A Sequential Model to Predict Frequency Variations over a Clock Network

点击翻译按钮获取中文摘要

Proper synchronization in a network is essential component to its operation. Inadequate synchronization**compromises quality of service, leading to such impairments as data retransmission, digital content distortion,**and severe degradation of encrypted services. Moreover, precise, accurate time enables virtually all**infrastructures such as data centers, wired and wireless communications, financial exchanges, industrial**networks, and other secure communications.**Today's networks are composed of thousands of independent clocks that produce data in time series format.**Keeping consistent time among different clocks in a network is a fundamental requirement of many networked**applications. Being able to model how clock frequency varies, two or more clocks will be in tighter sync**especially under conditions where the reference/master clock becomes unavailable. Currently, this problem is**not handled efficiently, as clocks will be synchronized once the master clock becomes available. This project**aims at building a data-driven solution for this problem.**In the proposed research, our goal is to build a framework to predict frequency variation of each clock over a**network of clocks and then evaluate it on a real network to empirically verify the predictions of the framework.**The ability to accurately predict the frequency variation of a clock even for a short period of time gives a**significant advantage of one solution over competitors. To do so with a flexible approach such that each clock**is modeled independently of other clocks would create a truly world-class solution to a difficult and important**problem. The solution will save a huge portion of synchronization time and costs and increases Microsemi's**competitive advantages. This will be directly beneficial to Canada as more revenue for Microsemi means more**investment from Microsemi in Canada and hiring more Canadian talent to expand existing products with**respect to the developed technology.