Sunday, May 27, 2007

Free and Open Source Software

In my last post: It's Not About Free Software, It's About Control and TCO, I made an argument that a model based on software given away for free with all of its source code was not a sustainable model. In the post, I used the term FOSS and defined it incorrectly. My thanks to gnuosphere for pointing this out in a civil manner befitting an education discussion. Gnuosphere also provided this link to the GNU definition of FOSS.

So going back to the original post and altering the definition of FOSS, I think that the model of simply giving away open source software for free is not sustainable, particularly in situations where frequent and disparate updates are required, a situation very common in education. The post was simply a statement that the real benefits are the open nature, the control that exists because of that openness and the lower total cost of ownership that can return real dollars to the instructional budgets of schools.

Watch this video as an example of an Italian organization of schools that replaced Windows and M$FT Office with Linux and OpenOffice to realize a lower TCO with free OSS and lower maintenance fees. Sounds a lot like the model I proposed.

Wednesday, May 23, 2007

It's Not About Free Software, It's About Control and TCO





Often times when discussing open source software, you hear the term FOSS, which stands for Free Open Source Software. To me personally, FOSS is very idealistic and somewhat unrealistic. In case most people have forgotten, we live in a capitalistic society that rewards those who create wealth. Ultimately, we must provide necessities for ourselves such as food, clothing and shelter. Writing free software and then giving it away does not lend itself to a supportable model. Even considering the power of community and collaboration, it is likely that those two factors are not enough to sustain a long term model producing FOSS.

So where does that leave us? Paying customers of course that enable a philanthropic software model to exist over time with sustainable results. That is not a bad thing and the reality is that is the way successful open source software works now. In my lowly opinion, OSS is really about control and total cost of ownership, not free software. Let's look at these two items separately.

We will start with TCO because it is very easy. For the record, I will say that M$FT produces very good products and for those education organizations interested in an integrated environment, M$SFT fits the bill for them. The caveat is that M$FT is in business to sell software licenses and over time, a M$FT district will be forced to upgrade products and pay additional costs for those upgrades. Often times, this means a requisite hardware upgrade as well. Anybody say Vista? This can become very costly and in large districts the sums of money spent can be staggering. Take Broward County Schools that two years ago tried to get approval to spend $68M on Apple Computers. That is a chunk of change to say the least.

Or take the FLDOE's Sunshine Connection project where M$FT gave them the software for free and some services to implement as well. I think that the total value could be around $20 million supposedly. Did M$FT really give away $20M or did they secure a long term annuity stream of revenue? I recently read in a white paper somewhere that Sharepoint 2007 has certain functionalities that require Office 2007. In other words, to enjoy the full potential of Sharepoint 2007 you must acquire Office 2007. So let's do some simple math. I am going to assume that the state of Florida's education system has about 1 million computers. Let's say at some point in time, that the FL DOE forces everyone to do at least one thing in Sunshine Connections that requires Office 2007. Assuming a per desktop license fee of $25 for Office 2007, that comes out to $25 million and it seems that over the long haul, M$FT is really giving Sunshine Connections away for only $5 million. And as the years go by and more upgrades are required and profits rise, you begin to realize that M$FT simply made an investment in FL to drive an incredible revenue stream over time.

OSS presents a situation where schools and districts can begin to take control and manage TCO to their benefit. Let's consider Linux as a start and specifically, let's look at thin client technology. For our example, we are going to consider Brandon Elementary School in Atlanta, GA. Brandon had approximately 1-2 computers per classroom running Windows XP and was experiencing a myriad of problems with those computers. Because it was an elementary school, Brandon did not have the proper technical resources and depended upon two parents for maintenance of their computers. These poor souls were spending a great deal of their time fixing these computers and decided to look for an alternative. They found K12LTSP and decided to give it a try. Using equipment donated from local businesses, they built an environment of 35 K12LTSP servers that are now managing approximately 250 Linux thin clients. They have raised the average number of computers in the classroom from 2 to 6 and have 8 or 9 in some classrooms. The uptime has been amazing and they are using thin clients running on 350MHz processors. Businesses are throwing away computers with that kind of processing power and would willingly give them away to schools. You can read the entire article here. Bottom line, Brandon estimates that they have saved about 90% costs per computer. That is a lot of money. Here is the direct quote from the article:

"Having Linux on the desktop has been a shot in the arm for the school in many ways. Howard and Fragakis estimate that installing used hardware with a cost-free operating system saved Brandon Elementary 90% of what it would have cost to install Windows XP PCs, and that doesn't include savings on operational and support costs. But the biggest benefit can't be counted in dollars."

As the years go on and they don't spend money on new hardware or new software, the cost savings begin to skyrocket and you can see that the TCO picture gets prettier over time. A side benefit: Brandon's test scores have risen and were the highest in the district and third in the State of Georgia.

What about control? Let's look at student information systems and small school districts. For this case, we will examine PowerSchool. PowerSchool is a SIS that was developed by Apple and later sold to Pearson. Pearson owns many SIS offerings and currently has greater than 50% market share for this particular application. In the small school districts, it is the dominant player by far. How much does PowerSchool cost? Here is a recent offer: http://www.pearsonschoolsystems.com/email/files/PremierBundle030708.htm

For the low one time licensing fee of $28 per student, you can acquire PowerSchool for your district of 1,000 - 3,000 students. These fees include implementation services, which I am guessing is installing the software and making sure it is working correctly. I can pretty accurately guess that fee does include services like data conversion from your old SIS. You can have those services performed for you at what appears to be a fee of $225 based on the Apple price list for Arkansas from 2005. So I have 3,000 kids, which means I pay $84,000 to acquire the software, have it installed and make sure it is running correctly. Should I require any additional services for something like data conversion, that costs and additional $225 an hour plus travel. For our case, let's assume that we will need one resource for one month (160 hours) to successfully convert all of our data, validate that it is correct and turn off the old SIS, That now adds another $44,000 assuming $2,000 per week travel costs. Now I am up to $128,000 total so far. At the end of year 1, I have to pay maintenance. For our example, we are going to assume that our district has purchased the Premier bundle, which includes a SIF agent for PowerSchool. According to the AK price list, maintenance comes to about $8 per student adding another $24,000. I am also going to assume that the three days of included training is not enough and add one more week. According to the AK price list, that will cost me $1,595 per day X 5 days = $7,975. My grand total is now $159,975.

What if I went with an open source SIS? There are multiple efforts underway. Centre, SchoolTool and OpenAdmin are examples. Of all the ones out there, I would guess that Centre is the most mature and has the most traction. Since it is open source, I immediately eliminate my one time licensing fee of $84,000. Let's assume that maintenance and support are equal. I checked Centre and they sell state reporting modules for $2,500. I don't know what they charge for data conversion of customization services, but am going to assume a flat rate of $100 plus travel. Using our earlier metric of 160 hours that adds $24,000 to the total. For training, I will assume the same $100 X 8 hours per day X 5 days = $4,000. Our grand total now comes to $54,500 representing a cost savings of $105,475 or 66%.

Over time, I would invest some of that savings into training resources in the open source technologies so that I could make my own changes to the source code. I don't get PowerSchool's source code. As an example, my state changes its reporting requirements and I have to add a field of change a field. I can have my resources do that instead of paying Pearson $225 an hour to make those changes. Over time, since I have access to the source code and because I have invested in my resources making them proficient in these technologies, I might choose to give up maintenance and support further saving myself $24,000 a year. And because I don't use M$FT products, I am not forced to upgrade my operating system and hardware saving large amounts of money.

While I think OSS has a ways to go, a 66% savings argument is very compelling for a small school district with very limited funds. While I talked about specifically about thin clients and SIS, there are a multitude of products out there that can run much a school district's major functions. In addition to the SIS, there is:

  • Moodle for Learning Management and On-line Courses
  • OpenBiblio to manage your library
  • A variety of open source portals to manage your web site (Joomla, DotNet Nuke, Websphere Community Edition from IBM)
  • Mambo for content management
  • OpenOffice or StarOffice
  • Edubuntu
The list could go on for quite some time, but you get the idea. Are you ready for lower TCO and better control of your IT environment? Time to investigate OSS.



Friday, May 11, 2007

Open Solutions in Education? - Let's Do the Math!

I recently was told by a friend that works at a district that they have been looking at options to replace their homegrown SIS. They are a small district and so PowerSchool was a natural choice. Pearson, the new owners of PowerSchool, quoted them $28 per student just for the software and data templates. That didn't include any help converting their existing data, modifying business processes or working on change management. In fact, Pearson quoted $225 per hour to perform these types of services. I am not sure about your small district, but my friend's district could hardly afford to pay a vendor $225 an hour plus travel expenses.

It occurs to me, that districts, especially smaller ones, should be considering the many open source options that are out there today and being used. Do you need a student information system or data warehouse? Try Centre (http://www.miller-group.com). Do you need a learning management system that can deliver courses on-line? Try Moodle. (http://www.moodle.org) Going Linux? Try the latest distribution of Edubuntu, based on the new Feisty Fawn release of Ubuntu. It comes pre-loaded with education specific software (http://www.edubuntu.org). How about a libary solution? Try OpenBiblio. (http://obiblio.sourceforge.net) How about Linux thin clients? Try K12LTSP (http://www.k12ltsp.org) You can use old PC's (remember those PC's businesses try to give you? Hint, Hint.) or buy brand new pre-configured clients for about $200.

The fact is, that for those districts that are willing to be pioneers and accept a few occasional bugs or forego some functionality for a period of time, open source solutions are coming into their own and maturing to a point where districts should consider using them in critical situations.

Many people always say get open source because it is free. That is not true and it is misleading. Open source must be supported and there are costs associated with that support. Open source often needs customization and there are costs associated with that effort. (Usually not $225 per hour). But over time, the total cost of ownership (TCO) tends to be far lower than that of a commercial alternative.

There are also the associated costs that can go away as well. Going open source can remove these licensing costs: (these are estimates only)

Windows Operating System - $50
Microsoft Office - $50 (use OpenOffice)
Windows Server - $2500 per server
Exchange Email - no idea but will guess $2,500 (use Zimbra. Comcast thinks Zimbra is good enough for its 12 million clients)
Microsoft SQL Server - $2,000 per server (use MySQL or PostgreSQL)
Many more could be listed.....

The real difference in open source is the word open. You get the source code and when you train people with the skills to modify the code, you have the flexibility that a commercial vendor takes away from you. State department changes a data collection requirement and you have PowerSchool? Guess what? $225 an hour to make the customization to fulfill the new requirement. You have an open source SIS? Only the cost of your IT guy and a fix in place much faster in most cases.

Open also means that your solutions can run on older HW, which can be a huge cost savings. I asked a friend in a large district how many PC's they had and he said about 35,000. I asked him if he had ever heard of OpenOffice. He said no, so I showed it to him and he was impressed that a free product had that many features. Let's do the math: 35,000 X $50 (M$FT Office license fee) = $1,750,000. So let's see at an average salary of $45,000 plus benefits, which probably comes to around $60,000, that buys me a teacher that is completely paid for for the next 20+ years assuming some raises along the way. By the way, that district must pay licensing every year for that M$FT software, so every year I add a new teacher that is paid for for the next 20+ years.

Also, by being able to use older HW, I can avoid the HW refresh cycles that are driven by new commercial products. Can you run Vista on your current desktop or laptop? Ummm, probably not and that is why HW vedors love M$FT. Imagine avoiding one HW refresh cycle for 10,000 computers like my friend just did. 10,000 X 1,000 = $10,000,000, which buys me 166 teachers for 8 years.

I can go on and on with numerous examples like this, but you should have the idea by now. Perhaps open source is worth a look.