May 03, 2006
The Fallacies of Distributed Computing
I heard about the 8 fallacies of distributed computing few years ago. Last year I finally tracked them down online at James Gosling's site. The fallacies, attributed to Peter Deutsch, were coined long before that--back in 1994, in fact.
- The network is reliable.
- Latency is zero.
- Bandwidth is infinite.
- The network is secure.
- Topology doesn't change.
- There is one administrator.
- Transport cost is zero.
- The network is homogeneous.
As the short preface to the fallacies on Gosling's site says, these are assumptions that almost anyone starting to build distributed systems is tempted to make. Yet all of them prove to be wrong in the long run--resulting in all sorts of troubles and pains for the solution and architects who made the assumptions. While (thankfully) I didn't assume all of them on the first couple of distributed systems I designed, I did assume some of them (like transport cost is zero and the network is reliable). To help you avoid this trouble, I will expand on each fallacy in the next few posts.
By the way, you may also want to take a look at some of the realities of distributed computing that I posted last year.
Posted by Arnon Rotem-Gal-Oz at 10:43 AM Permalink
|
