Slashdot Mirror
Driving Out Costs with Open Source Tools?
zO1inks asks: "I work for a custom web solutions company and I'm working on a white paper and presentation that outlines the merits of open source software, particularly the cost savings that can be achieved by using free (or low cost) open source software (OS's, Server, Utilities, etc.) vs typical (and more costly) Microsoft solutions. The material is targeted towards Fortune 500 companies that would have an interest in driving down the costs of custom web development, design, construction, migration, and support. I'm wondering where the largest gains are to be had and what supporting statistics show the validity of such strategies? (e.g. Bank xyz uses Linux to conduct X number of transactions per day with a downtime of only Y hours per month) Bottom Line: Is Fortune 500 ready to embrace Open Source?"
If I get an 800 number, and support your open source questions in better average time than MS et al, can you get your company to pay me big bucks?
(I know, this model hasn't really worked that well in the real world, but my overhead is lower since I have no buzzwords and am not planning an IPO.)
Then I went to work for a big-5 consulting firm, and on one of the jobs I was forced to drive an hour and a half each way to a client who was paying $175/hour for me to take that cable out of the plastic bag, attach one end to a UPS, and the other end to a server.
They refused to do that job themselves, because they had contracted with us to do it. No matter that the job was charged to them "Time and Expenses" at rates the top whores in the country would be embarrassed to charge.
That said, most Fortune 500 companies won't use open source (note how Apple, IBM etc. already do, however!). This makes them easy pickings for the likes of a Microsoft to seize control of all proprietary software and start draining money out of these Fortune 500 companies- which can happen for a _long_ time before the Fortune 500 companies are severely weakened.
In the bottom line it becomes a question of efficiency above all. Yes, huge corporations like to outsource this sort of thing, but if other huge corporations can get an advantage in margin or efficiency by taking on IT responsibility themselves, over the long run they will beat up on the corporations that are overspending on Microsoft licenses and continually wasting human resources keeping up with Microsoft's ever-increasing accounting demands. Complying with a Microsoft audit costs MONEY. For a huge corporation the human costs of labor and wasted time could be nearly as much money as the actual license fees in question- and the audits won't stop coming.
Basically, any Fortune 500 company sticking determinedly to proprietary software is supporting not only itself but is also supporting its vendor in a sort of charity. If that vendor is Microsoft, the cost is very impressive- even staggering. You have to include the accounting costs of audits, the 'runtime' of the relationship as well as the actual products themselves.
This could prove fatal for such companies in the long run if their direct competition includes more self-contained companies.
the sack of shit it is
This is inflammitory speech. This kind of language is designed to invoke an emotional response. People who know they do not have a logical argument use this kind of language. People with logical arguments do not need to result to this kind of name calling.
For the mission critical stuff [open source is] far too insecure and lacking in enterprise-level features.
Buzzword mania. Note how this poster tells us that all open-source software is insecure without backing up this claim with facts.
The facts are this:
- Sun, as an example of one of the expensive closed-sourced vendors this poster considers better than Linux, has 23 vulnerabilities reported in the year 2001.
- OpenBSD, in the same time period, has only had eight vulnerabilities reported.
Yet, we are supposed to believe that closed-source is always better than open-sourced systems.This person talks about vague "enterprise features" that open-source is supposibly missing without telling us exactly which enterprise features open-source is supposed to be missing.
In other words, this person is making a number of inflammitory emotional statements, and stating a number of opinions without backuping up those opinions with facts.
Moderators should not be moderating a post like this up.
The secret to enjoying Slashdot is to realize that it should not be taken too seriously.
The problem is, the guy stocking the store has no "real world" experience in Unix admin or production programming. Any idiot who can read can setup and install Apache/PostgreSQL/etc on a Linux box. While setting up Oracle and installing the ACS is more complex, it's still not that hard. Hell, I did it, and I'm a complete idiot.
Setting up a single machine, while somewhat impressive is not the same as building a functioning network out of a building full of oddball platforms and legacy applications -- which is very important in Fortune 500-type companies.
I'm not trying to put a damper on the guy's enthusiasm -- rather, I want to encourage him to expand his knowledge beyond simple administration of a single box. Make a trip to your local thrift store an pick up a few 486s on the cheap and integrate them into a home network. Mix and match -- don't focus on Linux only, stir in some Net/Free/OpenBSD, Windows and Mac (Mac SE/30s with a network card can be had for $30 or less, and can speak TCP/IP).
Then set up a remote access service to your home network via modem, then a VPN with a friend. Congratulations, you've just worked through the major things that Big Companies are looking for in IT workers.
Potato chips are a by-yourself food.
When the point-N-click works, it's great. It all begins to turn to pooh when one of these things happens:
Never confuse a plethora of choices with real freedom. Never confuse pretty with useful. Never confuse slick packaging with thorough testing.
Where OSS such as Linux wins is:
Let's not even bother exploring the ugliness represented by layer upon layer of legacies and idealogically incompatible subsystems in the various Microsoft products.
Got time? Spend some of it coding or testing
Unfortunately OSS engineers typically cost much more than MCSEs.
Not if you are comparing people of similar skill levels, and you should. If you hire a "dummy" to run your systems... then the problem is you have a "dummy" running your system, and you get what you pay for. If you hire a highly skilled UNIX or OSS engineer and a highly skilled MCSE, chances are the cost is not going to be that much different. If the rewards for building skills aren't there, why would anyone want to pay all that money to become an MCSE? Microsoft must be lying to someone -- they tell workers "become and MCSE and make buttloads of money" and they tell bosses "hire MCSE's, they don't make jack squat compared to real engineers".
While there may be a lot more MSCE's than UNIX or OSS engineers out there that are "dummies" because of the effort that has been made by Microsoft and their partners to push a lot of people through training classes and whatnot, I have to say "Who cares?". I don't know why you'd want that sort of person in your employ. And if you have to pay about the same amount per person, then UNIX or OSS typically wins because most people find that they need fewer people to support UNIX and OSS systems than they do Microsoft based systems because they break less often.
One of the big myths about Microsoft's stuff is that it is "easier" to administer or somehow less complex than *nix. I don't think that is true at all. It merely hides its complexity under a GUI that lulls people into a false sense of understanding. But when things break, the complexity underneath can rear its ugly head, and since so many things inside the black box are secrets people often find themselves unable to solve problems without help. *nix on the other hand puts everything right out there in the open, which can be daunting for a newbie, but by forcing people to do their homework up front, they build real understanding more quickly, and get to use and build on that knowledge faster.
If I want to get Sun's C++ compiler for SPARC Solaris, I need to get a purchase order, I need to wait for the software to arrive, I may need to wait for a CD-ROM to arrive, I certainly need to wait 24 hours for my host-specific licensing information to be processed by Sun. As the project grows, I need to ensure we have enough licenses; if we don't, I need to go through the whole deal all over again. That's a lot of my time. (I've done this on several projects.)
If I want to get g++, I download it.
I can assure you that at least two Fortune 500 companies use gcc/g++ as their production compiler for commercial software. (Only LGPL libraries are used; great care is taken to avoid GPL libraries.) Sorry, no names.
Stupid job ads, weird spam, occasional insight at
Bullshit. I work for a division of a fortune 100 company. We use lots of Open Source software including Linux.
We *also* use a lot of proprietary software, including Windows NT, AIX, Solaris, and Novell.
We use whatever we think works best for each particular situation. Some Open Source programs have compelling advantages, including price, customizability, and (for some programs, like OpenBSD) security.
Will this guys paper convince CEOs and CIOs to drop all of thier proprietary software? Hell no. But it might make help them to understand what their tech staff already knows. Open Source programs *can* sometimes be a valuable addition to any IT departments "bag of tricks".
most companies DON'T make money off of software.
you, like "real" companies that make "real" products. These kind of companies have accountants that get all hot and bothered because they can reduce the number of screws in a widget by 2, so they can save 3 cents per widget.
"closed source" is an expense. pure and simple.
one day, the bean counters are going to wake up, and say "geezus, we are now paying MS about $500/employee/year just so we our employees can send email, use MS Office and play MineSweeper(tm). On top of that, we have constant hardware upgrades, and our IT staff takes up 20% of our budget, and they are a bunch of high school educated people with a "certification" from microsoft, and no formal college training... and we're paying them on average $60,000 a year on average so they can point and click to keep our network running correctly"
For the VAST majority of businesses, IT is a cost center. It produces no revenue. In times of economic hardship, cost centers get reduced or eliminated.
The economic argument is a good one, dont underestimate it. Companies are not going to roll over and play nice forever. they bought MS's line of shit that using MS products will "make" them money. Watch. Its coming. I know of one Utility that is looking at their IT budget, thats at 20% of total expenditures and saying "wtf, we arent in the computer business, why are we spending so much money on it".
tagline
... hi bingo
however, you need 3-4 times the number of MSCE's, coupled with the fact that teh skill level and knowlege of MSCEs is so much lower than a *NIX guy.
but hey - thats what you get when you dumb down server admin to be point and clickable.
tagline
... hi bingo
Um, no.
If I want to get Sun's C++ compiler, I go to their try and buy page, and download it.
It stops working after about a month. That's more than enough time to decide whether you like it, and if you do, to go through the purchase order and whatnot.
24 hours for licensing information? It took me twelve minutes. Most of that was waiting for Timothy to get off the damn phone and let me call up.
(Note, I happen to be a regular contributor to GCC. So please don't think I'm slamming GCC. I just don't like the smell of fresh bullshit wafting from my monitor.)
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
Whenever I try to get my employer to try some open source app, the main thing they scream about is upport. If it breaks, they want someone to blame/fix it, and they are willing to pay big bucks for that.
Then contract with Cygnus Support!
That's what they DO!
(Besides upgrade many of the Gnu tools, cut distributions, and maintain archives for them - with money from the support contracts.)
Think of Cygnus as a software company that happens to put its products under the GPL. Or think of it as a service organization specializing in GPLed software tools.
And they're no pikers, either.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
I had to do a paper in school (a Unix Admin class at U. of Maryland) on this exact topic. I found out that a very good argument can be given for open source products such as Apache, Sendmail, Linux, Free BSD, etc. so long as you can get the proper TRAINED professionals to manage the software and as long as you can keep the TCO down by utilizing stable equiptment and software that is well tested. Fortune 500 companies dont really care about buzz words, they want products with support channels and proven performance. I think with companies like Red Hat and IBM supporting Open Source prioducts we will see more adoption in fortune 500 companies.
"If ignorance is bliss, why aren't there more happy people in the world?"
The MS Licensing schemes (or scams?) are their bread and butter. I can't put into words my disgust when I spec a server from Dell or Compaq and it looks really cheap until I factor in say 25 CALs and W2K Server. Pay-as-you-go? I'll be one of the first engineers to take a number for the angry complaint line if a server goes down because it's time to renew my subscription and I have many angry people wanting to know why...
...
I can see it now, the upper management will hold an emergency meeting to run the IT dept (or outsourced guys) through the wringer...
Stuffy Corp Exec: "How could this have happened? How long will it take to have the systems back on-line and functioning properly?"
Witty Network Admin: "That's just it, the systems are functioning properly. They're simply denying logons and not allowing anybody to access shares or printers. The Event Logs indicate that our licenses have expired!"
Stuffy Corp Exec #2: "Isn't it the responsibility of your department to maintain our equipment and keep us current on our licensing?"
Witty Network Admin: "It is indeed. We have kept this network running flawlessly. We made our request for licensing funds more than 2 months ago. I do have a question to pose: where in the grand scheme did my PO request for updated licensing fall?"
Idiot Numbers Guy: "It never made it past my desk. I didn't see any ROI (Return on Investment), when it came down to brass tacks. I don't see why it should have failed, we bought the software and licenses, we should be able to use it."
yea, I think we already know the way things go with open source and free software...the only money involved is for the hardware, and time to configure properly. Maintenance in a well-designed environment should be drastically less money bound. All it takes is the systems being given to the people with the right skills.
When you are talking databases, you can frankly ignore Microsoft. MS SQLServer has its features, but ultimately, you are stuck with an MS only product. Simply put, you don't want your database to be MS only. If your needs are simple, many opensrc databases can do the trick. If your needs are complex, you probably need some more support and should go with a proprietary job (Oracle, DB/2, etc). If your queries are simple, MySQL is pretty tough to beat. However, if you are doing more sofisticated transactions, working with over 1TB of data, and thinking about failover and cluster, then you are pretty much talking Oracle-time.
As for the middle-tier, services, etc, you definitly want to go open source. There just is no need to use any web server other than Apache. MS IIS makes me want to puke. Everything from Perl, to Java Servlets, to JSP will fill your middle-tier needs. Services are abundant, from SendMail to OpenLDAP. NovellDirectorService (free but not opensrc) is also not a bad F500 app. For NAS needs, definitly go with Samba. It performs just as well as win2k as a CIFS server. Where you will find trouble is with things like Exchange. If they want Exchange, you will have to go with a mixed system.
However, the real question shouldn't be opensrc code or proprietary code. Because, frankly F500 company's dont give a shit about paying for something. The real question is MS or *nix. MS has some great products, but they don't play nice with others. So for me, the answer is simple. Once you reach a certain size, you are going to have a few Linux boxes, some BSD boxes and maybe even a Solaris box. And you will run a healthy mix of opensrc and proprietary code on those systems.
Someone you trust is one of us.
The example you mentioned in parathesis sounds more like a backend thing. Backends, especially in the banking industry, tend to be so old that they predate both open source and Microsoft in the server domain. Those systems tend to be running AIX or SCO.
You're welcome to use my company, TrustCommerce, as an example. We use open source software for pretty much everything, both web stuff and backend. Our processor runs millions of transactions a month and hasn't had a single moment of downtime in over a year. All of our web and frontend stuff is using open source software as well, but there are plenty of good examples of that from companies that will be far better known than us.
"The new wave is not value-added; it's garbage-subtracted" - Esther Dyson, Dec 1994
Whenever I try to get my employer to try some open source app, the main thing they scream about is support. If it breaks, they want someone to blame/fix it, and they are willing to pay big bucks for that.
Believe it or not, not including the support issues mentioned earlier, the "open source/free is better" stance also will likely run into serious opposition from the IT/IS/product development managers within those Fortune 500 companies, because any savings generated by using open source tools would translate into their budgets being cut by the resulting savings (unless they spent it on something else). That's how the really strange world of the big company budgeting process works: if a manager spends less than his/her budget for a given year, that manager gets less money the next year, even though the manager has done what seems to be the right thing for the company. Then, by extension, that manager would lose that much political prestige/status/karma/whatever you want to call it within the company, since they then have a budget potentially smaller than their peers.
Even stranger, I've seen cases where managers new to a company demanded what was the most expensive tool for a given task, even though the free tool fit the developers' needs better. It's all about "how important is this project?"
I know it sounds bass-ackwards (at least it did to me when one of my old bosses explained it to me), but that's the way it works in a lot of places.
That's odd...
My students are quite capable with Apache, CGI, and databases, but they have incredible difficulty getting someone to pay them anything. They've trained themselves on Free/OpenSource software because it's available, and because they can work with it and learn from it.
These folk will work for dirt cheap.
I ask my friend Ross, marketing analyst and owner of a company, "Why are you installing IIS? It costs much more than Linux, and a well maintained Linux system is far more secure, and comes with good free database software, should you ever decide to grow that way." He says, "Because UNIX admins cost $100,000/yr, whereas an MCSE lackey costs half that."
I've got one student who has set up ArsDigita Community systems, Oracle databases, Linux, FreeBSD, reads kernel source, and has written some programs in C, tcl, Perl, and Python. He's working as a stocker for some odd store. He's configured systems left and right. He's dying to get a $38,000/year job somewhere doing anything. In his free time, he works on studying embedded systems and attending GSLUG meetings.
He's not an isolated case, I have many other well talented students who are working hard and doing well, developing their talents.
They want jobs. Any jobs, just to show off their talent and get experience.
So when I read: "Unfortunately OSS engineers typically cost much more than MCSEs," I wonder: What crack is everybody smoking? There's plenty of great UNIX people out there. who will work for cheap..!
Dear Slashdot,
I work for a custom web solutions company and I'm working on a white paper and presentation that outlines the merits of open source software. This looks to be a difficult project, and I really don't want to to work late this week. Please use your unpaid minions to help me get my research done so I can I can enjoy the weekend.
OSW: Open Source Work -- Same great salary, but with reduced workload.
-----
D. Fischer
ShoutingMan.com
I've worked for some large companies (Borland, Actuate, SST), and my perception does not match yours. I got 4 Linux machines into Borland back in 1997, before they "got the gospel." My current company is a chip company -- fairly cyclical industry -- and in downturns, they are always interested in saving a quarter-million. If I can show them a small, 3 machine Linux cluster put together for $15,000, and have it work as well as the $90,000 E450 box, well they're interested in that.
I'm not saying it is easy. I have to sell the ideas and talk to the right people. But I am suggesting that if this guy gets his report in front of the right people, buying patterns can be altered. I've done it.
My Greasemonkey scripts for Digg &
The cost of software is one of the lowest cost points for corporations. Among the highest are the salaries for the employees. So if you could furnish the company with low cost open source software and low cost open source engineers, you'll be doing good.
Unfortunately OSS engineers typically cost much more than MCSEs.
The costs don't balance out in favor of OSS.
Dancin Santa
Lets consider some facts now:
- Do the average Corp, the ability to modify software means precisly dick!
- It's often hard to get good support of OSS. Good in this case means timely. If one of our HP servers goes down, we can be on the phone with a HP tech in less than 5 minutes. Trying getting a tech that quick with OSS, without paying through the nose in advance. Lost time == Lost Revenue
- Retraining isn't free. If you're using Software XYZ, it's going to cost quite a bit to train your staff to use Software ABC
- If it ain't broke don't fix it. When our commericial software does what we need it to do, why SHOULD we switch, and risk failure? This ain't somebodys home Linux firewall box we're talking about here.
To sum up, IS IT REALLY CHEAPER? In a large number of cases (not saying all) this is NOT true. OSS software can be GREAT, but it isn't always the best tool for the job, and that, at the end of the day, is what REALLY matters.TODO: Something witty here...
For example, I recently proposed that, instead of sending me to a class to become intimately familiar with Solaris and Sun's hardware (my work involves Solaris on a daily basis), the company buy me a SunBlade 100 and a few hundred dollars worth of books. This would, of course, save considerable money for the company. My manager liked the idea, but that's pretty unimportant, because most Fortune 500 companies have a strict set of rules when it comes to computer usage/acquisition. I couldn't get a home-built computer for my desktop here if I wanted to, because corporate policy, for uniformity, dictates that we buy the Dell OptiPlex.
Unfortunately, it's much the same way with software. I do happen to run Linux here as much as possible, but I'm forced to dual-boot it with Windows 2000, because it's the company standard. Not because we were forced into it by Microsoft, but because we need to have that assurance that it's going to work and not need to worry about permissions.
I'm as big a fan of Linux as anyone, and I've been using it since 1996 now (good ol' Slackware 2.x days)... but the Fortune 500 just isn't ready for it.
Every once in a while I like to masturbate a new word into my vocabulary, even if I don't know what it means.
But there is nothing at all forcing you to keep upgrading to these newer versions. There's nothing to force you not to.
If your system works don't change it. What do you think this is, Windows? Are all old kernels going to disppear or suddenly stop working because new ones have been released? There's plenty of linux boxes around here, doing their job, day in, day out, with extremely minimal administration, on 2.0 kernels. The same is definitely not so for old Windows machines. You seemingly have little choice in upgrading Windows, simply to get ever closer to that much promised stability and power.
People who upgrade on every single kernel release are suckers for punishment and you don't do it on production servers unless there's a specific problem that's been fixed. In which case the constant release of bugfixes and patches is highly beneficial.
Open Source is proven. It's been around for far longer than proprietary solutions. As always it's pig ignorant management and kiddies who think VB is god's gift to programming who think otherwise.
If you run server systems and want good uptime and efficient support you pay for professional admins and quality hardware. People who reckon that just cos you're using a PC cluster+MSCE combination it's somehow some vast saving on the outlay for a qualified Unix admin? Utter tripe. Using cheapo hardware and clueless point-and-click trained admins loses you just as much if not more through downtime and inefficiency.
There's solid working practices that should be adhered to whether you use OSS or not. A cheaparse MS solution is just as bad as a cheaparse Unix one.
Frink (the little pig in the more costly to build proverbial house made of bricks who can sit back and relax and laugh at all the straw houses blowing down all over the place, silly, cheap piggies running around like crazy trying to keep them up)
"Don't get mad, get a monkey!"