Is CentOS Hurting Red Hat? 370
AlexGr writes "Jeff Gould raises an interesting question in Interop News:
Why does Red Hat tolerate CentOS? The Community ENTerprise Operating System is an identical binary clone of Red Hat Enterprise Linux (minus the trademarks), compiled from the source code RPMs that Red Hat conveniently provides on its FTP site. It is also completely free, as in beer. CentOS provides no paid support, but it does track Red Hat updates and patches closely, and usually makes them available within a few hours or at most a few days of the upstream provider, which it refers to for legal reasons as "a prominent North American Enterprise Linux vendor." Free support for CentOS can be found in numerous places around the web, and a few third parties offer modestly priced paid support for those who want it."
Centos brings back the 'play at home' (Score:5, Interesting)
What changed it for me was Centos. I found that I could use the free as in beer versions for all my personal/internal needs, and it was so dang close to OEL and RHEL it became a no-brainer for testing and some dev work. With the internal blessings from our side that our code would work, QA did the formal testing on the branded versions of Linux. Folks running our product, of course, would want OS support - so they purchased the formal 'supported' OS from the commercial vendors. I suspect Centos is saving RHEL/OEL sales that might have gone to Ubuntu or other variants.
Re:Simple: Support (Score:2, Interesting)
Looks like the kernel is not the only one who was "Dazed and confused"
Joshua SS Miller
It works both ways (Score:2, Interesting)
It doesn't matter how many times I say 'CentOS is 100% compatible, and FREE! (w00t)' to my boss. When a machine goes to a customer, it goes out with Red Hat. Even if no one ever calls Red Hat for support, that warm fuzzy CYOA feeling of having a big well known company behind your product is irreplaceable. At the same time, we have a stack of CentOS machines and VMs in the office for testing and development for no additional cost to us.
I can honestly say that CentOS made Red Hat a much more palatable choice as we switch away from our previous UNIX- SCO Openserver.
Redhat support (Score:5, Interesting)
A Very Tough Business (Score:5, Interesting)
RHEL, CentOS, and Fedora are all competing brands under the same umbrella. Fedora is great for cutting edge developers and home users. CentOS is good for people who desire the better tested software. RHEL is targeted at enterprises (hence the 'E' in the acronym) who need things working all the time (99.9999%). The three different markets are comparable to the different brands offered by Microsoft (Server, Workstation, Home). The only difference is that Red Hat doesn't make any money from CentOS or Fedora.
But take a step back and think about Microsoft a bit more. Imagine you have a business laptop which was provided to you by your company. It runs 2000 or XP or (god forbid) Vista and the company has a site license for you to run that software. Microsoft is happy to slash margins for the individual site license which you have as long as they can continue to service the servers and infrastructure which run the business critical systems of your company. Similarly, if you are a developer or home user... your copy of Windows came from an OEM or you pirated it. Sure, Microsoft gets money from Dell and the other OEMs... but (I imagine) so do the Linux companies who have been able to get involved in that method of distribution.
In the end, you help Red Hat by using CentOS or Fedora just like you help Microsoft by using pirated Windows. Simple enough?
Re:nope, doesn't hurt RH (Score:2, Interesting)
Re:Simple: Support (Score:3, Interesting)
Ah, so that explains how they've built a multi billion dollar business on providing software services and support. Wait, what?
Re:nope, doesn't hurt RH (Score:3, Interesting)
The problems among 20 identical servers might occasionally be the same, but Murphy's law always dictates that they won't be.
Cheers.
Re:nope, doesn't hurt RH (Score:5, Interesting)
Re:Simple: Support (Score:1, Interesting)
Red Hats support is not their selling point. Their software services and training may be; their updates and RHN another. Their so called support is a joke for all but the very largest enterprise customers. I should know, I manage the Red Hat relationship for a Fortune 500 company with thousands of RH seats. We have currently 100+ open support issues on the Evolution mail reader alone where Red Hat support have does absolutely nothing. Their standard response is that "it will be corrected in a future release" but that makes their service not what I would call support, it is accepting bug-reports.
Re:Doesn't hurt them at all..... (Score:5, Interesting)
Myself I've used Ubuntu series of Linux on my home machine because its better for desktops but if I were to run a server I'd probably choose CentOS for myself (or a small business), RHEL if I had a big budget in a major company.
Re:Because it's essentially advertising? (Score:2, Interesting)
After running Zoot (Redhat 6.2), I decided RH wasn't the distribution for me. I've run (and still are running) several other distributions after that, including Slackware 7-10, Debian Potato, SLES and OpenSuSE. Since the Novell-MS deal, I cannot trust SuSE enough, and I switched again... to CentOS. And now I'm considering the next servers to get a paid-for RedHat.
If it weren't for CentOS, i would not have bought anything from RH...
I guess I'm a little slow (Score:3, Interesting)
In fact, working mostly for very small companies, I've never seen any throat of upstream vendor take as much as a deep gulp. Even with fairly expensive software products, you still get a junior tech who usually insists for the first week (or more), despite comprehensive technical attachments to the contrary, that somehow you aren't using the expensive product correctly, because a product that expensive made by such a large and powerful corporation with such a long history couldn't possibly be that incredibly broken.
A recent nightmare that comes to mind was the Xilinx ISE Webpack ignoring pin constraints, claiming in most outputs to have satisfied them, but if you dug down deep enough, you could find a report that told you where the pin was actually bound. Of course, Webpack ISE is not an example of an expensive support contract, but even so, Xilinx has a $7B market cap., with a long history, and if they can't get something this basic correct, what exactly has their corporate stature done for you?
By the time the error in our prototype system was detected and fixed with much recriminations and pulling of hair, the damage was done, we had discovered our own work-around tweaking the source code, and what was to be gained by turning the thumb screw on Xilinx? In small companies, I've rarely had the level of support where I could choke a throat without first filling out an application form. Oh, the primal satisfaction.
In the cases where I've seen expensive support prove its worth at an engineering level, the companies involved have a deep enough business relationship to fly engineers back and forth for training, support, and knowledge exchange. Any support contract below this tier, no matter what metalic luster is applied by the marketrons, has good odds to waste more time and talent than it salvages. The exception where it can pay off is for orgs (esp. high-powered revolving-door consulting orgs) that hire at the bottom of the experience pool, people who can't actually use the software correctly, and where the money you save by hiring unqualified workers pays for the expensive support, with the added bonus that your unqualified workers acquire portable skills far more slowly than if they ever solved a problem themselves.
Where I have seen the implied law-of-the-jungle "choke the throat" aggression play out is mid-level a-hats in development meetings insinuate "we are paying tons of support money on this expensive support contract from this powerful company, so how come you can't get it to work?" This only serves to nourish all the standard corporate dysfunctions, drives the wedge between marketing and development ever deeper, mandates CYA behaviour from every quarter, and tilts the landscape in favour of those whose job descriptions are primarily political in nature over those who have technical obligations to complete, and who can't play politics on a full-time basis. What's not to like?
I've heard this meme for twenty years--usually tendered by the fetishists of corporate jargon--and I'm still asking myself "whose neck is actually being choked here?" It has always struck me that the primary function of these contracts is to make it easier for the suits to denigrate the geeks.
Here is a telling observation. When is the last time any bean-counting suit ever asked "did we get our money from that support contract or not?" Of course they got their value: the empowerment to drip condescension toward their technical staff over every snag and delay. "Of course these problems are not a management failure, we bought the gold platinum titanium support package from every vendor we sourced. It must be our own clunkhead engineers who can't get anything right."
Re:nope, doesn't hurt RH (Score:2, Interesting)
You hit the nail on the head with this one. No IT manager wants the hammer to fall on them when the shit-hits-the-fan and they have to explain to senior management that a mission critical application went down because the free (as in beer) operating system they were running offered no support. The cost of a RHEL license ($1200) is nominal when you look at the cost of downtime of a major application running in a large business. Not only in terms of revenue, but time spent by IT staff to correct the problem. Having someone to call (in this case RedHat) and say "Fix it", is just good business.
On somewhat related note...
I have to ask, who writes these articles? They really have no bloody clue what they are talking about. I swear I am going to find the dude who invented the "blog" and kick him in the nuts. It has resulted in nothing but an endless crapshoot of self-rightious wankers who get off listening to themselves spew garbage on topics which they have not the slightest clue. To make matters worse, there seems to be a plague of mass-stupidity running ramped through the internet where people actually read this shit, and even worse, submit it to Slashdot. Oh how I long for the days where there were these people called "journalists" who did "research" and wrote about "news". Somehow we've ended up somewhere where any old retard can become a "blogger" and pollute our news sites and RSS feeds with their diarrhea-of-the-mouth.
Re:nope, doesn't hurt RH (Score:2, Interesting)
Re:nope, doesn't hurt RH (Score:4, Interesting)
A big thanks to RH for continuing to support the community by not throwing a wrench into projects like CentOS, Whitebox, etc...
Re:nope, doesn't hurt RH (Score:2, Interesting)
making modifications to it."
Clearly YANAL. Red Hat cannot make excessive changes to the source that are obviously made to obstruct modifications or building. The GPL does indeed specific what constitutes source code for a work.
Re:nope, doesn't hurt RH (Score:2, Interesting)