What exactly is the "cloud".

SlimCini

KC and the Sunshine BANNED
How big is a server physically? How much volume or space would 10000tb of server storage take up exactly?
 

steaky360

Moderator
Moderator
How big is a server physically? How much volume or space would 10000tb of server storage take up exactly?

Datacentres can be huge, but a server doesn't need to have 10000Tb of storage to be a server - a server can be as small as a SFF desktop (or probably even smaller!!)

facebook-one_2588746k.jpg

Apparently that's FB's arctic circle datacentre
 

ubuysa

The BSOD Doctor
How big is a server physically? How much volume or space would 10000tb of server storage take up exactly?

A server is a single IP address, but the file storage it supports can be almost any size you can imagine. For example, when I connected to the PCS website my browser connected to the PCS web server (a single IP address). I would guess that the storage capacity of this server is quite modest, it only has to store the html of the web pages for example. When my email client connects to my mail server it also connects to a single IP address. The storage capacity of this server is probably quite large, since it has to store all the emails for all it's customers.

So, in pure TCP/IP terms, a server is a computer running server-side applications and bound to a single IP address. We usually reference that server via a domain name (pcspecialist.co.uk, for example) and use DNS to map that name to an IP address (198.41.206.44). The communication between the client and the server is a single one-to-one communication (often over a single TCP connection).

As an experiment I just connected to Dropbox (a cloud service) and uploaded a small file. My Dropbox client connected to four different servers to complete that process (I'm sniffing the network packets with Microsoft Network Monitor, that's how I know). So, in this example, Dropbox is not a single server, it's at least four, and they may well be in different geographical locations - even different countries (and I have no idea what's happening in the back-end, those servers could be contacting other servers for backup and/or resilience). That's the point of the cloud, it's not a simple client/server communication involving one client and a specific server, it's communication between one client and many servers, all of which are cooperating to provide a specific service. And that's just a simple example, other cloud services might involve dozens of servers in widely dispersed locations. We don't need to be aware of any of that though but to indicate that it's not simple client/server they have coined the term "cloud computing".
 
Last edited:

GeorgeHillier

Prolific Poster
That's what I'm saying - for your free 10GB Google Drive account, no chance. But, for businesses negotiating contracts and paying for 'cloud' storage, it must be an option so they can keep to data protection laws (or can minimise the number of different jurisdictions' laws they need to comply with).

I would assume the data would be hosted on all the servers in the country you pay from. e.g. an English company will have its data stored on English servers.

Also, I believe that your data is stored on a couple of the servers, so they would be on 2 or 3 servers within the company's country.
 

ubuysa

The BSOD Doctor
I would assume the data would be hosted on all the servers in the country you pay from. e.g. an English company will have its data stored on English servers.

Also, I believe that your data is stored on a couple of the servers, so they would be on 2 or 3 servers within the company's country.

That may be true, but the point about the cloud is that you don't know, and you don't need to know either. With traditional client/server you know which server you're talking to, but with the cloud you're likely talking to several servers (and possibly not even the same servers each time you store stuff in the cloud) and you have no idea where they are. That's the key difference.
 

Androcles

Rising Star
and you have no idea where they are. That's the key difference.

And the biggest reason to avoid using the cloud like the plague in my opinion. I want and need to know where my stuff is, I don't want to worry about it rolling about god knows where out there somewhere.
 

steaky360

Moderator
Moderator
Maybe its my ignorance but I don't see the difference (in real terms) between knowing some data is stored on a hard drive located 250 mile away from my house (over which I effectively have zero control) and a collection of hard drives potentially located anywhere...

Surely regardless of the physical location the data is just as safe? I guess I'd 'feel better' knowing where it was.. but when I think about it most of my 'online presence' is stored in places I have no idea about nor control over.

Its probably my ignorance... sorry for interjecting.
 

ubuysa

The BSOD Doctor
Maybe its my ignorance but I don't see the difference (in real terms) between knowing some data is stored on a hard drive located 250 mile away from my house (over which I effectively have zero control) and a collection of hard drives potentially located anywhere...

Surely regardless of the physical location the data is just as safe? I guess I'd 'feel better' knowing where it was.. but when I think about it most of my 'online presence' is stored in places I have no idea about nor control over.

Its probably my ignorance... sorry for interjecting.

Actually I agree with you. What you should care about is that your data is protected from unauthorised access and protected against loss. Where it is physically stored is not important.

In any case, if you have no control over the server (or data store) on which your data is stored it's pointless knowing where it is.....

BTW. The cloud is not new, we were talking about exactly this back in the 1990's, although we didn't call it "the cloud". The buzz-word back then was "thin client", the idea being that (given the Internet) we no longer needed operating systems on our computers, all we needed was a network connection and enough client code to request the applications and data we wanted to use from the Internet (very similar to Google's Chromebook). Your computer would thus be a very minimal platform, hence the term "thin client". Naturally Microsoft wasn't keen, for obvious reasons, and in any case the bandwidth back then wouldn't have supported it in practice. But the idea that all the end user needs is a basic platform and that the data and applications exist "in the cloud" is most certainly not new, it's simply that the bandwidth now exists to make part of the "thin client" idea (the storage of data in the cloud) a possibility.

You young-un's don't have a monopoly on good ideas you know.... ;)
 
Last edited:

Androcles

Rising Star
Maybe its my ignorance but I don't see the difference (in real terms) between knowing some data is stored on a hard drive located 250 mile away from my house (over which I effectively have zero control) and a collection of hard drives potentially located anywhere...

Surely regardless of the physical location the data is just as safe? I guess I'd 'feel better' knowing where it was.. but when I think about it most of my 'online presence' is stored in places I have no idea about nor control over.

Its probably my ignorance... sorry for interjecting.

I have a lot of control over the data on my hosted server, I can control my own security, adjust any firewall settings I like, partition my allocation how I like, assign DNS to different parts of it, have parts secure and parts public, host websites and voice servers, far more control than on the cloud, and I know it is somewhere being looked after by a company I can trust being backed up on a regular basis and not some random server that has a loose connection with some other service that may or may not have shoved a copy of my data on some other random server that I can't be sure is secure because I have no idea who it is and what tech they use.
 

ubuysa

The BSOD Doctor

Wow, he made that sound complicated, what rambling load of ....

He's reacting to what Vint Cerf said a few days ago (see my post on here https://www.pcspecialist.co.uk/forums/showthread.php?40311-Is-Vint-Cerf-wrong) about the problem that old data requires both an application that understands it, an OS that supports the application, and a hardware platform that supports the OS. My take on all this is in the other thread.
 
Top