An IEEE-1588 compatible RADclock

Davis, Matthew; Villain, Benjamin; Ridoux, Julien; Orgerie, Anne‐Cécile; Veitch, Darryl

doi:10.1109/ispcs.2012.6336624

Cited by 8 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most clock synchronization protocols (e.g., NTP and PTP) run in a time daemon, which periodically sends and receives UDP packets between a remote process (or a time server). Unfortunately, the overhead of system calls, buffering in kernel and network interfaces, and direct memory access transactions can all contribute to errors in delay [25], [27], [38]. To minimize the impact of measurement errors, a daemon can run in kernel space, or kernel bypassing can be employed.…”

Section: Problems Of Clock Synchronizationmentioning

confidence: 99%

See 1 more Smart Citation

Globally Synchronized Time via Datacenter Networks

Shrivastav

Lee

Wang

et al. 2019

IEEE/ACM Trans. Networking

View full text Add to dashboard Cite

Synchronized time is critical to distributed systems and network applications in a datacenter network. Unfortunately, many clock synchronization protocols in datacenter networks such as NTP and PTP are fundamentally limited by the characteristics of packet-switched networks. In particular, network jitter, packet buffering and scheduling in switches, and network stack overheads add non-deterministic variances to the round trip time, which must be accurately measured to synchronize clocks precisely. We present the Datacenter Time Protocol (DTP), a clock synchronization protocol that does not use packets at all, but is able to achieve nanosecond precision. In essence, the DTP uses the physical layer of network devices to implement a decentralized clock synchronization protocol. By doing so, the DTP eliminates most non-deterministic elements in clock synchronization protocols and has virtually zero protocol overhead since it does not add load at layer-2 or higher at all. It does require replacing network devices, which can be done incrementally and with very small amount of hardware resource consumption. We demonstrate that the precision provided by DTP in hardware is bounded by 4TD where D is the longest distance between any two nodes in a network in terms of number of hops and T is the period of the fastest clock. The precision can be further improved by combining DTP with frequency synchronization. By contrast, the precision of the state-of-theart protocol (PTP) is not bounded: The precision is hundreds of nanoseconds in an idle network and can decrease to hundreds of microseconds in a heavily congested network.

show abstract

Section: Problems Of Clock Synchronizationmentioning

confidence: 99%

“…Further, TSC counters are employed to estimate the frequency of the DTP counter. A TSC counter is a reliable and stable source to implement software clocks [25], [50], [55]. Modern systems support invariant TSC counters that are not affected by CPU power states [10].…”

Section: A Accessing Dtp Countersmentioning

confidence: 99%

Globally Synchronized Time via Datacenter Networks

Shrivastav

Lee

Wang

et al. 2019

IEEE/ACM Trans. Networking

View full text Add to dashboard Cite

show abstract

“…Further, TSC counters are employed to estimate the frequency of the DTP counter. A TSC counter is a reliable and stable source to implement software clocks [46,50,25]. Modern systems support in-variant TSC counters that are not affected by CPU power states [10].…”

Section: Accessing Dtp Countersmentioning

confidence: 99%

“…Because NTP normally does not provide precise clock synchronization in a local area network (LAN), much of the literature has focused improving NTP without extra hardware. One line of work was to use TSC instructions to implement precise software clocks called TSCclock, and later called RADclock [25,46,50]. It was designed to replace ntpd and ptpd (daemons that run NTP or PTP) and pro-vide sub-microsecond precision without any extra hardware support.…”

Section: Related Workmentioning

confidence: 99%

“…Most clock synchronization protocols (e.g., NTP and PTP) run in a time daemon, which periodically sends and receives UDP packets between a remote process (or a time server). Unfortunately, the overhead of system calls, buffering in kernel and network interfaces, and direct memory access transactions can all contribute to errors in delay [25,27,36]. To minimize the impact of measurement errors, a daemon can run in kernel space, or kernel bypassing can be employed.…”

Section: Precision Errors Introduced By Network Stackmentioning

confidence: 99%

See 1 more Smart Citation