Slashdot Mirror
Paid Support Not Critical For Linux Adoption
ruphus13 writes "At the LinuxWorld expo, an analyst for the 451 Group pointed to a growing trend in enterprise — the increase in adoption of community-supported Linux distros. From the article, 'Companies are increasingly choosing free community-driven Linux distributions instead of commercial offerings with conventional support options. Several factors are driving this trend, particularly dissatisfaction with the cost of support services from the major distributors. Companies that use and deploy Linux internally increasingly have enough in-house expertise to handle all of their technical needs and no longer have to rely on Red Hat or Novell.'"
When the person giving it to you knows what they are doing. If that person actually uses the software then they probably know alot about it. This is why community-driven support works, if you manage to keep the "kiddies" out so that they don't clog up the forums with lots of repeated/redundant questions then everything goes quite smoothly. Arch Linux does a very good job of this; it's a simple distro to use for the experienced user, so you get alot of good questions being asked with lots of good answers. Community support > paid support any day.
Obligatory blog plug: http://www.caseybanner.ca/
The paid support is for businesses who can't waste their time scouring the Internet and posting in forums for solutions. Time is money, and the sooner they get the help they need, the better. The same is true for Windows. You think Microsoft doesn't have expensive paid support? Guess again. They basically have a monopoly on it, whereas with Linux, any company can support the software competently, since the source code is available.
We started with RHEL3, especially since we ordered a Dell Server and it could come with the server, thus I knew RH would just work on it. Never seemed to get my money's worth out of support (if you are going to administer it, you might as well learn it, so I answered most questions by myself.) A year or so later, instead of going RHEL4 I went to CentOS 4 next, as it had the same necessary apps and updates, support didn't matter so I had the OS without the bothersome RedHat Network license validation nag screens.
About a year after that I got tired of CentOS - when I started looking at options for a cross-platform backup solution, CentOS was the low man on the compatible distribution totem pole, sometimes not even there at all, most support requestes ended with some vague problem with dependencies and an 'oh well'.
Also learned to shy away from SuSE then too, as I noticed around that time any Novell associated projects usually dropped any non-SuSE binaries (i.e. iFolder).
But Ubuntu had just about everything there, was well updated, and a lot of forums with solutions. Granted, Ubuntu lacked the nice SAMBA admin program (GSAMBAD needs help), but I never have any problems finding apps or resolving installation issues quickly.
"Enjoy what you're doing! If it becomes drudgery, you're doing it wrong!" - Jim Butterfield
How did Ubuntu get such a huge community so quickly? I remember hearing about Ubuntu shortly after I installed Xandros on my system, about three or four years ago. I began looking into Ubuntu, and its community was exploding, and still seems to be. I wouldn't be surprised if alot of enterprises are installing this distro now, based on its community. Yet still: why Ubuntu? Why not one of the other similar distros? Is it the name? The slogan? The color scheme? Mark Shuttleworth? What's the deal?
Harold
And while we *do* pay for support and it has come in handy on occasion, I have found that google is a far more valuable tool than their support services. First off, it doesn't take 2 days to get a response when you are using google. Second, you aren't forced to do a sysreport by some 1st tier keyboard jockey in Bangalore before they will even consider thinking about the problem you are reporting.
Now, having said that, when you manage to escalate your problem to someone high enough up, you do get quality support. you just have to jump through hoops to get there, which really does IMO make the value of the paid support rather questionable.
Feed the need: Digitaladdiction.net
Why buy the milk and eggs bundled, when you can easily feed the chickens yourself, thus making the bundle price, overpriced, when compared to the cost of milking the cows yourself.
It really depends on the business. However, the advantage that Linux has is that it has become 'common'. Linux admins and programmers are fairly easy to find these days; as opposed to, say, HP/UX (and others) that is harder, arguably, to find competent people to admin and program for. On the flipside, if you're a large enough business to pay for several sysadmins and programmers, then I would guess that the annual support fee is worth it--in effect the paid support is an ex-situ 'employee' that is available 24/7 and is not ONE employee but a team of employees. The ex-situ employee is not going to decide to go work for somebody else either...
Don't forget though that the ability of bigger enterprise-driven companies like Redhat and Novell to pay full-time linux programmers has had a tremendously postive effect on community distros.
It is hard to imagine what the linux desktop would look like today without the contribution of Redhat and Novell programmers during the last 5 years.
So how do Red Hat and Novell pay full-time Linux programmers in the future when no one needs paid support any more because community support is better going forward?
Why? It's easy to install. It's easy to administer. It's Open, available and Free. It's secure by default (no open ports). It has repositories for thousands of useful and free apps that you can get from their repositories instead of downloading them from random Internet sites. It supports nearly all the hardware you've ever heard of. Server is free. Client is free. Thin client with servers is free. Clustering is free. Did I mention that client licenses are free? You can boot it from nearly any readable media. Boot time is swift even in ways you wouldn't expect it to be (pen?). It's easy to upgrade and paths are easy too -- and free. With Open Office it reads all the common Office formats, for free. It's extensible, adoptable, and free. The BSA will not be beating your door down over this one because they want you to use it.
The better question is: "Why not Ubuntu?"
Help stamp out iliturcy.
Small to mid-sized shops who get by with less than a dozen SAs and who don't have WAN volume replication concerns might go this route, but there is too much risk for Fortune 500. It mostly boils-down to 3rd party applications, hardware and drivers. If you're a F500, you probably have proprietary storage of some sort and you probably rely on volume replication across the WAN. You want to hook into that storage from Linux, you need a "certified" platform and that ain't going to be an arbitrary set of Ubuntu packages. Sure it will probably work from Ubuntu, until you get kernel panics under load. Then your in-house Linux "experts" call support for the storage vendor and they ask what distro version and driver you're using. When you say "Gutsy Gibbon recent" they laugh and refuse to support you. At that point, your idea of community support doesn't look quite so hot considering nobody in the community can repro your hardware/driver issue.
My shop has several hundred Red Hat boxes. What do we do with the money we pay Red Hat? Primarily it is to have access to their web site and get the ISOs for different Red Hat versions, as well as individual packages if needed. Or we use up2date/yum on the machine itself to grab packages.
One thing I can say in favor of Red Hat. I used to use Debian at home (now I use Gnewsense, a knockoff of Ubuntu, which is a knockoff of Debian). For many months, the "search the contents of a package" feature was disabled on Debian's website. So if I wanted the program "sftp" but didn't know it was in package openssh-client, I could search there and discover that. But Debian just decided to take it down for a few months. Red Hat would not do that for so long, if at all, and if they did I could call and complain.
One problem with Red Hat versus Sun is if a kernel panics or whatever with Solaris, I can send the core dump to Sun and that's it - the control the OS, they control the architecture (except for Host Bus Adapters and the like), and that not only makes core dumps easier (netdump seems to be preferred on Red Hat, which I think blows), but makes them easier to diagnose - it is all coming from one source. With Red Hat you don't know if is Red Hat that did something, or your hardware vendor (Dell/HP/etc.) Which means they can point fingers at one another, with Sun can not do as it is all coming from one source. OS and hardware all from one source has its advantages. Also, the usual answer from Red Hat and the hardware vendors is we should have everything patched to the latest version, which we never do, so reporting it is pointless. Even if we had everything patched, since unlike Solaris it won't be dumping core to a local disk, we would have to go through the effort of a project where all machines could netdump somewhere. As we only have a few systems go out a year, and do not have the resources to keep all machines up to the latest patch levels, system crashes are often a mystery, which irks me, but due to our limited resources and the shortcomings of the Red Hat model, is just how it is.
The altruistic comments mentioned are silly I think. My boss is not going to shell out money to Red Hat because it goes to "the greater good". If I could get my company to send money somewhere, it would be to the Free Software Foundation.
One thought that occurred to me is companies like Red Hat might be transitional in some ways. Companies wanting to move to something open want hand holding at first. I can think of many examples like this in my career. I worked at a company where we hired Java developers and started using a professional Java application server, which we became unhappy with and then began using Tomcat. The developers said their confidence with being able to develop for the professional server is what let them try Tomcat, which worked out very well for us. The move from Solaris to Red Hat to free as in beer Linux is another example. I see another example with MySQL recently - looking to save money, a division is going to use MySQL for a new project as opposed to Oracle, which they traditionally use. After a few years, might the DBAs drop professional MySQL and go with a non-supported MySQL? Who knows?
I think the companies like Red Hat and MySQL, if they are adaptive and fine tune their business strategies, can survive this transitional stuff. The more traditional companies, the Microsofts and Oracles and Suns are who should be worried.
Support is not where the money is for free software IMHO. And actually, although I'm not in the loop for these companies I don't think either of them make most of their money from "commercial" style support.
The big money is either in custom distribution builds or in custom software development (or both). Usually you sell a "support contract" with it too, but it's more of an extended warranty than a real support contract.
I once had an interesting talk with a salesman from Novell (who is a big free software fan). He told me that he doesn't try to sell support contracts for Linux. Instead he's more interested in providing upgrade paths for existing Netware customers. These products run on Linux and to compete against Microsoft's offerings they need a full package deal (office suite, email, etc, etc). In fact, from his description of what they were doing, I got the impression that the support side was still being run as a "loss center" rather than a "profit center".
To make a long story shorter, successful free software companies will make money providing specific solutions to customers. Those that rely on "generic" (IMHO, useless) end user support will die an ugly death. However, I don't believe that any of Red Hat, Novell, Canonical, IBM, Sun, etc, etc are trying to base their business on end user support.
So we can expect to see more of the same.
At work we would only be happy to pay for commercial support and updates, but we choose to use Fedora instead of RedHat Enterprise simply because Fedora is a better product for what we do, and Redhat does not offer commercial support for it. The enterprise version is geared toward network administration and services, but for a development shop, having access to the add-on Fedora repositories like livna, more up to date software versions, and the greater user base makes Fedora a far better platform.
Seems like RedHat missed the boat on desktop Linux, and Ubuntu ate its lunch in that market. I wonder if they will ever try to make a comeback, or if they will be happy in the network niche.