Globally Synchronized Time via Datacenter Networks

Ki Suh Lee, Han Wang, Vishal Shrivastav, Hakim Weatherspoon

2016 · 78 citations · 32 references

Concepts

TL;DR

Current state‑of‑the‑art protocols lack bounded precision, achieving only hundreds of nanoseconds when idle and degrading to hundreds of microseconds under congestion. This work introduces Datacenter Time Protocol (DTP), a packet‑free clock synchronization protocol that attains nanosecond‑level precision. DTP implements a decentralized synchronization scheme on the physical layer, using control messages to exchange hundreds of thousands of protocol packets without affecting higher‑layer traffic, and can be deployed incrementally by replacing network devices. DTP eliminates non‑deterministic elements, incurs virtually zero overhead, and achieves precision bounded by 25.6 ns for directly connected nodes and by 4 T D (≈ 6.4 ns × 4 D) in general, with software daemons accessing the clock at better than 4 T precision and end‑to‑end accuracy better than 4 T D + 8 T nanoseconds.

Abstract

In this paper, we present Datacenter Time Protocol (DTP), a clock synchronization protocol that does not use packets at all, but is able to achieve nanosecond precision. In essence, DTP uses the physical layer of network devices to implement a decentralized clock synchronization protocol. By doing so, DTP eliminates most non-deterministic elements in clock synchronization protocols. Further, DTP uses control messages in the physical layer for communicating hundreds of thousands of protocol messages without interfering with higher layer packets. Thus, DTP has virtually zero overhead since it does not add load at layers 2 or higher layers. It does require replacing network devices, which can be done incrementally. We demonstrate that the precision provided by DTP is bounded by 25.6 nanoseconds for directly connected nodes, and in general, is bounded by 4TD where D is the longest distance between any two servers in a network in terms of number of hops and T is the period of the fastest clock (≈ 6.4ns). Moreover, in software, a DTP daemon can access the DTP clock with usually better than 4T (≈ 25.6ns) precision. As a result, the end-to-end precision can be better than 4T D + 8T nanoseconds. By contrast, the precision of the state of the art protocol is not bounded: The precision is hundreds of nanoseconds when a network is idle and can decrease to hundreds of microseconds when a network is heavily congested.

References

32