3 ms·
This actually has nothing to do with the web in general, or the internet, or politics, or personal liberties, or peer to peer networks. This is about designing
by peterwwillis 13y ago
This actually has nothing to do with the web in general, or the internet, or politics, or personal liberties, or peer to peer networks. This is about designing stable distributed applications.
Almost every really big, really stable website (or network service) is built on a set of distributed applications. Global data redundancy is just one of the considerations. If you want your application available everywhere, all the time, you have to design it to withstand faults, to distribute data and computation, and to do this over long distances, and still perform the same actions the same way everywhere. That's all a de-centralized web needs to look like, and it's actually already implemented in many places.
What I think Tim is saying is that we need to move away from concepts that centralize data and computation and use existing proven models to make them more stable in a global way. And i'm totally behind that. But if you think we're going to get there with p2p, self-hosted solutions, new protocols, new tools, new paradigms, or a new internet, you've completely missed the boat. We have had everything we need to accomplish what Tim wants for a while. It just has to be used properly [which some companies actually do].
But good luck convincing most companies to spend the time and money doing that...
- dclara 13y agoI partially agree with you in the sense that we don't need new network with p2p or self-hosted solutions, but need stable distributed applications. How? I cannot agree with you to realize it via data redundancy and replication. The way you mentioned is just what Google is doing right now, which is not true distributed computing. Google's architecture relies on replicating services across many different machines to achieve distributed computing/storage and fault-tolerance across the world. See the reference here: http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/googlecluster-ieee.pdf http://static.googleusercontent.com/external_content/untrust... The real distributed computing solution for providing a well-organized web has been proposed to build a layer on top of the existing physical infrastructure of the internet to present the web in a de-centralized manner under a systematic way. http://bit.ly/MwT4rx http://bit.ly/MwT4rx
- peterwwillis 13y agoAs far as i'm aware there's no such thing as "true" or "real" distributed computing. It's a wide field of study with many different overlapping concepts. The paper you linked to does describe how queries are resolved by Google, and not quite the whole picture required for a typical distributed computing model. But to say it's not distributed computing is... confusing. Clearly they must support this model over many sites the world over, and we know from other documents that they geo-locate data to make it more convenient, meaning it's portable and partitionable. Sounds like a distributed computing system to me.
- dclara 13y agoYes, I agree that it's hard to define what is true distributed computing. That's only from my personal opinion. A good reference could be from Wikipedia: a distributed system is a software system in which components located on networked computers communicate and coordinate their actions by passing messages. The components interact with each other in order to achieve a common goal. You can find from the blog page here about distributed computing: http://bit.ly/1jkySqU http://bit.ly/1jkySqU