Distributed fiber sensing possesses the unique ability to measure the distributed profile of an environmental quantity along many tens of kilometers with spatial resolutions in the meter or even centimeter scale. This feature enables distributed sensors to provide a large number of resolved points using a single optical fiber. However, in current systems, this number has remained constrained to a few hundreds of thousands due to the finite signal-to-noise ratio (SNR) of the measurements, which imposes significant challenges in the development of more performing sensors. Here, we propose and experimentally demonstrate an ultimately optimized distributed fiber sensor capable of resolving 2100000 independent points, which corresponds to a one-order-of-magnitude improvement compared to the state-of-the-art. Using a Brillouin distributed fiber sensor based on phase-modulation correlation-domain analysis combined with temporal gating of the pump and time-domain acquisition, a spatial resolution of 8.3[thinsp]mm is demonstrated over a distance of 17.5 km. The sensor design addresses the most relevant factors impacting the SNR and the performance of medium-to-long range sensors as well as of sub-meter spatial resolution schemes. This step record in the number of resolved points could be reached due to two theoretical models proposed and experimentally validated in this study: one model describes the spatial resolution of the system and its relation with the sampling interval, and the other describes the amplitude response of the sensor, providing an accurate estimation of the SNR of the measurements.