SaaS For IT Management
This blog is all about using Software as a Service (SaaS) for IT Management.
-
Is SaaS “Greener” Than On-Premise Software?
Posted on September 22nd, 2009 No commentsOver on the SoftwareAdvice blog, Chris Thorman has an article entitled “SaaS v. On-premises Software: Which One is More Green?” Rather than simply speaking in generalities, Chris actually crunches the numbers and gives examples:
On-Premise Energy Consumption:
We’ll use the example of a typical physician practice, since electronic medical records (EMR) software is a market we know well. The “On Premises” side of the graphic below illustrates a four-physician medical practice, running EMR software on their own in-house server.
The HP ProLiant DL server, one of the most commercially popular servers on the market today, will consume 7,008 KW of server energy per year. That’s running 24 hours a day, 365 days a year.
In addition, each user is using a Dell desktop 546, Dell’s most popular starter desktop. A single 546 Dell desktop will consume 600 KW of energy a year, running 8 hours a day for 250 days a year (an average work year).
A four-physician practice will consume 9,408 KW of power each year just to run EMR software on-premise. Each user will personally consume 2,352 KW of power each year.
Now, the SaaS model:
Now let’s see how the energy consumption of SaaS software stacks up. Rackspace, one of the largest providers of cloud computing hosting services, lists the Dell PowerEdge 2950 III as one of it’s most popular server choices. And since a data center would have a redundant server in addition to the PowerEdge, our SaaS applications are powered by two of these servers.
Running 24 hours a day for 365 days a year, the total energy consumption for these two Dell servers running SaaS applications would be approximately 6,570 KW/yr each, or 13,140 KW/yr total.
Running this server in a SaaS data center allows the SaaS EMR vendor to tens or hundreds of customers on one server. When a new customer goes line with the application, the incremental computing requires – an power consumption – increase only marginally.
How does this affect energy consumption by our physicians?
Now our physicians are only using 131.4 KW (1/100th) of the Dell PowerEdge server energy each year because of the multi-tenant architecture. Also, because SaaS applications require less computing power on the client, the physicians are able to switch to more efficient Dell netbooks, which only consume 120 KW of power each per year.
Using SaaS, our four physician practice now only consumes 1091.4 KW per year running their EMR software. That’s 272.85 KW per year, per physician.
That’s a 88% reduction in overall energy consumption for a four physician practice using SaaS EMR software over on-premise software!
So the lesson here is that even in a small, four physician practice, switching to a SaaS infrastructure can reduce energy consumption dramatically. And when you scale up the numbers, the incremental computing resources and power consumption only increase marginally.
Chris also points out that there are other “green” benefits to SaaS, inlcuding:
- Remote IT support. Whether or not your IT support is in-house, they’re going to consume energy traveling to and from an office to perform maintenance and fix problems. Since there really isn’t much of anything to maintain at the office, SaaS vendors are able to provide remote IT support, reducing travel and CO2 emissions.
- Less frequent replacement of PCs. Given that SaaS applications just require a web browser on the client machine, you really don’t need a very powerful PC. SaaS customers can keep their old machines in place or get a longer life from any new machines they buy. This compares to on-premises software, where customer will often upgrade hardware to support the computing resource requirements of new client software.
- Telecommuting. Accessing on-premise software remotely is typically slower and more technologically complex than a SaaS application. With SaaS applications accessible from any computer with an Internet connection, employees can work remotely, saving fuel and energy costs in the process.
I really enjoyed the post by Chris, and love the fact that he used actual data to make his points. And like every good post, it brings up some questions:
- At what point do the cost savings on power consumption alone make moving to a SaaS model the smart option?
- How do these numbers compare in an enterprise infrastructure with thousands of end-users and multiple apps?
- Is there a good way to associate cost with reduction of power, reduced travel and the cost of upgrading hardware and replacing end-user machines? It seems like the SaaS model has multiple environmental benefits (which are excellent), but showing cost savings are critical in getting business stakeholders to make the jump to SaaS. Are there standard metrics out there to help show the cost savings?
Again, excellent article from Chris. Can’t wait to hear what you think.
Leave a reply
Sponsored By Aprigo
