I see more and more reference to the Cloud, especially in IBM ads. Sure would like to know what it is. Thanks.
There have been some recent threads about this that might be worth looking up.
But the basic idea is that “the cloud” means that you are using IT hardware, network infrastructure and/or applications through the Internet so that you don’t know and don’t care where they are physically hosted or by who. Rather than buy your own server with its own hardware, applications, etc. you are using a part of a server hosted by someone else, such as IBM.
It’s a rather vague term referring to software applications which are hosted in a shared environment by a dedicated provider over the Internet. (The Internet being represented by a cloud icon on network diagrams.)
“Cloud” services cover a pretty broad range. Examples include:
Amazon EC2, which allows you to order up virtual machine instances that run in Amazon’s data center and run whatever software you like on them.
Google App Engine, which allows you to write web applications against their specific API, which are then automatically run on their distributed architecture which scales on-demand automatically.
Hosted applications, such as Oracle on Demand, which allows you to pay for Oracle database capacity on an as-needed basis (you can also set up your own private Oracle clouds for your company’s needs.)
Et cetera.
Furthermore, this assumes that you will always have sufficient reliable network connectivity that you can just assume that the network is there, and that your connection will have sufficient bandwidth to support what you are doing.
The extreme endpoint of this is a situation where your personal computer has no local storage. All your videos, letters, memos, business documents, etc, are only stored in The Cloud, and you pull them across the network whenever you want to even look at them. This the way things like Google Docs work.
Not necessarily the greatest idea, especially in rural areas. Plus, if you keep your work documents in “the cloud”, and you lose your connection, you’re screwed if you don’t have a local copy.
Just a few hours ago I was watching the playoff game, and one of those “to the cloud” commercials came on. I note how stupid there were since they don’t give you any indication about “the cloud” actually is, or why you can’t do whatever it is they are touting without “the cloud”.
Webpages and any other networked service (like XBox Live or Arena Net) requires computers called servers to be online which the person who is using the service can connect to.
Historically, companies would set aside a room and fill it with servers. They would employ people to sit there and watch the servers on the off-chance that one goes down and needs to be rebooted.
Now, on lots of diagrams of how things work, the internet would be shown as a cloud. I’m not sure exactly why – though probably because Visio or some similar diagramming software included a cloud as an emblem for the internet – but there it is.
Your average diagram would show the client with a dotted line connecting it to the cloud, and then on the opposite side a dotted line connecting the cloud to your server.
If you let someone else run the servers for you, then it’s like putting your servers in the cloud, instead of behind it. You stick your software up into some random place on the internet, and your clients can access it there directly without having to come to you.
So that’s what a Cloud Solution is. It’s a company which runs a bunch of servers and charges you rent for putting your stuff on their machines.
Like Sage Rat says, the name comes from network diagrams where a cloud is used to represent “other stuff we’re not worrying about right now” like the internet or the rest of your company’s architecture. It’s like in a graph when you put that squiggle in the axis to mean “we’re not showing this part because there’s nothing interesting there.”
Is this not what Larry Ellison (Oracle) was promoting in the late 80s/early 90’s?
There are few truly new ideas in computer technology. Things that sound very similar to cloud technology have been around since mainframes were invented. However, it is all about the details and implementation which is why I believe that the cloud concept is different than all of the similar ideas in the past.
It isn’t just forgoing your own server to rent one somewhere in Asia where it is cheaper. The best parts of the cloud mean that you can access your applications and data from many different types of devices from desktops to smart phones and you don’t need to worry about where the servers are physically located anywhere in the world. Enterprise cloud computing would also include globally distributed failover protection and the data being stored in different places so you don’t have to worry about losing your data because of a fire or flood in a local server room. It is all protected and distributed and this is a very big deal to larger businesses. The cloud also means that you can dynamically scale up or scale down computing power and bandwidth based on demand because it is shared.
It is similar to having a generator powering just a few houses in your neighborhood and that is all they have even if it fails versus distributing the load across thousands of power stations that are all interconnected to serve a huge area.
The cloud synergizes collaborative architectural platforms on a decentralized platform, leveraging bleeding edge game changing applications across content delivery networks.
I read up a bit on the Amazon EC2 product, it seems the vast majority of companies using this will be hosting website with open source programs in the traditional LAMP setup. Question: regarding the HPC option, I currently use NX Nastran at work with our puny cluster (and decent desktops) for engineering simulation, how does the cloud affect licensing arrangements between me and my proprietary software provider? Are they ok me running jobs on the cloud, as long as the instances check out the licenses at my private license server? Does amazon license professional engineering software from the developers and sublet the licenses? Is any of this possible at all? Do they have any engineering simulation softwarenrunning the GPGPUs?
“The cloud” refers to several things, and is mostly a marketing ploy (IMHO). As stated by a previous poster, it’s essentially SaaS (Software as a Service). But, since everyone wants to be trendy, there are private vs. public clouds (hosting a cloud ((usually in your own datacenter but managed by a third-party)) or ((using a third-party for comms, but hosting your own super-secret intra-intranet))).
In IT-World, there is a fluctuation between server-side management and host-side management. If you save documents and such to the hard drive on your machine, you’re host-side; if it all goes up to a server, you’re server side. “Cloud computing” is a trendy way of saying “server-side IT management.” That’s the trend right now.
Despite the unavoidable sarcasm (I get sarcastic about both models), I don’t intend to imply that it’s bad. It’s just exactly what it is - higher data availability. But that comes at the expense of higher risk if the repository is compromised (aka “hacked/cracked”).
Amazon doesn’t do any licensing. You can put whatever you want on your EC2 instances, but if you need per-host licenses then you have to figure that out with the licensor yourself.
Hell, I was using this back in the 1970’s. With dumb terminals, on a mainframe.
Much of the mainframe systems operated this way. I could sit at a terminal at my desk, and work on files that were stored on disks at the mainframe site. And when I submitted a job to run, it might run on the mainframe in the basement, on the one a few blocks away, on on the one a few hundred miles away in the next state. I could choose which one, or let the system decide where to run the job. But it just happened somewhere out there. And if one of the mainframes was down, things just automatically went to another one. That’s really a lot like the current explanations of “the cloud”.
At some of the later consulting jobs I had, I did not even know where the mainframe was physically located at all. Sometimes I could figure out the state, based on the area code to reach the operators. But I didn’t know the actual town or address – and I didn’t really have to know. It was somewhere out there, at the other end of the wires.
Not much is really new in this business.
Like a few years back, when people were pushing ‘thin clients’ connected to ‘blade servers’ – I could hardly see any difference between dumb terminals connected to a mainframe.
Well, blade servers are not really analogous to mainframes in any particular way. It’s a hardware configuration that allows you to pack more compute power into a given amount of rack space by moving common components like power supply and networking to a shared chassis and then sticking small, specialized servers (the blades) in as needed.
I’m not sure that anyone has yet nailed it. That probably reflects the different cloud solutions available. 
IMHO at its core, the cloud is an abstraction of service, just as a SAN is an abstraction of data storage, and as a DFS allows invisible redirection of storage.
It’s a marketing term like the “Information Super Highway” or “It puts a tiger in your tank”.
The question is where the information you use is stored: On your local computer, or in some service. For example, Flickr is a “cloud”. You store your pictures on Flickr, and their accessible anywhere. If you have a Mac, iPhoto is not a “cloud”. You store your photos in iPhoto, and they’re only available on that computer.
The whole idea of cloud computing is coming up because many companies are now building massive computer facilities and they plan on using them.
Plus, we are now more use to using and storing information “in the cloud”. This message board is “in the cloud”. Google docs are “in the cloud”. Google is pushing ChromeOS that will turn your computer OS into a mere web browser. Not only is all of your data stored “in the cloud” but so are all of your computer applications.
The advantage of storing information on a remote service is that its available everywhere. Chrome OS can detect viruses and reinstall its operating system. If your ChromeOS computer is stolen, your data and applications are safe. Simply buy a new one, and log back into “the cloud”.
The disadvantage is if that service is closed down, you could lose your information forever. For example, Delicious a service by Yahoo to store web bookmarks maybe closing down.
Then, there are services like Dropbox. Dropbox creates a folder on your computer called Dropbox. Store a file there, and it’s physically on your computer. However, all files stored there are automatically copied to Dropbox’s server (which use Amazon’s S3 service). Thus, you can visit the Dropbox webpage from any computer and access those files. Even better, if you sync that computer to your Dropbox account, it will download those files to that computer Thus, being both “cloud” and “non-cloud”.
It’s all just marketing.
Isn’t it a way for software companies to make sure they can charge you for the use of their programs?
Although they are advertising the “cloud” to consumers, the primary initial target of these types of services is businesses.
There are 2 primary advantages over having local infrastructure:
- Don’t need expertise to purchase, setup and maintain a bunch of servers
- Changes in required computing power can be handled flexibly without having to swap out a bunch of servers, etc.
Well, your data is safe from loss, in that the ‘cloud’ company probably is more careful about backups than most of us are.
But it’s no safer against theft – identity theft, credit card number theft, etc. Probably less safe, since a big ‘cloud’ company is a more attractive target for hackers or thieves than your little PC. Despite the security efforts of these companies, there are regular reports of thousands of credit card numbers being compromised at one of these companies. Often by rouge employees or contractors.
So “safe” is rather nebulous here.