreeBSD scheduler, then describes the new ULE scheduler. The article does not describe the realtime scheduler that is also available in FreeBSD.

O ne of the more surprising things about digital systems-and, in particular, modern computers-is how poorly they keep time. When most programs ran on a single system this was not a significant issue for the majority of software developers, but once software moved into the distributed-systems realm this inaccuracy became a significant challenge. Few programmers have read the most important paper in this area, Leslie Lamport's "Time, Clocks, and the Ordering of Events in a Distributed System" (1978), 2 and only a few more have come to appreciate the problems they face once they move into the world of distributed systems. Any discussion of time should center around two different measurements: synchronization and syntonization. Synchronization, loosely defined, is how close two different clocks are to each other at any particular instant. If two clocks, or computers, claim that it is 15:30 at exactly the same moment, then they are considered to be in sync. The definition of "exactly the same moment" is where the first difficulty arises. Do you care about the same minute (15:30), the same second (15:30:00), millisecond, microsecond, nanosecond, etc.? The level of accuracy you wish to achieve Time is an illusion. Lunchtime doubly so.

My work group has just been given approval to hire four new programmers, and now all of us have to interview people, both on the phone and in person. I hate interviewing people. I never know what to ask. I’ve also noticed that people tend to be careless with the truth when writing their resumes. We’re considering a programming test for our next round of interviewees, because we realized that some previous candidates clearly couldn’t program their way out of a paper bag. There have to be tricks to speeding up hiring without compromising whom we hire.