Early Thoughts on Cloud Computing

The topic covered by this post is one which you've either been hearing a lot about, or will in the near future. It's an important subject that we all be familiar with, as we will need to address it with our clients and partners, and sooner, rather than later.

Let's start off with a definition:

Cloud Computing is "any situation in which computing is done in a remote location (out in the clouds), rather than on the desktop or handheld device." (Business Week ).

Or

In general, cloud computing builds upon the general concepts of Web 2.0 or software as a service (SaaS) for users to rely on the Internet to satisfy their computing needs.

Examples you may be familiar with are Salesforce.com, Amazon's "Elastic Cloud" or EC2, Google Apps, Microsoft Office Live, and our own LloydCare™ Hosted.

What's the "cloud" mean?

I think it's a cool term developed by the some marketing firm.

Don't get me wrong, we are in for a paradigm shift in how people use, access and pay for data and data processing.

In this new paradigm:

Your data, or a portion of your data, for the most part doesn't reside local to your workstation or servers. Although it may, because sometimes you need to work offline, and sometimes your applications demand that they be local, for either performance or licensing purposes. A really good example of this is the Amazon Kindle. I automatically download the newspapers, magazines, and books I want to read, and they are available to me offline or on. However, if I delete them from my local device, they are still available, theoretically forever, on Amazon's servers, for me to access or download again at will.
Another good example is Google Apps (a whole lot more on this in future posts) working with their Gears application allows you to work both online and take your data offline.
I work the same way with SharePoint, which syncs with my Outlook so I can work offline.
Your applications may not need to reside locally, they run instead in a browser, and therefore have the potential to be device independent. This means it doesn't matter if you're using a PC, a MAC, an iPhone, a thin-client (read cheap PC), or Blackberry. Or even perhaps an Amazon Kindle.
In the long run, you will:
Keep your data where it is cheapest to keep that data. Example, old archive documents someplace cheap, with a higher level of risk of loss of access, keep mission critical data someplace more expensive.
Be able to access that data from anywhere.
Buy application access as you need it, and
You will use a greater variety of applications:
Will you stop using Microsoft Word in exchange for Google Apps? Probably not, I think you'll use both, depending on what's cost, power and reliability are more useful to the given task, or what works best in interfacing with the partners you are working with.
Companies will save a tremendous amount of Capital Expenditures, slightly expand their Operating Expenses, and have a lot more money available to invest in really generating returns from IT.
And in that is a paradox, because I believe that if firms don't reinvest some of that money in improving productivity, their value proposition, or competitive advantage for IT, they lose all ability to differentiate themselves from their competitors, and overall improve their bottom line, through the use of IT. If we are all using Google Apps, and no one is customizing the applications or interfacing them with processes and tools you create specific to you, then you risk commoditization and generating lower returns for your stakeholders.
 

So, some facts:

This cloud computing thing has been tried before under various circumstances (Time-Sharing, ASP's), why is it now a possibility?
Most of you are familiar with Moore's law, which states that advances in technology will allow manufacturers to double the number of transistors on a given circuit board every two years, thus doubling (theoretically) the processing power. There is a similar law known as Grove's law, which states that the amount of bandwidth available for voice and data doubles every hundred years or so.
The "DOT Bomb" boom broke Grove's law (at least temporarily). During the late 1990's and early 21st Century telecom and other companies laid fiber optic cable. Lots of it. Verizon's desire to stay relevant is laying even more. At this point, all pretty much unprofitably.
Of course, the Obama Stimulus puts out billions more for broadband, so Grove's law will probably stay in place for the next several years, at least.
What fiber optics does for computing is what alternating-current did for electricity, it makes the location of the equipment, data and applications to be unimportant to the user, and allows disparate and formerly incompatible machine to operate together as a single system.
One interesting fact – the two biggest users of bandwidth right now are YouTube and Hulu. If this increases perhaps Grove's law will eventually come back into play.
Because of the above, the "PC Age" is giving way to the "Utility Age" ("cloud" computing")
Virtualization technology is here. For years the processing power and memory on computers has, for the most part, exceeded the users needs, for the majority of the time. In other words, a user might need 10% of their servers processing power at any given time, only once in a while increasing their usage anywhere near capacity. Virtualization technology allows one physical machine to act like multiple machines, and controls which "virtual machine" gets access to processors, memory, and storage based on current needs. We are making extensive use of this in our hosted offering.
Parallel processing is here. Thank you Google. The way Google works is to take thousands (millions) of computers, tie them together, and get them to act like one giant computer, running many things in parallel.
Big companies such as Amazon.Com and Google have built out huge datacenters, with the capacity to handle business at their peak, with redundancy built in. For example, Amazon uses 10% of its capacity most of the year, only using more around peak shopping periods like Christmas. They are now making this excess capacity available, at a price, to the end user.
Web 2.0, which really means that the Web is 2-way. There has been a lot of buzz about Web 2.0.
Web 1.0 went one-way. You published a page, someone read the page.
Web 2.0 is about interaction. The user can publish or modify. Wikipedia is one example, blogging or Facebook another. One great example to look at is a firm called Threadless.Com, where they've used Web 2.0 to take business process (and cost) and lay it on the customer.
Current competitors in the "cloud computing" space include, among others:
Google
Microsoft
Amazon
Lloyd Group
Billions and billions and billions of dollars are being spent, mostly to build datacenters near dams and other cheap sources of power.
You are going to see the biggest shift in how IT budgets are spent since the PC was invented.
What's the difference between each of the competitors?
Google's offering is centered around Google Apps, a competitive line of products to Microsoft Office and a host (no pun intended) of other applications. It includes storage and comes in both free and premium offerings. To become a Google Authorized Partner we at Lloyd are investing in a Premium Subscription and will look at using Google Apps for some of our business processes. I've begun testing them personally, and am using them to handle some of the functions within my family, communication and planning with my mentee at HBS, and to manage my role as Forum Chair for the NYC YPO Chapter. I'll let you know how these go, and ask for your feedback on how you find the applications to be when used internally.
Microsoft has many flavors, including traditional hosted Exchange and their new Live offering, which is designed to compete with Google. This is a huge effort in which they are investing billions of dollars.
Amazon's offering, or EC2, basically rents out server and storage space (virtualized) inexpensively to companies.
The Lloyd Group has built a hosted offering based on Microsoft Technology. We use virtualization in a way similar to Amazon, and manage our clients virtual environments the way we manage their physical environments today. More information and details on how our offering works will be published over the coming weeks. In the long-term, our offering will evolve to include not only what it does today, but to interface with each of the offerings mentioned above.
Who's going to win? I think everyone, but mostly the end user. In my opinion, there is no "network effect" in "cloud" computing. (A network effect is a phenomenon where there is only one winner, because so many users migrate to that one offering that even if it isn't the best or cheapest there is no way to compete against it.)
More on this topic, especially on Google Apps, in upcoming posts.