As a SharePoint farm grows beyond 80 million indexed items, co-locating index, crawl, and content processing components on the same servers creates I/O contention and crawl queue bottlenecks. This post covers the large farm reference architecture — dedicated index servers, distributed index partitions, and isolated crawl databases — and how to implement it with PowerShell.