Autonomy or interdependence in distributed systems

You want your own processors and memory because 1) you want administrative

control over its scheduling policy and 2) having your own makes it

potentially more private. Having administrative control means you

can optimize for low latency rather than high throughput and that

you can be sure to have computing at 2am Monday morning as well as

at 2pm Thursday afternoon. If you are not careful, having administrative control turns you into

a "system manager", a very unpleasant prospect if the system is, say,

Unix; not that much better if i t ' s a Macintosh (installing new software

and I/O devices is time consuming and mysterious). One strong argument

for interdependent systems is that a site can hire a single system

manager to manage all workstations. But he will only be able to do

that if they all use a fairly standard configurations of hardware

and software. Even if they all run "stand-alone", the common management

reduces autonomy. You don't ever want your own name server, file server or authenticat

In  Proceedings of the 3rd workshop on ACM SIGOPS European workshop: Autonomy or interdependence in distributed systems?

Publisher  Association for Computing Machinery, Inc.
Copyright © 2007 by the Association for Computing Machinery, Inc. Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from Publications Dept, ACM Inc., fax +1 (212) 869-0481, or The definitive version of this paper can be found at ACM’s Digital Library --


> Publications > Autonomy or interdependence in distributed systems