Software as a Service vs Infrastructure as a Service

In an earlier post, I talked about two challenges which fuel cloud computing: the capital and support costs of network infrastructures and desktop anarchy. I want to pick up here where I left off, with more discussion of three cloud computing opportunities.

Software as a Service probably has the most familiarity. The wispy tendrils of Software as a Service reached into my Thanksgiving this year. Usually, we don’t host, and this year we will. To reduce operational costs, frustrated with the support costs involved with email, I put the family invite list, arrival and departure logistics and food contributions up on doodle.com.

It’s off my desktop and phone system into their system--simple, collaborative, low cost, centrally maintained, reliably backed-up, eminently scalable (hopefully won’t need that for our thanksgiving), platform independent, secure (enough for this), environmentally sustainable (carbon offsets for how much we may eat).

Doodle is a light weight alternative to meetingwizard.com or more elaborate reservation systems, and though its light-weightedness caused some other problems, it helped.

Meanwhile, Linda tapped into foodie SaaS epicurious.com to supplement the menu planning. Yes, we also have the 3-ring binders of family recipes and a shelf of cookbooks. Epicurious (OK, I prefer allrecipes.com) brings a world of collaboration, search, on-demand, pay as you go to the meal planning type of project. Print magazine co-sponsor Gourmet magazine is about to close its doors, but no doubt SaaS Epicurious will continue.

Holiday photo-sharing? Facebook, Flickr, Google web albums and not bring-your-own laptop, flash drive or photo album.

At the consumer level, Google has championed this approach more than any other company. Google entices millions to leave the desktop behind and do our daily work with on Gmail, Google Documents, Calendar, YouTube, Maps, Groups and more. Yahoo and others add to the offerings.

Once you get comfortable, you may find yourself surrounding yourself with other services. You may well use a hosted pay as you go service like Basecamp, Zoho or Central Desktop to mange projects. You may use Mozy, Carbonite, Dropbox or other web based services for back-up. You may use webex, adobe connect, or logmein to provide hosted screen sharing or web conferencing services.



Going back ten years, at the corporate level, Salesforce championed the software as a service approach to managing contacts, leads, business relationships, and all things related to the sales process (and now support and other business functions). Instead of network installing Act, Goldmine, Microsoft stuff at the low end or Siebel and other systems at the high end, they invited customers to use their web service. Pay according to what you need month-to-month, and let them take care of security, back-up, support, upgrades and the like.

The SF foundation charitable donation and “starter pack” has extended Salesforce’s appeal to nonprofits as well.

As I suggested earlier, these types of services typically may get justified on cost basis. Specifically, they transform long term capital and human resources costs into operating expenses. Today’s recession favors this calculation. Instead of figuring out the total cost of ownership of a new server, new software, support costs, licensing, periodic version updates, desktop upgrades, back-up, security and other ancillary utility purchases, SaaS lets organizational planners focus on one thing—the monthly usage cost. They also favor collaborative approaches to getting things done that are not as easy or natural on traditional network installed systems.

As critical applications move up to the cloud, the desktop does not matter very much. All that matters is having a standard browser.

Yet, to get to those web software as a services, you still have you to boot up that familiar personal computer and land on your personal desktop. Google hopes to go further, turning its Chrome browser into an entire web-facing operating system. More about that another time.

For now, for most of us, the desktop still matters a lot. Most of us still do our daily work there—word processing, maybe email, maybe other stuff. It matters for installed donor and contact management systems, network accounting software, critical long time installed databases whether Microsoft Access or some giant client server application. It matters when you want to print, scan, archive, share files, and avoid malware all the while. When these issues come up, the costs in time and support in monitoring and managing desktops come into play.

IT as a service (or Infrastructure as a Service) looks at things from this end. If the premise of SAAS is to make the desktop matter less, with IT as a Service, the idea is to tame it.

IT as a service has levels and gradations. It often begins with combining and rationalizing network servers using virtualization. VMWare and other companies have popularized the idea of running mutual virtual machines on the same hardware. Many businesses and organizations today have more than one server. They may have a basic server for sharing files and printing. Another for Exchange email management. Another for their fund-raising software or other special applications. Virtualizing servers takes advantage of increased computing power to reduce the total number of servers, and focus on keeping those few up and running well.

A related trend is to centralize and standardize the user experience. You start up your machine and then connect via Microsoft Terminal Services or Citrix into central server which then gives you a new, fresh, controlled desktop. For larger companies, creating that standardized network infrastructure makes it easier to then have IT infrastructure specialists manage everything from afar using elaborate diagnostic tools such as Kasaya.

The higher level next step is to move the whole infrastructure up to the web. If everything is centralized, virtualized, standardized, why not seek even higher efficiencies of scales by running the whole operation on someone else’s servers, where bunches of other organizations also are getting managed at the same time in the same place.

Ask you local network support organization if they have this kind of offering and compare prices with what you are doing and planning now. I recently got a major taste of this approach at a day-long business summit run by Zumasys, a West Coast infrastructure company which has helped some of our larger clients, and I see other partners heading in this direction.

Of course, like SaaS, IT as a Service come at a cost. Both point in the direction of flexibly supporting organizational expansion, especially in a recessionary era, reducing support costs, and reducing user woes and problems, and also reducing the cost of a desktop computer.

The goal is usually not to reduce on-site IT staff, but to keep that staff from burning out and to keep it focused on strategic support and innovation. By commodifying basic support issues, IT staff can focus on what is critical and distinctive about what you need and do.

Next up, the newer “aaS,” Platform as a Service.

Cloud computing and taming the desktop

Cloud computing is one of those buzz phrases that has come to mean everything and nothing depending on your perspective.



Cloud computing is neither good or bad: it is increasingly an element of planning and strategy even for small to medium organizations. It is new technology that happens to correspond to, accelerate and enable us to respond to larger economic and social trends. At the risk of oversimplifying, I tend to break it down to two challenges and three opportunities.


Two Challenges

Cloud computing wrestles with two fundamental challenges of the modern computing era: infrastructure and the desktop.

Without a doubt, the bits and pieces of organizational infrastructure have gotten cheaper. Putting all those hardware and software pieces together, managing them, securing them, updating them, backing them up, and ensuring 24x7 global up-time remains challenging and expensive.

In the current recession, cloud computing has become the means to translate technology from capital costs to operating costs. That is, instead of investing in longer terms cycles of network servers, other office hardware and software, cloud computing says, pay as you go over the Internet. Business planning articles have commented on this as an imperative about this recession—with which the maturing of cloud computing happens to coincide.

Formerly, organizations planned on upgrading locally installed network servers every three to four years. Or should have. The cost of not doing so might be a catastrophic failure with days of down time. The cost of an upgrade includes new hardware, usually new network software to be compatible with the new software, ancillary updates to security systems, back-up software, email software and so on. The cost also includes migration, planned instead of emergency downtime, reconnecting desktops and lots more. Big costs, hard to consistently budget for, particularly in a recession, and particularly for smaller businesses and nonprofits in recessionary environments.

These one-time costs can be enormous even if the pure cost of a server has gone down. Cloud computing aims to turn these capital outlays into ongoing operational costs. They can be metered as usage grows or retracts instead of planned over multiple year period.

Cloud computing also can alleviate the budgeting for support costs. It’s not that there’s necessarily less to watch or do. It’s more the scale and scheduling of support these days. Even if system consoles have become somewhat more straightforward to manage, the burden of responsibility may not have. More organizations support staff working at home or in the field at odd hours. More organizations have multiple offices or close partners sharing systems. More organizations maintain connections between internal data systems and their website. These all imply closer and closer to 24x7 support, even if an organization nominally has traditional office hours.

The other big challenge on the mind of IT strategists today is the desktop. Twenty five years or so into the “PC Revolution,” we’re grappling with what to do with the desktop. We all wanted its promise; now we have to contend with the consequences. Desktops have become more and more powerful. As they do, they also become more and more difficult to tame and manage.

Before PCs, everything ran off of servers, and desktops were relatively cheap, dumb, easy to maintain. Support was centralized on the server. While those servers and related infrastructure cost more, support costs were perhaps more predictable.

Even as purchase costs for a new office computer drop, analysts still consistently talk about total costs of ownership to a business or organization dwarfing that initial cost. (I’ve recently heard 15% as current ratio of initial to total cost.) The total cost includes acquiring and configuring the computer, maintaining it with security and software updates, supporting its users, and then managing its replacement down the road. Windows-Mac-Ubuntu each have their proponents in this regard. Travelling around to different offices, I see universal complexity on the typical staff person’s desktop and unending monthly support requirements.

Jonathan Zittrain’s excellent “The Future of the Internet—And How to Stop It” (http://futureoftheinternet.org/ ) writes about the fundamental openness of the PC environment. And this is true whether Windows, Mac or Linux. By their nature, unlike earlier devices that sat on office workers’ desks, they provide endless choices for installing software and configuring things. Endless choices mean endless headaches for IT staff.

Troubleshooting why some configuration no longer works on a particular machine (it is hardware? Software? Malware? User error?...) is probably the least fun part of an IT dept’s week. Zittrain argues for keeping that creative openness against the forces that want to put computer usage back into “walled gardens.” (This is the “future of the Internet he wants to stop.) Yet he also recognizes the pressures facing IT departments in wanting to lock things down or limit choices.
It would be hard to impossible to imagine giving up the computing power now at everyone’s finger tips on their desktop. The challenge is to provide that power with less support complexity.

Three Opportunities

From two challenges come three opportunities: “Software as a Service,” “IT as a service,” and “Platform as a Service.”

Software as a Service means taking strategic business or organizational software applications and moving them up to the web as pay as you go services. For example, on a small scale—Constant Contact. On a large Scale—Salesforce. Things run in the browser and operate like household utilities.

Platform as a Service takes a similar model and provides software developers the tools to build complete applications. For example, instead of doing your own Drupal thing, you develop complex sites from within the Acquia Drupal Platform.

IT as a Service (or Infrastructure as a Service) aims to move your whole network infrastructure out of your office and up onto the web somewhere. Instead of worrying about configuring Outlook on everyone’s desktop, everyone uses the same desktop software image centrally maintained and managed.

The first two of these meet the challenge of the desktop by minimizing what runs there versus what runs in your browser. IT as a Service meets the challenge by taming and centralizing what runs on the desktop. All three aim to reduce the capital expense footprint of IT. There is more to say. (For one thing, I've oversimplified a bit what Acquia and Salesforce offer). In part II (yes, sorry, there is a part II of this), I’ll assess these three a bit more.