David L. Mills

Internet time synchronization: the network time protocol

The network time protocol (NTP), which is designed to distribute time information in a large, diverse system, is described. It uses a symmetric architecture in which a distributed subnet of time servers operating in a self-organizing, hierarchical configuration synchronizes local clocks within the subnet and to national time standards via wire, radio, or calibrated atomic clock. The servers can also redistribute time information within a network via local routing algorithms and time daemons. The NTP synchronization system, which has been in regular operation in the Internet for the last several years, is described, along with performance data which show that timekeeping accuracy throughout most portions of the Internet can be ordinarily maintained to within a few milliseconds, even in cases of failure or disruption of clocks, time servers, or networks. >

Improved algorithms for synchronizing computer network clocks

The Network Time Protocol (NTP) is widely deployed in the Internet to synchronize computer clocks to each other and to international standards via telephone modem, radio and satellite. The protocols and algorithms have evolved over more than a decade to produce the present NTP Version 3 specification and implementations. Most of the estimated deployment of 100000 NTP servers and clients enjoy synchronization to within a few tens of milliseconds in the Internet of today. This paper describes specific improvements developed for NTP Version 3 which have resulted in increased accuracy, stability and reliability in both local-area and wide-area networks. These include engineered refinements of several algorithms used to measure time differences between a local clock and a number of peer clocks in the network, as well as to select the best subset from among an ensemble of peer clocks and combine their differences to produce a local dock accuracy better than any in the ensemble. This paper also describes engineered refinements of the algorithms used to adjust the time and frequency of the local clock, which functions as a disciplined oscillator. The refinements provide automatic adjustment of algorithm parameters in response to prevailing network conditions, in order to minimize network traffic between clients and busy servers while maintaining the best accuracy. Finally, this paper describes certain enhancements to the Unix operating system kernel software in order to realize submillisecond accuracies with fast workstations and networks. >

Precision synchronization of computer network clocks

This paper builds on previous work involving the Network Time Protocol, which is used to synchronize computer clocks in the Internet. It describes a series of incremental improvements in system hardware and software which result in significantly better accuracy and stability, especially in primary time servers directly synchronized to radio or satellite time services. These improvements include novel interfacing techniques and operating system features. The goal in this effort is to improve the synchronization accuracy for fast computers and networks from the tens of milliseconds regime of the present technology to the submillisecond regime of the future.In order to assess how well these improvements work, a series of experiments is described in which the error contributions of various modern Unix system hardware and software components are calibrated. These experiments define the accuracy and stability expectations of the computer clock and establish its design parameters with respect to time and frequency error tolerances. The paper concludes that submillisecond accuracies are indeed practical, but that further improvements will be possible only through the use of temperature-compensated local clock oscillators.

The DARPA internet protocol suite

THE MILITARY requirement for computer communications between heterogeneous computers on heterogeneous networks has driven the development of a standard suite of protocols to permit such communications to take place in a robust and flexible manner. These protocols support an architecture consisting of multiple packet switched networks interconnected by gateways. The DARPA experimental internet system consists of satellite, terrestrial, radio, and local networks, all interconnected through a system of gateways and a set of common protocols.

Adaptive hybrid clock discipline algorithm for the network time protocol

This paper describes the analysis, implementation, and performance of a new algorithm engineered to discipline a computer clock to a source of standard time, such as a GPS receiver or another computer synchronized to such a source. The algorithm is intended for the network time protocol (NTP), which is in widespread use to synchronize computer clocks in the global Internet, or with another functionally equivalent protocol such as DTSS or PCS. It controls the computer clock time and frequency using an adaptive-parameter hybrid phase/frequency lock feedback loop. Compared with the current NTP Version 3 algorithm, the new algorithm developed for NTP Version 4 provides improved accuracy and reduced network overhead, especially when per-packet or per-call charges are involved. The algorithm has been implemented in a special-purpose NTP simulator, which also includes the entire suite of NTP algorithms. The performance has been verified using this simulator and both synthetic data and real data from Internet time servers in Europe, Asia, and the Americas.

On the Accuracy and Stablility of Clocks Synchronized by the Network Time Protocol in the Internet System

This paper describes a series of experiments involving over 100,000 hosts of the Internet system and located in the U.S., Europe and the Pacific. The experiments are designed to evaluate the availability, accuracy and reliability of international standard time distribution using the Internet and the Network Time Protocol (NTP), which has been designated an Internet Standard protocol. NTP is designed specifically for use in a large, diverse internet system operating at speeds from mundane to lightwave. In NTP a distributed subnet of time servers operating in a self-organizing, hierarchical, master-slave configuration exchange precision timestamps in order to synchronize host clocks to each other and national time standards via wire or radio.The experiments are designed to locate Internet hosts and gateways that provide time by one of three time distribution protocols and evaluate the accuracy of their indications. For those hosts that support NTP, the experiments determine the distribution of errors and other statistics over paths spanning major portions of the globe. Finally, the experiments evaluate the accuracy and reliability of precision timekeeping using NTP and typical Internet paths involving ARPANET, NSFNET and regional networks. The experiments demonstrate that timekeeping throughout most portions of the Internet can be maintained to an accuracy of a few tens of milliseconds and a stability of a few milliseconds per day, even in cases of failure or disruption of clocks, time servers or networks.

Jitter-based delay-boundary prediction of wide-area networks

The delay-boundary prediction algorithms currently implemented by transport protocols are lowpass filters based on autoregressive and moving average (ARMA) models. However, recent studies have revealed a fractal-like structure of delay sequences, which may not be well suited to ARMA models. In this paper, we propose a novel delay-boundary prediction algorithm based on a deviation-lag function (DLF) to characterize end-to-end delay variations. Compared to conventional algorithms derived from ARMA models, the new algorithm can adapt to delay variations more rapidly and share delays robust high-order statistical information (jitter deviation) among competing connections along a common network path. Preliminary experiments show that it outperforms Jacobsons algorithm, which is based on an ARMA model, by significantly reducing the prediction error rate. To show the practical feasibility of the DLF algorithm, we also propose a skeleton implementation model.

On the long-range dependence of packet round-trip delays in Internet

Many previous studies on network traffic demonstrate that long-range dependence (LRD) is a ubiquitous property of traffic both in a local area network (LAN) and a wide area network (WAN). So the performance of the network should be dominated by this property. The packet round-trip delay is an important measurement of network performance. We present evidence that LRD exists in packet round-trip delays. This discovery has serious implications for understanding the impact on network performance of LRD network traffic, and the design of transport control protocols for special applications, i.e., of teleconferencing. Statistical analyses show that the complementary probability distribution of packet round-trip delays decays more slowly than at an exponential rate; this fact probably justifies the studies on the prediction of the queue length distribution with LRD network traffic. We also tentatively use a multi-queueing system to interpret the existence of LRD in the packet round-trip delay process, which we believe is caused by the LRD of Internet traffic.

A brief history of NTP time: memoirs of an Internet timekeeper

This paper traces the origins and evolution of the Network Time Protocol (NTP) over two decades of continuous operation. The technology has been continuously improved from hundreds of milliseconds in the rowdy Internet of the early 1980s to tens of nanoseconds in the Internet of the new century. It includes a blend of history lessons, technology milestones and series of experiments that shape, define and record the early history of the Internet and NTP.This narrative is decidedly personal, since the job description for an Internet timekeeper is highly individualized and invites very few applicants. There is no attempt here to present a comprehensive tutorial, only a almanac of personal observations, eclectic minutiae and fireside chat. Many souls have contributed to the technology, some of which are individually acknowledged in this paper, the rest too numerous left to write their own memoirs.

The fuzzball

The Fuzzball is an operating system and applications library designed for the PDP11 family of computers. It was intended as a development platform and research pipewrench for the DARPA/NSF Internet, but has occasionally escaped to earn revenue in commercial service. It was designed, implemented and evolved over a seventeen-year era spanning the development of the ARPANET and TCP/IP protocol suites and can today be found at Internet outposts from Hawaii to Italy standing watch for adventurous applications and enduring experiments. This paper describes the Fuzzball and its applications, including a description of its novel congestion avoidance/control and timekeeping mechanisms.