Learn how the environment impacts timing and system performance, how oscillator choice can boost profits, why traditional timing parameters may no longer apply, and recommendations for evaluating oscillators for today’s applications.
Traditional metrics of timing performance, such as stability, drift, aging, Allan deviation, jitter, and so on, are well established in the literature and quoted in datasheets. But as our reliance on accurate data deepens and becomes more pervasive, the environment in which we store, compute, transfer and consume this data becomes more distributed and dynamic. The traditional metrics are no longer adequate to predict real-world performance. Modern systems, such as those used to enable autonomous driving, edge data centers, and 5G densification, demand a new class of timing products that can perform in real-world environments and not those carved out in datasheets. A need for durable timing is here, in which devices maintain precision when subjected to real-world conditions, including shock, vibration, electrical noise, thermal gradients, acceleration, pressure, etc. This talk explores the need for durable timing in data centers and 5G RAN applications. We'll discuss how the environment impacts each case's timing and subsequent system performance. For example, learn how the proper choice of the oscillator can boost data center profits by $5M annually. We'll discuss why the traditional parameters used to select timing devices may not apply and recommend those to evaluate instead.
Jagdeep Bal – Director, Technical System Customer Engineering, SiTime