How Big US Firms Use Open Source Software
Diomidis Spinellis writes "We hear a lot about the adoption of open source software, but when I was asked to provide hard evidence there was little I could find. In a recent article we tried to fill this gap by examining the type of software the U.S. Fortune 1000 companies use in their web-facing operations. Our study shows that the adoption of OSS in large U.S. companies is significant and is increasing over time through a low-churn transition, advancing from applications to platforms, and influenced by network effects. The adoption is likelier in larger organizations and is associated with IT and knowledge-intensive work, operating efficiencies, and less productive employees. Yet, the results were not what I was expecting."
Part of their results are based on what they host their company websites on. I don't know about the top 1000. But when I worked at an ISP, several large clients that colo-ed several racks of equipment from us hosted their website on our hosted servers. If a company website doesn't do anything interactive besides send an email to someone in sales or marketing then thats probably what said company does.
Also, its really more interesting what the internal systems in a corporation are running, not the company website, which is usually not handled by IT.
--- Justin Dearing http://www.justaprogrammer.net/ We're just programmers.
Most companies are afraid to derive products from projects with GPL license, in fear that they will have to share all their code (even unrelated) with customers, and that exact obligation from license is unclear, and might change in court.
Now, article seems to be more about using SW tools developed with GPL license; not developing their own products from GPL components. That is lesser issue.
lke svn, twiki, mysql, memcached, boost, postgreql, hudson -and we're mostly a windows shop -although we are trying to move some of our heavier workloads to Linux.
-I'm just sayin'
I think this is the best way to convert people to OSS. Start slowly by showing them how can run free, non-Microsoft, non-Apple software. Then after a year or two, transition them to Linux.
And even if they never abandon Windows or MacOS, they're still hurting those corporations by not buying their major products, and using free altenatives instead.
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
We don't often use open source products directly, instead we use tools supported by 3rd parties that are built on them...
For example:
Firewalls are based on BSD but since BSD licensing allows it they are closed systems forked form BSD a long time ago.
Our firewall management platform runs on Linux and contains many open source packages, you even have the option of running the management tool on your own linux but we don't, we purchased a rack-able appliance that is maintained as a whole. We get "releases" that update the whole app, server services and kernel as a working supported package..
Our ANTI spam package runs on linux and is based on spam assassin at the lowest level, however again, we purchase a racked supported appliance that gets frequent updates so we don't waste time trying to piece together all the little things.
Hell even our desk phones run linux under neath but do I care? no I wan a phone that just works, so I never touch the open source part..
If you are doing a survey on open source and you are looking at desktop apps and web-servers in an Enterprise, you are missing the open source software right under your nose.
EA David Gardner -"... but the consumers have proven that actually what they want is fun."
I think your frosty machine is broken. It's spewing diarrhea all over the floor, instead of delicious icy goodness.
For large sets, this will be our guide even unto death, for the LORD will work for each type of data it is applied to...
This part surprised me, til I read this:
"Open source software is often less polished than its proprietary alternatives; version proliferation and poor usability are two often-reported problems [Nichols and Twidale, 2003,Krishnamurthy, 2005,Viorres et al., 2007]. Highly-paid employees, like knowledge workers, may argue that the fit of the OSS [Thompson et al., 1991], the service quality it offers [DeLone and McLean, 2003], or the perceived behavioral control they have over it [Ajzen, 1991] is worse than that of its proprietary alternative. The key factors for resisting such change can be classified into people-oriented, system-oriented, and interaction theories [Jiang et al., 2000]. As the cost of the software used by highly productive workers forms a small percentage of their total employment cost and the software's quality reflects a lot on their productivity, spending on industry-standard proprietary software may be a rational decision. Consequently, we could expect that the relative advantage of OSS viewed as an innovation [Moore and Benbasat, 1991,Rogers, 2003] will be marginal. As an example, traders with seven figure incomes are unlikely to skimp on the operating system running on their PCs.
--> "Conversely, in Fortune 1000 companies with numerous but less productive employees adoption of cheaper though less polished OSS can offer significant cost advantages, and therefore management can easier mandate its use. For instance, we can easily imagine the cost savings associated with thousands of service desks running Linux and the Thunderbird mail client."
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
Can you really discuss the current offerings of software based on studies or surveys done from 1991 - 2007?
Pretty much any Java project these days probably uses Spring and Hibernate or iBatis and a pile of other libraries.
This study seems to be asking what browsers are being used by corporations, and what servers are providing content. This really doesn't even scratch the surface of how corporations are using OSS.
The biggest obstacle to using OSS in seriously large corporations is support.
I'm a lead architect on a seriously big project for a global scale logistics company right now. We use CouchDB, HudsonCI, Membase, EhCache and some other OSS stuff. Its used as core infrastructure on a global deployment to over 100 countries, supporting the entire platform, and handling in the tens-of-millions of messages per day - and I can say from bitter experience that while most managers will buy into the use of this stuff, unless they can buy 24/7 Gold Support with a 4 Hour response, it makes not a shred of difference how good this stuff is technically.
It all boils down to CYA - and no support means no CYA.....
Wibble-Wobble, Wibble-Wobble, jelly on a plate
How is this surprising? Highly paid knowledge workers are under heavy demand to perform. Their entire job can be measured in hours of "non-productive" time, which would include learning a new workflow process (Linux desktop and software) so they can do the job that they are already doing quite well. If some tool written under windows is the tool they need, then they get that tool on the latest version. Not giving it to them means the inflow of money stops.
Same can't be said for most of the other staff. Linux is fantastic for call centers, admin assistants, and the like. Most of their work can be done through a browser. A desktop admin can remotely lock down their machines, choose which options of which applications to enable remotely, push and roll back software updates- all with free tools. It cuts down on the amount of staff needed to keep those parts of a company running without much pushback. No more need for licensed backup, anti-virus, update, software push, or upgrade software, and all of the admin headaches that come along with that.
Most decision makers in DN, DA, DAF run to the hand-holders [AKA: Proprietary Software Marketeers (PSM)], because DD components are dominated by highly certified and very poorly qualified technology-project managers. One project a colleague brought in to the company was promoted and supported by the government worker-bees and pack-mules, but after 12+ months of unsuccessful requesting/pleading for ".mil" domain dev/test and production environments ... The PSM came in to save the day and help produce a monolithic silo CMS/BMS/BWf/IDE... system to rule them all.
The information (data/content) will never be "FREE" from vendor formats, hooks, ownership ....
Unaccountable leaders are masters, and unrepresented people are slaves. How do US and EU fare?
Most trading companies have huge numbers of Linux servers feeding data to high end trader desktops. I suspect that many of the quants have both Linux systems for work, and Windows for bureaucracy.(email, expenses, etc.). Regular institutional traders use tons of open source software, but likely don't realize it. Sure, the desktop OS is likely Windows, the office suite is likely MS Office, but the browser is Chrome/Firefox. There is also all the stuff on the back end. If the database isn't open source, it is still more likely to be running on Linux than Windows (Db2 and Oracle), with the exception of MSSQL. Web servers, internal and external are more likely to be Nginex or Apache than IIS or other proprietary offerings. Commercial application servers have a small market share compared to open source ones. I can go on, but open source software is already heavily adopted by most large corporations. You won't see it on the desktop extensively, but it is there.
ROFL. Should read "less well-paid." The legions of workers that actually make products that people use are somehow "less productive" than the traders with "seven figured incomes." Really.
The company that I work for is thinking of moving off of MS Office because their license is expiring soon and moving towards LibreOffice and Thunderbird. I have been doing the testing for our company trying to work all the "bugs" out and seeing how difficult the migration will be. Now, the company I work for is not very big, just under 50 employees, but it is still a company none the less.
As an example, traders with seven figure incomes are unlikely to skimp on the operating system running on their PCs.
Traders with seven figure incomes don't select the OS and apps running on their desktop. That's what the IT department is for. And when selecting a system for their seven figure income trader, cost is much less of an issue than for the call center employee. Just buy them the newest, shiniest rig at the PC shop. And if it quits, just buy a brand new one and scrap last year's model. At call centers, someone is counting every nickel.
One more point: Some of those citations are pretty old (1991). Things have changed since then. Principally, the visibility of open source. In 1991, the face of OSS was RMS. Now its Google.
Have gnu, will travel.
My small company runs a webserver on Ubuntu, we have an internet facing exchange server, and an SSL VPN appliance.
Internally we have 30 or so windows/mac servers.
Externally it appears that we are using lots of open source software, but internally we are using practically none. Still, 10 years ago we were using no open source software, so OSS use is growing - at least at my company.
-ted
Not so long ago. Say four years ago, I was asked to come up with four open source alternatives to the components in Hummingbird Exceed for the Engineering group of a F500 company. Spinning ahead to the future the result was XMing, Putty /ssh, wc3270, WinSCP, Filezilla. The installers were packaged as a 32 bit chained installer. Each tool was configured for the environment. The install just over a minute, The uninstall completed in 37 seconds with no reboot.
So, to answer the questions about 'proof' I can speak of several apps that took the place of one very expensive app.
Also, they were accepted (blessed by Enterprise Applications Council) on the enterprise level.
Sometimes I wish there were less bickering on SD, and more group work towards implementing better solutions.
False. First I was asking directly about the reports and why they are cited.
Second, open source has a huge advantage of making people more efficient. Lets take a frim I have been consulting with as an example:
First meeting of the day was about licensing. We had 10 people in a room to discuss what licenses are in use, which ones are going away, and what we need to plan to spend next year. Many other people spent the last week gathering information, creating charts, and writing reports for this meeting. Lost productivity: about 120 hours * 15 people. Next, we worked on trying to mitigate upgrades that two vendors are requiring, leaving them with an unusable system, 4 people assigned full time for the last year. Add in the fact that the closed source vendor has a bug in their software, and our million dollar support contract doesnt cover vendor bugs if they are to fixed in some upcoming version, the ticket is closed and we can suck it. They cant go anywhere else easily, they are locked in nice and tight, the data cant get out and they have convinced management that training is always more costly then change. Next we review how two other offices have reduced thier support labor from 20 people / 200 desks down to 2 per 500 desks using Linux thin clients and open source apps. The users are more productive as the apps are tailored to their workflow, not some clump of apps slapped onto Windows like their counterparts.
So closed source apps and proprietary data formats are the big labor wasters.
In hardware development where I work (Fortune 500 company) Linux is predominating OS. MS server installations also are there but that is only a small small fraction of that. Solaris comes distant third. One sys-admin can manage tens if not hundreds of Linux servers. Linux OS is stable, so that problems that come are either networking or server hardware, such as HP's power supplies. We're replacing HP server-blades with Cisco blades, none of that server farm is Dell, some IBMs, yes. So, Cisco is a big winner in what is going on in server-blades market. I can say that much. The other winner is RedHat.
...to spend all that money trying to get something for nothing... There's a German guy out there somewhere who should own the rights to that report since he obviously loaned you the money to live on while you were playing scientist, and Zeus knows you never planned to pay him back in the first place!
We all know that is what it is like in real life. The MBA/PHP view of the world is not tainted by contact with reality
Sent from my ASR33 using ASCII
You have a point here. And you haven't mentioned the huge cost associated with procurement processes for proprietary software, especially in the public sector. These can drag on for months. In contrast, acquiring an open-source product is often simply a matter of a one-click download. Even if the organization's legal has trouble understanding open source licenses, this is a hurdle you have to overcome just once.
traders with seven figure incomes are unlikely to skimp on the operating system running on their PCs.
This is a particularly silly remark. First, as someone's already pointed out, it's the IT department that specifies, installs, and configures the PCs and their operating systems, certainly not an individual trader, who has better things to do with his time even if he did care about such things. The doctrine of comparative advantage, and all that.
More importantly, though, the operating system running on a trader's PC is only the tip of a vast infrastructure. Sure, he may be running Windows, using Excel for pricing, and maybe one or more proprietary Java-based trading system GUIs. But the bulk of the work will be done on hundreds or maybe thousands of back-end servers, almost all of which are pretty much guaranteed to be running Linux, with software developed using gcc, Eclipse, emacs, and all the usual OSS suspects. (Even the GUI itself might be on a Linux box, with only an X server or VNC or equivalent running on the Windows PC.)
Finally, note the specious attempt to suggest a positive correlation between the cost of software and its quality. Anyone reading this site should be able to come up with endless counterexamples.
Will someone please tell me the Difference between OSS and F/LOSS in the business community ???
The rules of academic publishing are that you have to cite relevant related work. This includes both fresh results and old classics. Where possible, we tried to cite the most recent studies. Some studies that are appear dated indicate a research opportunity to update the corresponding area. Also, it would be wrong to dismiss a paper because of its age. Some of the older studies we cite present theoretical frameworks of enduring value and importance, demonstrated by the thousands of citations they have received over the years. For instance, the 2003 study by Venkatesh and his colleagues on the user acceptance of information technology, which we cite, has received almost five thousand citations. It would be wrong to ignore it, just because of its age.
So the claim here is that a high paid worker is a very productive one?
Why is there no citation for that little gem?
perhaps gestational or nascent might be a better term than abysmal. When did the first VB come out? I remember using QB4.5 and Basic PDS in 1990/1991 right around the DOS/Windows transition...
ANY changes are a problem in this respect. It's not just limited to Linux. Merely upgrading to the newest version of Windows or Office will pose the same challenges. It's been this way since the early 90s.
Dealing with the "market leader products" is not free. Upgrading and maintaining them is not free. Never has been.
A Pirate and a Puritan look the same on a balance sheet.
that there is no training. There IS training for FOSS products available, but you may have to actually look for it ... of course, the MS reps and shills deny it.
Redpill Linpro here makes money giving FOSS courses in several cities. The ad supplements to the IT journals carry course schedules. And no, the courses are not free, similar price per day as other computer courses.
Google much?
OSS - Open Source Software
F(L)OSS - Free (Libre) Open Source Software
Businesses don't care much about software freedom in the sense of the FSF, but they do enjoy using open source projects or languages, provided that they increase workflow and make things more productive.
The difference can be found at fsf.org, in te "words to avoid" section, I believe.
Basically don't worry about it too much, since most businesses still have a stigma against OSS and FOSS products, even if they are unfounded.
"You can trash MS all you want but their development IDE and related services are pretty good compared to the others", cavreader
As compared to what, I would have thought the EclipseSDK was well up to the job?
AccountKiller