Slashdot Mirror
Open Source Adoption by Corporations?
shakuni asks: "I work for a large network equipment vendor in the operations software business unit. One of the questions that I have been asking all my customers (large telecom service providers) is their position on adoption of open source software in their operations environment. The customers that I have interviewed don't comprise a a large enough sample to make sweeping statements. However, most large service providers (who have probably more than 80% of $1 trillion telecom market worldwide) seem very wary of open source, even though the high cost commercial software is hitting on them hard. How is open source adoption being encouraged amongst the financial and telecom behemoths, who are averse to taking risks with their IT systems? Are there specific organizations out there that actively address the IT manager and CTOs concerns about open source software? In other words, is there an enterprise Open Source initiative that pro-actively helps companies move in this direction?"
Bill Gates, with his Windows operating systems, is the best salesman for open-source software that there is.
Windows is to worms/trojans like a bug-zapper is to bugs (if the bug zapper has a broken zapper and can't kill them: just attract them).
Hi. I'm Troy McClure. You might remember me from such pro-OSS propaganda films as "Open Source, Open for Business", "Boardroom Penguins IV: Electric Boogaloo", "Explorer Madness" and "The Triumph of the Gnu"
I am sure a lot of these big companies weight the money saved by OS solutions against the additional cost in things such as air filtration, hygenic costs, loss of employees due to new stank OS developers, the high cost of snacks (fritos, jolt, or whatever the agerage OS developer constantly stuffs in his snackhole) and many other similar things. In the end, the TCO of OS software is a lot higher. Sadly, this is completely due to the upkeep of the developers and admins, not the software.
Those large companies are in many ways in the best position to take advantage of open source software. They can afford to spend the money on in-house programmers to do support, add features and fix bugs in open source software instead of paying Mircosoft and others large amounts of money for bad support, useless features and more bugs.
The risks are much higher when buying closed source software because you never know if the company you are buying from will add the features you want, fix your bugs or even stay in business. Open source software allows you to be in full control.
Why are they leary of OSS?
Without knowing this, the discussion seems somewhat pointless.
At least at the large financial corp. I work for what matters first (even before functionality sometimes!) is the contract, i.e. who has liability for what, followed by cost and vendor reputation.
I'm not talking desktop OS here as obviously most niche vendors with desktop products for $FinanceBusinessFunction require the desktop to run Windows.
But with regard to encouraging Open Source in server, backoffice, data center, mid range, etc etc etc, the decision makers don't really care if the code can be looked at by anyone, as long as it's as secure as possible and the contract is drawn up in such a way that the vendor shoulders as much liability as the collective lawyers can agree on.
So from where I sit the question of encouraging Open Source is sort of like asking what's being done to encourage more yellow and red colors on the company intranet... Who cares as long as the job is getting done and the price and $Lawyer-Stuff is right.
That's not my answer but that is the Corporate Answer.
Operator, give me the number for 911!
So I would change your question a little. What groups are out there discovering companies' needs, then communicating those needs to OSS developers? We shouldn't fall into the trap so many companies do: writing software for the developer and not the customer. This is especially dangerous for OSS developers, because their own needs are usually the reason the software gets written in the first place.
"Anybody But Bush 2004 [theatlantic.com]"
Nice article, but Soros lies several times per paragraph. Soros is the best argument there is against people having political influence just because they are rich.
I can't speak for telecoms.
But what I've seen is open source deployment at the grass-roots level. Acceptance is a gradual, building thing with exposure working its way slowly upon the organizational hierarchy.
Smaller company CIO's and smaller organizational subunits in large corporations are willing to take gradually increasing potential risks by utilizing open source.
The irony is stealth deployment cuts both ways.
One of the reasons it's easier to take that risk with open source is that deployment doesn't require visible commitment of dollars. That Samba or Apache server just cranks away, no invoices come in, no need to count licenses to be compliant, etc. And it sure doesn't hurt that many open source applications are as reliable as death and taxes; they don't drop service causing the CIO to fume about not being able to get service.
But by the same token, those open source deployments are largely invisible to people higher up; those people are less familiar with the successes and failures and are therefore not yet ready to jump in the water headfirst.
"Provided by the management for your protection."
However, nearly every developer here uses open source tools daily. JUnit and Ant are everywhere, as well as NUnit, NDoc, and NAnt for the .NET folks. Eclipse is gaining ground, and Emacs use is pretty common.
For one project, the dev team created a post-project list of all the software used during development. Out of about 30 programs (including DB, OS, etc), 2/3rds were open source.
I'm looking for a HEPA media filter for my TV. I'm alergic to reality shows.
It was simply common sense. We had three independant firms calculate the cost out over the next 5 years. It basically panned out like this:
Microsoft: 15 million
Sun: 10 million
Redhat ES 3.0 on DL380's: 2 million
We had briefly toyed with the idea (seriously though) of using Debian instead of Red Hat, but some of our proprietary hardware was only supported by Red Hat. I can safely say that the level of technical acumen and common sense here made SCO's hilarious blatherings have exactly zero impact in any of our decision making.
For every annoying gentoo user, are three even more annoying anti-gentoo crybabies. Take Yosh from #Gimp for example.
Don't underestimate the power of a perceived dependability of cost. This is a HUGE motivator. There are corporate people who LIKE the whole software as a service/new Microsoft licenses kind of thing, because they perceive that under such a regime their costs will be predictable going forward.
That's why the focus of the SCO debate on end-user costs is so insidious. The perception of an uncertain outcome threatens the comfort level of those who think they have all of their foreseeable upcoming future corporate costs all mapped out.
I think what you're looking for is Red Hat.
They may not be the owners of the software, but last I remember they'll enter consultancy contracts with anyone and they'll make it better if something's wrong.
As a network administrator for a medium-sized medical office, I have recently been migrating users of Microsoft Office 2000 to OpenOffice.org 1.1.0, and MSSQL to MySQL. So far adoption of OpenOffice.org has been going better than I had expected.
Remember your history. Remember how long it look for Windows to get accepted by the large corporations. Heck, even the desktop PC was a "courageous" decision by some one.
Windows "trickled" up from the home into the small business, and from there to the medium and large corporations. Don't worry about the big buys.
Don't blame me, I didn't vote for either of them!
When discussing these kinds of questions, one needs to realize that the "average" slashdot reader differs significantly from the "average" senior corporate IT manager (yes, I realize there's a Gaussian distribution centered on both of these averages and therefore it will always be possible to find some number of exceptions to any statement based on averages).
Again, speaking in general terms, the geeks focus on the nuts and bolts of the software while the senior IT people (and the average non-technical computing user) see software as a *means to an end* not an end in and of itself.
In this context, what's important is Total Cost of Ownership and user productivity. There have been several studies that categorize the components of TCO and the majority of them agree on the following:
* The single largest component of TCO (>60%) is day-to-day administration, maintenance, etc.
* The second largest component (~25%) is the cost of downtime.
* Software and hardware purchase cost *combined* are 10% of TCO.
What are the implications of this data?
1. Unless you're a Dick Stallman on a personal jihad to destroy the Evil Empire in Redmond et al, free software *in and of itself* isn't going to have a significant impact on the market, the world, etc. (with one possible exception discussed below). Yes, the geeks will love it, and more power to 'em, but geeks are a tiny percentage of all software buyers and users.
2. The Big Bang in software will come from designing software systems that minimize administration overhead, minimize downtime, and maximize productivity (anyone who doesn't realize that there are significant advances yet to be made on these fronts is on serious crack). These terms in the value functions of software buyers and users far outweigh the terms for either purchase cost or making some religious/altruistic statement about "freedom".
Now for the exception to point 1: One of the larger driving factors behind the presence of OSS in corporate America (and the world) is that OSS is being used by some corporations to put pressure on the single most dominant player in software (billg and company). Commoditizing the OS and the two highest revenue-generating apps (word processing and spreadsheets), causes the most damage to the company with the largest market share. If you're a competitor to billg you like this. If you're a buyer of software you like it too because it increases competition. In other words, no small percentage of the people/corporations supporting OSS are doing it not because they believe in freedom but because it is to their economic advantage (in turn not because of zero-purchase-cost but because it impairs a monopolist).
Note that I'm not faulting them for doing this, it's a rational thing to do even if they're not doing it for the betterment of all mankind.
I like that open source exists because it provides a mechanism for anyone to learn and later improve the inner workings of important software systems that otherwise would not be accessible. If the OSS movement does nothing more than this I deem it a success.
Unlike some of the OSS jihadis, rather than talk about what I want to destroy, I'd much rather talk about what I want to create.
I work at a fortune 150ish company and it's been interesting
to see how the adoption of OSS has progressed in the years
I've been employed.
Depending on the department, running Linux or BSD on your
personal computer was tolerated to various degrees (for a
while I had to dual boot in order to use MS Project for
staff meetings, but that requirement was quietly dropped
during a reorg). Recently, however, there has been growing
infrastructure support for running non-Windows operating
systems on desktops and a slow migration of development
desktops over to Linux and BSD.
In production, OSS tools were tolerated at first as long as
the platform was either Solaris or Windows. Then servers
running Linux began popping up beside the Solaris servers.
Eventually the standards setting people admitted that it
was OK to use linux for some purposes, and many new projects
are developed and deployed on Linux. Even some old projects
are being quietly ported (mostly from Solaris, but occasionally
from Windows).
Most of the resistance to Linux and BSD seems to be coming from
the support groups since they're used to Solaris and Windows and
don't seem very interested in adapting. I think this will change
over time, but it's hard to predict.
*sigh* back to work...
Well first problem could be the system administrators.
Generally system administrator don't like dealing with different types of platforms, Linux mixed with Unix mixed with Windows Mixed with novel. Having a the mixture creates a lot of problems and switching cold turkey to OSS is risky and irresponsible because it means a major problems in dealing with business until things are completed. 12 buisness hours of downtime could mean a million of dollars in loss sales. compared to play 3,000 a year for a license.
Next you get the subadiquate sysadmins who's primary job is to keep their job, which means keep busy and seem compantant. Switching to a different platform will humble them and make them a newbie again. Thus making him seem more expendable, also if everything is working smoothly then he is no-longer busy. So his job is in serious threat. So no need to switch there.
Next is company policy, a policy could be 10 or so years old dealing with IT or the same thing with a number change in the version. So if your unit breaks the policy then you need to get an other unit to talk to you and you mention that your not using the policy then they will just reject your request. Changing the policy means you will need to get all units to agree with the change and that can include the subadiquate sysadmins.
The PHB. Well they are not always dumb people but in most cases the general rule is you get what you pay for. Some OSS breaks like Linux that rule but not all of it. So you try to sell them Free they will think well it can't be that good if they are giving it away.
What do most people do. As americans we grew up in democratic ideas. That what most people want is generally the best solution. (while I disagree with it) A lot of people don't and they find is as a good safety zone for them because they can say well I am not the only one.
Fear of personal dependence. A lot of these companies got burned in the 80's and 90's where the developers of some of the company code decided to quite or ask for a huge pay increase. Because only them can take care of the company. And all of its Fortran 77 code. Now your proposing to install an extramly flexible programs in the company. Now it will seem to the manger that this is a great way to configure the system so only you can administer it thus making you irreplaceable, thus able to hold the company for hostage when you want better pay.
What about application X, most companies are running some application that may no longer work on and OSS platform. This application could be custom written many years ago and no plans to remake it because it would be expensive and retraining people would be horrible.
Talking to other companies. Well you want to be able to talk to other companies and share information with their systems as well you need to know that you can do that.
So there are a lot of major problems. Yes I am sure that there are OSS solutions to the problems but switching is not as easy as back up the server data. Format the drives as ext3 and install linux on top of it. It will take a lot of time and work to get everything going. Players like IBM and Novell who are pushing Linux and OSS will help ease some of the pain. Also a lot of smaller venders are willing to help out too. But it not a technical problem but an emotional problem.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Starting in 2005 IBM will begin moving all employees' desktops/laptops to a Linux-based OS (probably RedHat). That's quite an adoption of OSS!
Define "corporations". My wife is incorporated. So is the company I work for, which currently employs over 50,000 people.
I believe the key factor is simply that it costs too much to retool in any short timeframe. My company, for example, owns something like 65,000 end-user desktop machines (actually, probably more; I doubt even the accountants have an exact count). Imagine the nightmare of trying to change operating systems while ensuring that everyone's work is reasonably uninterrupted and that everyone has some OSS-panacea -version of whatever applications their work requires. Hell, our janitors (oops -- sysadmins) can't even reliably deploy simple patches and updates without borking the job roughly half of the time.
Ok, so we backpedal and ignore the desktop for the moment, and just look at the servers. Migrating the servers is a bit more realistic, but we're still talking about roughly 6,000 servers. Even if you could get a reasonably large number of the thousands of admins to agree on what to use (and you won't), the process is simply too huge to tackle with any speed.
The ONLY way truly large companies (we are a large financial company, but we are only medium-sized when you consider the full spectrum of companies out there) will migrate is S-L-O-W-L-Y. The traditionally-touted benefits of OSS are simply irrelevant in comparison to the labor, cost, and risk of doing anything but gradual implementation.
In other words: most OSS thinking is still very small-time. Only companies like Sun and IBM seem to be making any serious attempt to change that in the near term -- and then the end-result tends to be roughly in the same price range for such large accounts, which means any potential long-term cost benefits aren't terribly relevant.
Slashdot quality declines as the number of hot grits posts decreases. - Provolt's Law, Apr-09-2005
They are already using open source, but many probably don't know it. Case in point, TCL/Expect and Perl are very popular in telecoms, both for testing and production scripting. I had written a 10k-line TCL/TK/Expect script for automating production maintenance activities back in 1999, and it ran on an 8-CPU system running Linux. Similarly, Cisco provided TCL scripts for many production maintenance activities (which we ran from the big Linux machine). However, most managers and execs only see the web-based front ends and GUIs, so they are completely clueless about what keeps the networks running.
Can You Say Linux? I Knew That You Could.