Wouldn’t you consider a vault a server? It’s serving encrypted chunks of data to the app that requested it. I’m asking this question because I feel it might cause some people who are familiar with how the current internet works to dismiss the project.
I guess it depends how technically strict you want to be, I would say average people associate a ‘server’ with a centralised server, don’t they?
I wanted to see what would come up for results when searching for “no servers” and found this recent article - http://www.wsj.com/articles/look-ma-no-server-corporate-it-expects-serverless-computing-to-trigger-big-changes-1477509674 Yet there is still centralized servers being used.
EDIT:When I originally posted the link it loaded fine and now is asking me to sign up. :-\
I’ve been trying to get my head around serverless - it basically just means that as a developer you don’t have to deploy a virtual machine in the cloud to run code - you just upload the code and it runs. It’s the next step on in terms of abstraction from containers (Docker and the like). As you rightly say, servers are still very much there in the background.
“Serving” data is one function of a vault, but it is nothing like a server really.
I think it is more confusing to refer to a vault as a server. It is pretty clear that when people talk about client-server they aren’t referring to a peer-to-peer configuration.
I’ve been saying “no servers” for some time and don’t think anyone is actually confused by it. One or two people have raised this point, but I think only after they already understood what the network is. So saying “no servers” seems more of an invitation to find out how that can be done, and what that means. For someone who doesn’t understand network architectures, you can always elaborate by saying…
“well, the network behaves like one enormous server, but it is very different to the current internet because nobody controls it, it is literally everywhere rather than in a big building somewhere, and so impossible to hack, blow up, or suffer a power outage etc - unless a large part of the planet such as a continent is affected, and even then it should be able to recover with no loss of data. So no servers or every computer taking part is a part of one enormous server!”
But starting off I think its fine to say “no servers”. Worth discussing though, so just my opinion
“No servers” good “serverless” bad…
I think a vault is a server on steroids. Servers today do more then serve data(Email, DNS, VPN 's, etc.) also.
It could be looked at as an invitation by some and push others away. I guess it all comes down to the individual.
In web development/design, a ‘server’ usually refers to a central computer that houses user data, and one that can be queried directly for this data. In other words, the relationship between clients and servers is many-to-one. In maidsafe, this is not so. Key differences being that: one computer DOES NOT hold the entirety of the data; rather, it is split into multiple encrypted chunks and sent to different computers in a untraceable fashion. Furthermore, the relationship between server and clients shifts from 1 to many to many to many. Even further still, vaults cannot be queried directly for data by users/application programs, only the routing algorithms can decipher (pun intended) the location of user data, stitch it back together and return it to the owner. In short, this is a vastly different software architecture than a traditional client server system. And really they are not close in their implementation or design thus justifying the phrase ‘no servers’.
Yeah, SAFE has potentially many billions of servers everything’s a server! yay
Server-client is how the term servers is being used on the internet. “dropbox” and storage system is the term used for data/file storage.
So vaults are not classed as “internet” servers, but they are storage/dropbox systems.
The OPs use of vaults being servers is the older usage in local networks where server-client architecture wasn’t really in use and servers were the storage systems. But now storage systems are their own term and servers are the “server” in server-client systems
I think it’s fine, there needs to be a wording that resembles the differences. A lot might think steem is decentralised, when it fact it relies on servers.
I agree. I think the key meaning of “no server” is in the sense of the centralised client-server-architecture which stands in contrast with the peer-2-peer-architecture we have. While you’d still have an entity serving another one data, in P2P the roles do constantly change. And while one vault might be serving you data at some point – being a server in that moment, it might as well be recipient or relay other times.
Most importantly, in this architecture anyone can be the server. So there is “no server” in the meaning of the famous quote that if everyone is special then no one is.
Seems like it would simpler and more accurate to just always stick with the singular, ‘no server’. Not that I think there will be much confusion tbh, but that keeps it tied to the idea behind it… no centralisation.
Yeah, I think “no server” is fine, just avoid “serverless” which has come to mean something else entirely, ie Amazon Lambda, Microsoft Azure Functions, and IBM OpenWhisk
Would you call torrent nodes servers?
Torrent software can be made to run on a server, as can safe vaults. But what participates in the network is a node.
Thanks for everybody’s input.
I post these questions to clarify my line of thinking and give me ammo to reply to people I talk with about the SAFE Network.
Clarity has been restored.