The Complete Definition of The “Cloud” It’s actually the Internet, but with additional capabilities. Essentially, cloud computing means that information is no longer stranded on your own individual machines (desktops, laptops, smartphones, tablets) but is instead combined into one digital “cloud” available anywhere on any number of devices simultaneously, managed in cyberspace on “someone else’s” computer(s). This is much like the way we access our electricity through an electrical “grid” which distributes power from a huge electrical plant located somewhere else (See The Best Analogy, below). This may be all the explanation you need - but for what this means to you as a consumer (as opposed to a business), see below. See the sidebar for the full (perfect, but incomprehensible) NIST definition. For a more detailed discussion which you may require as a business, read on:
In addition to browsing websites or sending e-mail, the “Cloud” can provide (free or leased) programs (for example, word processing or databases, perhaps through “Google Apps”) over the Internet, rather than requiring the installation of each program on every one of your devices. Moreover, it can store and share on the Web all of the data you create using those programs. There are three basic types (a/k/a “development models”) of clouds - public, private, and hybrid (e.g. where a company may run most of its business in-house, but outsources some particularly burdensome tasks, like data analysis):
Also, there is what is called “community” clouds, which are not quite hybrid clouds, but rather are more of a specialized cloud available strictly for member businesses. [For example, the New York Stock Exchange has built one of the first community clouds, called the Capital Markets Community Platform, which provides all member companies with not only tremendous resources and speed, but also specialized services, apps and analytics to suit the particular needs of each of the members, as needed.] UPDATE: See also the definition for cloudlet, at the edge of the Internet for mobile devices.
The upside: Your programs are always available anytime, anywhere, so long as you have a computer or other device and an Internet connection. And it is probably less costly than purchasing all of the programs and storage hardware outright for each device. For enterprises, the Internet has always been a way to store huge amounts of data on an outside server farm. Now, for individuals, as the amount of data generated increases during our digital lives, it is becoming a more attractive storage alternative. From simple off-site file services like Drop Box to full off-site drive backup services like Acronis, the cloud is becoming more attractive to home and small office users as well.
The downside: You must have an active Internet connection, usually broadband speed, to access the program and the data, and there’s always the security worry you have when someone else is storing your data for you. If you decide not to use the Cloud sometime later, you don’t own any software. If you have a large organization and your host goes out of business (as did Coghead in 2009, data transferred to SAP), it can take weeks or months to transfer terabytes of data and, at say 10 cents per gigabyte, this can add up to tens of thousands of dollars.
And even the enormous cloud providers can periodically go down: Witness - Sony (April 20 through May 14, 2011 and after); Hotmail (December, 2010 - when it came back on line, 17,355 e-mail accounts went missing [eventually returned]); Windows Phone Marketplace (April, 2011 down for extended maintenance - isn’t Microsoft large enough to build in maintenance without disruption?); Windows Azure, which failed to recognize leap day February 29, 2012 and was down worldwide on October 30 and 31, 2013 ; the June 30, 2012 Amazon server outage which took down Netflix, Instagram, Pinterest and others (brought about by a thunder and lightning storm in Virginia) [this was after the 2011 Easter weekend outage the previous year] and again on October 22, 2012 which knocked out popular sites like Reddit, Pinterest, TMZ, Instagram, Flipboard, AirBnB, Imgur and GetGlue; the 2007 Rackspace outage when a transformer was hit by a car; Data wiped out for those who got hit by superstorm Sandy in 2012; Dropbox’s security breach in July, 2012 and Amazon’s EC2 cloud service (which struggled for four days in April, 2011 to get its customer data and databases back on line after a human error occurred as the result of a configuration error during a scheduled update during the middle of the night - don’t fall for their excuse that the system was still up 99.5% ‘cause it doesn’t matter if your data isn’t available!). Even in 2013, cloud failures have been experienced by Amazon (1/31/13, 49 min.), Dropbox (1/10/13, 16 hrs.; 5/30/13 1.5 hrs.; seems to be a pattern here), Facebook (1/28/13, 3 hrs, 6/18/14, complete, all day), Microsoft (2/1-2/13, 2 hrs.), Google Drive (3/18-19/13, 17 hrs.), CloudFlare (3/3/13, half hour), Fail Whale [Twitter] (6/3/13, 45 min.), Google (7/10/13, 1.5 hrs.); 2014: G-mail, Google+, and YouTube all went down most of Friday January 24, 2014, as did Azure for most of a week in August, 2014. In fact, during 2014, the following cloud servers were down: Dropbox (1/10/14 - 2 days; 3/14/14 - 1 hr.); Google (1/24/14 - 1 hr.; 3/17/14 - 3.5 hrs.); Basecamp (3/14/14 - 2 hrs.); Samsung Smart TV (4/20/14 - 4.5 hrs.); Evernote (6/10/14 - 10 hrs.); Adobe Creative Cloud (5/16/14 - 28 hrs.); Feedly (6/11/14 - 3 days); iCloud - 6/12/14 - hours); MS Lync - 6/23/14 - hours); MS Exchange - 6/24/14 - 9 hrs); and many other shorter outages at Verizon, Google Drive, FaceBook, etc. You get the picture. And, while many were missing out on instantaneously knowing what their long lost high school flame ate for breakfast, the effect on sales for a corporation like Amazon can amount to millions of dollars a minute. Time really is money. Even worse, your cloud data could be completely erased, gone forever. Take Journalspace - All of its files were deleted by a vengeful ex-employee of the company. All of it blog posts, photos and other attachments, spanning years of lives, simply disappeared. As did the company shortly thereafter. There was no backup, no fail safe, no off-site archive.
How does this happen? Despite the astronomical odds of failure, simultaneous and cascading failures occur in the most unpredictable ways. Click HERE for the amazing stories about how random events cause some of the catastrophic failures noted above.
And this doesn’t include the effects of intentionally hacking cloud servers. In early 2013, it was reported that Microsoft, Apple and Facebook were victims of hackers. In February, 2013, customer support tool Zendesk reported that it was hacked, exposing the e-mail addresses of users of three of its clients - Tumblr, Pinterest and Twitter. On March 4, 2013, Evernote, purveyor of note-taking pens and software, announced that it had been hacked as well, exposing its fifty million users to possible exposure of their e-mail addresses and user names, although the passwords, being robustly encrypted, were probably not at risk. If your data is stored on a public cloud, it can be hacked more easily than if it’s under lock and key at your own physical location. Ask the celebrities like Jennifer Lawrence who had their nude photos hacked from iCloud on September 1, 2014 (which begs the question: Why even save this kind of stuff in the cloud, anyway?). Basically, the cloud isn’t much different from the storage server you may have at home or the office, with the key exception that cloud servers are rented while traditional servers are owned. And because they are rented, they are under someone else’s control and not yours. If you own the books in your library, you have complete control over them. But, if they’re stored in the cloud, not so. You can only get to your stuff by using that other person’s security controls. And there’s the problem, the loss of control and vulnerability to intrusion. There’s not much you can do about these hacks, except perhaps to change your password, and watch out for any communications requesting verification of personal or billing information or passwords. This, however, still is and must remain one of the main downsides to cloud computing.
Moreover, don’t forget that some types of data (i.e. medical, defense, technological) must be stored and/or transmitted by law only within the U.S., which can preclude those out-of-the-country cloud server farms. And, as clouds become deployed for more than just test data, it appears that the cost of full deployment can ramp up very quickly. For example, one cloud user which backed up a public cloud cluster of some 250 computers reported a charge of $23,400, over ten times what was planned. Consequently, there is an entire cottage industry of vendors like Cloud Cruiser, 6Fusion and Uptime Software which collect and interpret cloud bills, developing ways to cut the increasing costs of cloud computing for larger companies.
Also there’s the general security issue: If you store everything in the cloud and your computer is hacked (see below) or your laptop stolen, it’s a simple matter for someone to get all of your data, something that might not happen if you store your data on external devices such as tapes, disks or flash drives (see Media). Case in point: All those nude celebrity photos that were hacked from the cloud and posted in September, 2014.
And the above illustrates a dark side of the cloud: Because it is anonymous as well as powerful, hackers can use it to propagate viruses over the Internet without fear of being caught and at very little expense, sometimes only a dollar or two per hour. And, not to be too pessimistic, but what if there is a war or cyber-terrorism brings down the net. Or a natural disaster. Or your provider goes belly up. Where would you be then?
Caution: If you’re a business, you’ll probably find that vendors are all jumping on the bandwagon, claiming that their services and apps are “clouds”. Unless they fit the definition (below), they’re just repackaging their products and services. Cloud computing isn’t really a completely new concept, but the most recent scale expansion of so-called “hosted software” (see ASP), which has been around for over a decade. The difference between cloud and ASP is that, with ASP, an outside firm hosts an instance of an on-premises application for you. True cloud providers are both “multitenant” (that is, all clients run on the same vendor stack) and “version now” (every one is on the current version). So don’t confuse ASP with cloud. Also, don’t confuse the “elasticity” model” with the “utility” model. An in-house virtualized computer environment, while it centrally serves multiple computers, is not a cloud, because it is not a utility hosted over the Internet (see below). See also D2C (D2D2C).
At its highest (most abstract) level, cloud computing is defined as computing power or services that are delivered over the Internet by an outside company as a utility and are billed based upon usage. (The old SaaS model rented entire packages of software primarily to larger business users and charged hefty monthly or yearly server fees; but small businesses and consumers can’t afford and won’t pay for this. The new pay-for-use-only model has been driven by their needs. Smaller businesses, even startups, can now compete by using computer power that would have previously been unimaginable for their size. And even larger businesses are happy to pay only to level off spikes in usage, when and as needed.) Governments, as well, have found it advantageous to use cloud computers rather then incur the large expenses involved in upgrading computer systems and software.
The Best Analogy: In 2008, Nicholas Carr wrote a book titled “The Big Switch” which traced the way that companies which started out with their own giant electrical generators in-house eventually switched to purchasing only the electricity they required when they needed it from a select group of large providers, freeing themselves up to focus on their core business. Arguably, computer power has reached the same level and appears to be resolving itself the same way through the use of cloud computing, a pay-per-drink provision of computing power freeing businesses to focus on their core business while large computing suppliers provide the heavy lifting. Another similar analogy would be the sharing of broadband internet - not everyone can have their own asynchronous satellite orbiting the earth, but we all can share the ones owned by Comcast and Verizon.
Presently, the big three hosts are Microsoft (Azure), Amazon (Amazon Web Services, a/k/a “AWS”) and Google. Below them are the incumbent providers, IBM, AT&T, Verizon, HP and the like. To be safe, it’s best to get a cloud provider that has SSAE 16 (formerly SAS 70) certification.
Cloud services fall into three main ”service models”: SaaS (“software as a service”), such as salesforce.com or Gmail; PaaS (“platform as a service”), such as force.com or Google Apps and IaaS (“infrastructure as a service”), such as Amazon web services). [See the chart below.] [Naturally, as of late, DaaS (Data as a Service), for the storage of databases in the cloud, is coming into its own as well as Big Data emerges as a force.] “Software As A Service” [pronounced “sass”], refers to a practice of delivering software to a customer over the internet on a rental basis, rather than the traditional “pay and own” model. Users tap software from the “cloud” rather than buying it outright and loading it onto their computer, often paying for usage in increments of time. This term has generally replaced the earlier terms ASP (Application Service Provider) and On-Demand. These “cloud” offerings include databases [Google, IBM, Yahoo, Amazon], backup [Mozy/EMC], storage [IDrive], virtual servers [Terramark], online storefronts, payment processing, CRM and other vertical applications. In addition, web giants such as Amazon [EC2 - “Elastic Computer Cloud 2”/Cloudfront Simple Storage Service (“S3”)], Google and EMC are now making their vast infrastructures available to a variety of web-accessable services for a very reasonable charge (e.g. 10 cents/hr, 10 cents/Gb). In October, 2008 Microsoft announced its own cloud computing system, Azure. [As of 2016, Amazon Web Services (“AWS”) dominates the market with a 32% share, followed by Microsoft Azure at 13%, IBM at 7% and Google at 6%.]
Unfortunately, as the result of the various issues described above, business cloud users have now found it necessary to create a “backup of the (cloud) backup” as part of their disaster recovery program, just in case the cloud servers are down. [Is this really progress??]
Offshoots of SaaS are PaaS [pronounced “pass”](a/k/a cloudware) where users are renting not just the individual web-based applications, but an operating system to run those applications, a network as well as a service provider that will perform basic maintenance on that network when necessary; and IaaS [pronounced “I as”] for Infrastructure as a Service, where the user is renting hourly or monthly the entire system: Virtual servers, storage space, virtual routers, switches and other hardware, networking capabilities, an operating system and applications, with the cloud provider (“Host”) offering even more extensive maintenance and services. All this, with the ability to “scale” up or down whenever the requirements change. Clouds can be “bare metal” (non-virtuaized infrastructure) or virtualized in nature. Even the Federal Government has experimented with cloud computing: The Defense Information Systems Agency (“DISA”) has awarded contracts for on-demand computing services.
Computer clouds are rapidly reshaping the way businesses access IT services. Cloud computing differs significantly from the earlier type of less-than-successful network and web based services, known as MSPs (Managed Service Providers), which typically required customers to procure processing and storage capacity in discreet units of a particular type of server over a set contract period. Also, there is a huge difference in scale, because Google, Microsoft, Yahoo and others can offer such vast data centers, and therefore computing power, at a magnitude never before available. There are different types of clouds - while Google may be ideal for sifting through data, salesforce.com may be better for running business applications like customer management and accounting software, even allowing companies to write their own programs to run on its servers. In 2008, IBM (Blue Business Platform) and Google announced a plan to roll out a worldwide network of servers from which consumers and businesses will tap everything from online soccer schedules to advanced engineering applications through the cloud model. Google has also introduced the Chrome browser, which is cloud friendly. And, not to be outdone, Microsoft developed its own cloud applications, known as Azure, in early 2009. A year later, Microsoft introduced the beta of Office 365, which is an amalgam of Exchange Server, SharePoint Server and Lync, for a monthly cost between $10 and $25 per user, which gives small businesses (say, 25 or so employees) the benefits of large users. In 2007, when Microsoft introduced its “Live” series of services, it introduced Skydrive (originally “Windows Live Folders,” now “OneDrive”), a free cloud storage service for up to 25 GB of personal storage with a 100 Mb file limit, which was expanded with the introduction of Windows 8. With the introduction of Office 2013, the cloud became a permanent embodiment of Office 365, although you could still purchase the program separately as a stand-alone app.
There’s also a related service known as “cloudbursting,” which is the ability to increase capacity by provisioning resources from an additional cloud, a kind of “spillover” effect between clouds. And beware of cloudwashing, the attempt by a vendor to repurpose an old product or service by associating the buzzword “cloud” with it, a deceptive practice designed to increase sales.
Confused? When I wrote this description in 2009, almost no one knew what the cloud was. Surveys show that even now 80% of people are still confused by the term. However, they are probably using it on a daily basis (see figures on sidebar), especially if they own an Apple or Android device, engage in social networking (Twitter, Instagram, Facebook, Reddit and the like) use Office 365 or Google Docs or download books, magazines or other content. And way back in the 1970s, AOL was the original cloud provider; if your computer crashed, merely signing in to your account on a new computer restored your address book, e-mails, even your stock portfolios. Users’ e-mail, tax preparation software and online gaming may come through the cloud, even though users don’t necessarily think about it. So, too, cloud storage and file sharing programs like Dropbox (see below), iCloud, OneDrive and Google Drive. [Click HERE for a 2016 comparison of these four providers.] Almost every computer backup services offer on-line, scheduled cloud backups as well. [Think about the magnitude of it: Dropbox, for example, has roughly 50 million users who (according to the site) upload a billion files to the service every 48 hours; That’s gotta be a lot of files to protect!] And office suites like Google Docs and Office 365 are cloud offerings. Much of the discussion above, however, is for those businesses who are making the move to provide such customers cloud-based services, and not the residential users who access them. See the chart below for a graphic explanation: