UCLA

Internet traffic has been growing exponentially due to the increasing popularity of bandwidth-hungry applications among mobile and cloud platforms, as well as the increased penetration of broadband technologies. The advancements in fiber optic communication technology has enabled the scaling of data rates to very high bandwidths to support the increasing global data traffic growth. This has come at the cost of making the optical network infrastructure too rigid to changes in demand, difficult to service and maintain, and energy inefficient. The next-generation of optical networks needs to be more agile so that it can dynamically scale network resources based on demand, identify and locate impairments in the network automatically, be resilient to network disaster and cybersecurity attacks, and be more energy efficient. This necessitates the development of software-defined networking at the physical layer of optical networks with rapid optical performance monitoring capabilities to be able to dynamically modify the network and its parameters to ensure high quality of service and energy efficiency. Rapid optical performance monitoring requires high-throughput instruments that can analyze high-speed optical signals accurately with low power consumption.

The increasing demand for higher data bandwidth is leading the communication industry to increase the operating frequency of components and systems. Performing accurate and rapid measurements of the various characteristics of electronic, electro-optic, and opto-electronic devices, modules, and systems at these high bandwidths is very challenging. With increasing operating frequencies, the conventional test equipment cost, size and power consumption also scales up proportionately. Longer test times to achieve increased accuracy would result in significantly higher test costs, hence innovations in faster testing of high bandwidth components and systems is a breakthrough that the test engineering community is anxiously awaiting for.

Electronic digitizers are an integral part of modern digital instrumentation systems. The resolution and speed limits of analog-to-digital converters presents the bottleneck in the development of high-throughput, high bandwidth instruments. Photonic time-stretch technology boosts effective bandwidth and sampling rate of the conventional electronic digitizers by 50 to 100 times by slowing down the high bandwidth signals prior to digitization and signal processing. Time-stretch technology is a fundamentally different approach to broadband digitization technology. Instead of increasing the speed of back-end analog-to-digital converters and digital signal processors to keep up with ever accelerating data rates, time-stretch slows down incoming signals before digitization, reducing the analog bandwidth of the signal. By employing this technique, lower speed, higher resolution, more energy efficient digitizers and signal processors can be used to capture and process full wide-band signals in real-time. In addition, the time-stretch architecture scales with digitizer and signal processor technology, continually improving in resolution, speed, and energy efficiency as the electronic back-end technology progresses.

In the first part of this dissertation, I discuss my work in the development of a novel photonic time-stretch accelerated processor for in-service, real-time burst-mode optical performance monitoring to enable the implementation of agile optical networks. This instrument

with an equivalent 2 Terabit/s burst mode processing capability measures the bit-error rate, an important metric for the quality of transmission, of 40 Gigabit/s data with a 28 microsecond acquisition time. With a low power consumption of 50 W, this instrument has a real-time burst sampling throughput of 250 Giga-samples/s using an electronic digitizer with only 2.5 Giga-samples/s sampling rate at 2 GHz analog bandwidth. The measured bit-error rate is transmitted as a feedback to the software-defined networking controller to automatically take corrective actions in case of impairments or network disasters. I also discuss the various demonstrations using this instrument such as in-service analysis of streaming video packets at 10 Gigabit/s on a commercial optical networking platform, characterization of various non-linear effects in optical fiber networks, and an ultra-fast instantaneous frequency measurement system.

In the second part of this dissertation, I discuss the development of an innovative high-performance single-shot network analyzer employing photonic time-stretch for extremely fast frequency response measurement of high bandwidth electronic, opto-electronic, and electro-optic devices, modules, and systems. This single-shot network analyzer has an effective sampling throughput of 2.5 Tera-samples/s at 40 GHz analog bandwidth. This instrument also features an extremely fast acquisition time of 27 nanosecond, a measurement jitter of 5.4 femtosecond, and power consumption of 50 W.