Yea so the infrastructure comments really shocked me but it makes complete sense. They had to do massive scaling early on and a lot of what they use now is from

Yea so the infrastructure comments really shocked me but it makes complete sense. They had to do massive scaling early on and a lot of what they use now is from that era. I can imagine even at google they have some sort of "if its not broken dont fix it" mentality. There is a reason why megastore is sluggish right? Its tries to uphold CAP theorem across datacenters and because of that write performance suffers.