Definition
NALSD (Non-Abstract Large System Design, sometimes written non-abstract large-scale system design) is the system design methodology used by Google's Site Reliability Engineering organization and popularized by the Google SRE Workbook. The defining idea is in the name: a design is not finished while it is abstract. Boxes labeled "database" and "load balancer" must be iterated into concrete resources — how many machines, how much disk and RAM per machine, how many queries per second per replica — until the numbers prove the design actually works at the required scale.
The method iterates through a fixed ladder of questions: Is the basic design possible at all? Can we do better? Is it feasible at the real scale — and what does it cost? Is it resilient to the failure of a machine, a rack, a datacenter? Each pass forces back-of-the-envelope calculation: starting from requests per second and payload sizes, deriving storage, bandwidth, and machine counts, and letting those numbers reject or reshape the architecture.
NALSD-style rounds are a standard part of Google SRE and infrastructure interview loops, and the style increasingly appears in senior system design interviews elsewhere — interviewers push past the block diagram with "how many machines is that?" whether or not they use the acronym.
How it comes up in interviews
In a NALSD-style round, the architecture diagram is the warm-up; the evaluation is whether your numbers hold up. Interviewers listen for stated assumptions, unit discipline, per-machine limits (a disk does ~100 MB/s sequential, a server holds N QPS), and designs revised because a calculation demanded it. Practicing estimation until it is automatic is the single highest-leverage preparation.