![]() Registered Member ![]()
|
What do you think true cloud computing would look like? Do you think Google's approach is what real cloud computing looks like, or do you think it would work differently, somehow?
Madman, proud to be a member of KDE forums since 2008-Oct.
|
![]() Registered Member ![]()
|
After a decent bit of thought (prompted by running networked X applications through SSH), this is what I think cloud computing would look like, if I were a fat guy with a massive wallet and a Swiss bank account:
You would register your username and password online with a (my?) server, and the server would create a Linux account under your name (and possibly a Windows account). From there, you would use a cloud-enabled device - in this case, I'm going to say a netbook as described below. Netbooks would come with 16GB -> 32GB SSD hard drives and between 512MB -> 1GB of RAM, as well as a high-speed (VERY high-speed) mobile broadband connection (these would be contract devices, like phones, or possibly offer pay-as-you-go services for internet connections). It might include a separate tiny SSD drive, like 512MB, for the, "miniature OS" (as explained below). The SSD hard drive wouldn't actually hold any data what-so-ever when the computer's switched off. When switched on, it would be an instant-on device that would ask you for a username and password, which would be registered with a (my?) server. Other then that, the log-in screen would have a menu, similar to the one currently in GDM and KDM, asking what kind of session you'd like to start - and it would have stuff like KDE, KDE 3.5, Gnome, TTS and possibly Windows. When you finish typing in your username and password, the netbook authenticates with the server, then, depending on which session you chose, the netbook would download the session to the SSD (using a tiny GNU/Linux base tuned for the laptop and discluding some unnecessary software or CLI software that standard users aren't likely to use (unless you selected TTS, or otherwise somehow specified that you wanted to download the extra stuff to your laptop)), plus the desktop environment you chose: on first start-up, KDE would look something like it does in Kubuntu, providing only the most useful applications (and KOffice instead of OpenOffice - this is the future, can we imagine that we could replace OOo with KOffice? Please?!). Add and remove programs wouldn't actually, "install" or, "remove" applications - it would download the application from the server and store (somewhere - probably on the server, for reasons I'll give below) the fact that you, "installed" or removed the application (Get firefox and remove Konqueror, for example). This setting would be used next time you log in, for intelligently downloading only the applications you chose to use on the netbook. All document management would be done locally: the directory tree would look like, again, the default of some KDE distribution, with stuff like /home/[username]/Documents etc.). When you save a document in the Documents folder, it stays there. Say, heaven forbid, the thing crashes, it's all right: the document's still on your hard-drive, so you can still get it. Say, your laptop battery dies: it's all right, you can charge it up, switch it on and skip the log-in process - the start-up/log-in screen miniature OS thingy looks at the other SSD and says, "Hey! You've still got stuff there," letting you carry on where you left off, safe in the knowledge that your documents are still intact and retrievable locally (similar to sleep/hybernate?). So lets say you've stuck a wad of music on the device, have several documents and other files on it. When you log off (note: safely), the netbook makes the connection to the server again, overwriting your old files and folders with the new one and saving the, "installed" applications in some configuration file somewhere. Then, the SSD drive whipes itself and the netbook shuts down. The reasoning behind the last part is that the netbook is essentially useless without log-on credentials to access the server: even with access to the hard-drive, someone that steals your netbook won't have access to any of your data. Additionally important, though, is that the server never overwrites data on the netbook's SSD, the netbook's SSD always overwrites the server's stored data (if it has permissions to the directories - no overwriting new applications with old ones, thank you!) and data is only downloaded from your home directory on the server if the SSD is already empty (hence, no overwriting). This is making the presumption that the version of the document on your SSD will always be more up-to-date then the version on the server, until you log out, in which case no copy exists on the SSD at all. It also means that, as long as you've logged in, the binary files on your netbook will never be overwritten by more recent versions (due to updates) on the server, which will avoid some funky behaviour (imagine if Firefox decided to just crash or freeze because suddenly, the libraries it's using have been updated but the binary hasn't). This is scary and controversial, but there are also many benefits: 1. You can use your own server. With the above system, you can configure the default server to log-in to, allowing, for example, businesses to do a single install of an application and make it instantly available to all the netbooks everywhere, or the opposite: no netbooks are updated, as the version of applications on the server isn't updated, garuenteeing interoprability all the time. It also means that data you can configure the client/server connection such that the data never leaves the local network, and all data is stored on a single server. Additionally, if the log-in screen provided switching between servers/services, you can actually use the same device for both business and home use and keep them completely separate: it would be as if using two difference computers. Additionally, all these netbooks would connect to the internet via a single connection: through that server. I suppose some other method would be used for on-line non-local servers - a direct connection to the internet, probably, rather then going through the server 2. security, safety and ease of maintenance. Stuff like virus scans, file system checks etc. would happen on the server (though what happens to the clients when a scheduled file-system scan is started would be interesting... I presume there'd be some fail-safe or some kind of communication from the server to the clients, saying, "Don't send me anything yet! Just keep it locally for now" if they try logging off. What happens when they try logging in, though, is a different matter...). Similarly, backing-up files would be done entirely server-side, twice or three times: if something goes arwy on the server, the netbook could still connect to some back-up server and get your documents/applications back. 3. "Home" servers could be provided as well, allowing one household to have four or five of these netbooks but have all files, maintenance and such on the home server hard drive. This would also ensure that the documents are always accessible, and that, regardless which of these netbooks you picked up, you could always log-in and retrieve the same documents over again. *edit(s): corrections. >.<'
Madman, proud to be a member of KDE forums since 2008-Oct.
|
![]() Global Moderator ![]()
|
Have you seen Jolicloud?
Moult, proud to be a member of KDE forums since 2008-Oct.
thinkMoult - source for tech, art, and animation: hilarity and interest ensured! WIPUP.org - a unique system to share, critique and track your works-in-progress projects. |
Registered users: Bing [Bot], blue_bullet, Google [Bot], Yahoo [Bot]