The economics of distributed computing

This is referenced in Tim Bray’s grid summary, but if you didn’t read that, you missed it. Jim Gray (Microsoft Research) published an interesting paper in 2003 that appears to be holding up well to the passage of time.

The abstract summarizes the state of things in a provocative way

Today there is rough price parity between (1) one database access, (2) ten bytes of network traffic, (3) 100,000 instructions, (4) 10 bytes of disk storage, and (5) a megabyte of disk bandwidth.

This is a cost driven analysis, and remains true as long as the cost of telecommunications doesn’t decrease faster than Moore’s law (which it didn’t, at least between 2003 and 2006). The paper concludes

The recurrent theme of this analysis is that “On Demand” computing is only economical for very cpu-intensive (100,000 instructions per byte or a cpu-day-per gigabyte of network traffic) applications. Pre-provisioned computing is likely to be more economical for most applications—especially data-intensive ones.