Question abount milvus sizing tools

[hjlee9182]

Why is the difference between 10m and 11m so big?
I think it's weird that the storage size is 10 times different, so can I know why?

[yhmo]

This page lists the use case of lite/standalone/distributed:
https://milvus.io/docs/install-overview.md

Milvus Standalone is suitable for medium-sized datasets, ranging from 1M to 100M, with an average 10M.

In the sizing tool, if the "Number of Vectors" exceeds 10M, it automatically switch from "Standalone" to "Distributed"

For distributed, all the dependencies(etcd, minio, pulsar/kafka) are independent services. They all have replica configurations, the required resources is also much larger than standalone.