Slashdot Mirror
Scientific American Article: Internet-Spanning OS
Hell O'World writes: "Interesting article on Scientific American outlining what they call an Internet-scale operating system (ISOS). 'The Internet-resource paradigm can increase the bounds of what is possible (such as higher speeds or larger data sets) for some applications, whereas for others it can lower the cost.'"
This article makes one fatal assumption: Consumers will always purchase more powerful equipment than they need.
.NET and dotGNU waves are likely to make thin clients much more realistic.
The time of super fast home-PCs is likely to not last very long. The incoming
There is absolutely no reason for 'Mary' to have so much computing power since she doesn't need it. The only real limiting factor today is bandwidth which this article assumes anyway.
What is probably likely in the future though is a more distributed OS. One that is truely network transparent in every facet of operation. I believe there are some rumors floating around about MIT working on something to this effect...
int func(int a);
func((b += 3, b));
Devoting compute cycles to specific, worthy causes is great, but the point of an ISOS would be to make all connected hosts more powerful and efficient. If I want to factor a large prime or predict the weather, I might have hundreds or maybe thousands of otherwise idle computers available to help with the task. So each processor is constantly busy.
Privacy is very important but can certainly be worked out. For one thing, data could be stored in "bit stripes" so that each byte of your data is split into 8 separate streams but stored in more than 8 foreign hosts for redundancy and availability reasons. In that way no one could reconstruct any portion of your data from fragments on their drive and no laws could be broken by storing chains of bits.
Also private and public space could be partitioned off so that things you want kept on your system would stay there and only data associated with your weather predicting program would get stored on the ISOS. And quotas would need to be enforced so that if you donate 100GB to the ISOS storage then you may store, say 30GB (due to redundancy) in the distributed system yourself.
And perhaps your CPU's MIPS rating and uptime could be tracked to keep things fair. Then it would be almost like your computer storing up its processor cycles and getting them back all at once when you have a job to run. Grid computing makes sense and a World Wide Grid could make sense if it is feasible and the logistics could be worked out. Imagine everyone everywhere having the power of a supercomputer at their disposal.
The Helsinki user is no worse off in this scenario than if Mary's machine were a web server.
We all know that such "just in cases" do actually occur. The only solution to data-loss is redundant copies of the data, maintained either manually (explicit backups) or automatically (transparent mirroring or replication). The authors' idea is to go for automatic replication, and once you have that you might as well use the replicas to improve performance by allowing them to serve data to nearby nodes. This can actually result in less overall bandwidth than traditional approaches, because each node is going somewhere relatively close to get data instead of choking up a central server.
That actually highlights a flaw in the example as given in the article. It would be quite abnormal for someone in Helsinki to be going half-way around the world to get the data, because there should be a nearer replica. It would be more accurate, though perhaps less compelling, to say that Mary's machine was being used as a "staging area" for other local users watching the same movie from Helsinki that Mary just watched ten minutes ago. That would IMO convey the idea of an ISOS (actually the data-store part of it) actually reducing network bandwidth while also improving robustness.
Slashdot - News for Herds. Stuff that Splatters.