Conversation

@MarkCallaghanDB

From the NewSQL paper by

and

"it it is affordable to store all but the largest OLTP databases entirely in memory" I suspect s/affordable/feasible/ but am curious about the data. RAM isn't free and to get more RAM in the cloud you also pay for more CPU.

5

3

23

Peter Geoghegan

@petervgeoghegan

Replying to

@MarkCallaghanDB

and

A 2018 paper written by David Lomet ("Cost/Performance in Modern Data Stores") argues this same point very convincingly IMV. Perhaps you're familiar with it already. If not, see: ci.spdk.io/download/paper

4:22 PM · Nov 18, 2021Twitter Web App

@MarkCallaghanDB

Replying to

@petervgeoghegan

and

I knew of it, put it on my (too long) todo list and forgot. Now it is back at the top of the list and looks interesting. Thank you.

1

Peter Geoghegan

@petervgeoghegan

Replying to

@MarkCallaghanDB

and

There's a talk too, for what it's worth:

Cost/Performance in Modern Data Stores – How Data Cashing Systems Succeed - Microsoft Research

Data in traditional “caching” data systems resides on secondary storage, and is read into main memory only when operated on. This limits system performance. Main memory data stores with data always...

1

2

Show replies