Where is your source for 'probabally' not running more than 20k nodes..
That's not what I said. I said "Most 20,000 node networks don't run on open source", go back and look. So you've found 1 that might - Google - but if the directory they're using isn't an open source product and proprietary to Google or someone else then it's still not a pure open source network. So you're yet to prove a single one, which is a far cry from "most".