Ask Red Hat CEO Matthew Szulik

Red Hat has made several changes in how they run their business, notably concentrating more (perhaps one might say "entirely") on enterprise-level Linux users. Some of Red Hat's moves have upset long-time users, and many people seem to have trouble understanding exactly where Fedora fits into all this. Red Hat CEO Matthew Szulik has offered to answer your questions and clear things up, so ask away. Please don't ask questions he's answered in recent interviews and statements, and try -- hard though this may be for some -- to ask only one question per post. We'll forward 10 or 12 of the highest-moderated questions to Szulik tomorrow, and run his answers when he gets them back to us.

Comments

I have a question...

[FraggleMI]

What kind of a business model is it to lure everyone into using the RedHat Desktop, and then drop support. Seems like the Microsoft model of forced upgrades.

Fedora

[radixvir]

Don't you think if more users are using an operating system they will be more implied to use that same operating system at the workplace or recommend it to others. In that case, why did you recommend windows for desktop users?

Re:Fedora

[Acidic_Diarrhea]

The people who are in charge of purchasing decisions for a server aren't basing those decisions on what desktop OS they use. People who make those types of decisions should have some type of technical background and should be familiar with Windows, UNIX, and Linux variants. And Red Hat is pushing Linux as the choice in this case. If I use Windows at home because I like playing games (and I have the technical expertise to run a server) do you honestly think I'm going to choose Windows? A better line of thought is that people who run private servers at home using Y Operating System are more likely to suggest that Y Operating System be used at work. But they're still two different markets.

Re:Fedora

[dAzED1]

how about a freakin clue, bub.

We're talking about OS's. The Director of Information Technology was once a help desk guy, or at least a network admin. HE is the one making the OS recommendations, for servers and clients alike.

And guess what - he uses a computer at home.

More importantly, his replacement 5 years from now uses a computer at home.

The upper and middle IT management of today are the network admins of 5 years ago. 5 year plans, people. 5 years ago, these folks started playing with linux - many of them, redhat.

5 years from now, there won't be the grassroots component. Grassroots is what MADE redhat, and every other linux distro. It is their backbone. And regardless whether they think otherwise, not even linux in general (regardless of distro) is free from needing to keep the grassroots strong - that's the reason Fedora exists at all.

So...same operating system? Learn to build your own damn kernel. The apps are either the same, or available. RedHat is still barely more than package management - the only difference between "fedora" and "redhat advanced server" is certifications, and default capabilities. So yes, you can get the latest kernel source, build it, and YOU TOO can have smp support, >4gb ram support, or whatever else is supposedly "missing" in fedora.

Its the same OS.

Re:up2date

[jimpop]

If the up2date service is NOT going to continue, will a refund be made to those whose pre-paid service agreement is prematurely terminated?

personal OS choice?

[BigGerman]

Which OS and desktop environments you, your colleagues and friends use every day?
thanks in advance for your honest and direct answer.

Question to mr Szulik

[motox]

Why the OpenSource community should contribute to what's effectively your open beta test ( Fedora ), a project that will never be stable, and then have to buy back from RedHat their own work ? Don't you think the users will simply shift to other distributions, especially given the ludicrous price you set for an operating system that is 99% based on free software ?

Small Business Market?

[gamartin]

Question: Why has Red Hat never articulated a strategy appropriate for the small business market?

Example: My small business has 8 workstations and 2 servers; here's what's important to me:

1. Moderate release cycle -- small businesses don't care about bleeding edge features
2. Security/errata updates -- small businesses need someone else to monitor these complex issues
3. Support available for at least 3 years -- small businesses do care about stability and hate forced upgrades
4. No per-machine licensing restrictions -- small businesses look to linux for cost savings and will not tolerate per-machine licensing; product must be installable on multiple machines to realize cost savings
5. Metered support options -- small businesses are willing to pay for actual support services used
6. No compliance audits -- small businesses do not have time for that type of crap

I'm willing to pay roughly $200/year for standard support services for these machines plus per-incident costs if they arise. I have been running Red Hat 7.3 with 2 Red Hat Network subscriptions and manually propagating updates to the other machines (which is annoying but tolerable since N is small).

I have been a paying customer, and I'm basically amenable to any sort of metered service system where payment is for services used. However, now I am being jettisoned as a Red Hat customer: Fedora has no support, and Red Hat Enterprise Linux is too expensive. Red Hat has all the resources already in place to support my needs, yet is unwilling to do so.

Why is Red Hat unable to support this type of revenue stream which seems perfect for linux?

Re:What's next?

[Anonymous Coward]

That is patently false. For the average person, there is no Linux. And if someone has heard of or even uses Linux, then they are certainly knowledgeable enough in the area to know that there are many distros. If they know Red Hat exists, they certainly know of others. I challenge you to find one person who use(d) Red Hat Linux on the desktop that didn't know they had a choice.

Support from Redhat

[LonelyKindGuy]

Greetings Mr. Szulik!

I would rate as one of your best assets the old Cygnus organization which is providing excellent support on the GNU gcc/g++ compiler. My company was a Cygnus client for years and has also gotten great support from the Redhat team in this area.

Enterprise is a word that always sounds big to me. How much support can you give to a smaller company? What kind of price levels/support levels (approximate) is Redhat now offering?

SOHO Support?

[soloport]

I'm not asking for much. Since I've already switched to Fedora Core, I've noticed that up2date still/already works with Fedora. What struck me was that there was no sign-up process! The packages downloaded without a hitch. Will this service continue forever? If your plan is to discontinue up2date support for Fedora, why?! Why not just keep charging for a RHN-like service?

I have at least a half-dozen entitlements -- faithfully renewed each year. I've offered a few of my paid-for entitlements to clients, for free, as part of my service. My plan has been to expand this to more of my clients in the near future. But now, I feel stuck.

These are mom & pop shops (in the dozens) who will NEVER be able to afford your Enterprise offer. They wouldn't know how to keep their Red Hat, back-office server up-to-date if it meant saving their business. I make a living by saving these people from hours and hours of servicing Microsoft patches, updates and malware. If you will not be effectively supporting the SOHO market (including my clients), what do you recommend?!

SOHOs know "Red Hat". I will have to teach them "Mandrake", "SuSE", or perhaps maybe not so much "Novell", instead. I believe today's SOHOs are tomorrow's Enterprise buyers. What do you believe?

Re:SOHO Support?

[pivo]

I make a living by saving these people from hours and hours of servicing Microsoft patches, updates and malware.

If you didn't make money doing this, would you still do it? RedHat didn't make enough money providing updates for their desktop distribution, so they stopped doing it. Seems rational to me.

There's no sign up process for Fedora because it's no longer required, not because it's going away. What has changed is who's responsible for providing updates. It's now not RedHat's sole responsiblity, it's a community process just like Debian.

You need to take a deep breath, relax, and read the information at http://fedora.redhat.com

Explain the strategy?

[duffbeer703]

How is Red Hat going to deal with the multiple free distributions that are bound to start eating at your market/mindshare?

The only thing the makes your "Enterprise" sustainable is the support of commercial software vendors like Oracle, IBM, etc.

What happens when Oracle decides that it's easier to provide their OWN distro for running Oracle?

It seems to me like RedHat is turning its back on the community and throwing itself to the wolves.

If you know what the answer will be...

[Jerf]

you shouldn't ask(/waste) the question.

This, and a number of other highly-rated questions where the answer is "Fedora" (followed by what will boil down to some hype for Fedora), should probably be moderated "Overrated" in the interest of presenting questions for which the answers the Red Hat CEO will give are not immediately obvious.

(Normally I wouldn't question moderation, but in interviews mods are more like votes, so this is a valid opinion.)

(And of course, in the event this gets rated highly it does not constitute a question.)

Re:Emphasis on Enterprise-level Linux?

[mr_z_beeblebrox]

Mr. Szulik, I am a desktop user of Red Hat, and your recent emphasis on Enterprise-level Linux leads me to ask if you know where I can get the best price on a copy of Windows XP?

I knew this question would show up quick. Let me sum up how this appears to me...Mr. Szulik, I am a laoyal Linux advocate and longtime user of Red Hat Software. I have downloaded the OS that you put together with your high paid developers (using your expensive bandwidth) ever since RH6.2. I can not understand why you are selling out and abandoning us....we got you where you are today.
Okay, so that may not be quite fair. However, I am guessing that the desktop was a financial loss for Red Hat. It was one that they cleverly supported, but a loss none the less. The fact that they supported it made a larger Linux base etc....and they benefited intangibly, but a board of directors will not tolerate intangible bennies only for long. A corporation is a math machine work plus money = more money that equation MUST be satisfied. Red Hat is going a natural route. When Linux is entrenched in many small-mid size corps then the desktop will be opened up. For now it is GENEROUS of Red Hat to support Fedora.

Why stab everyone else in the back?

[jazman]

I appreciate that corporate goals change, and after supporting Linux extremely well for many years now you have decided to focus on corporate customers and drop your support for the consumer market. What I don't understand is why you said what you did, that Windows was better for this market? Perhaps it is at the moment, but there are other distributions still trying to change this, and I feel that your statement has now given them a major competitive disadvantage against Microsoft; all Microsoft needs to do is to state to any customer (by which I mean, for example, a PC retailer who sells PCs to the public preloaded with Windows and is now considering Linux as well) that even Red Hat, who should know, don't think that Linux is appropriate for their computers, and the relevant Linux vendor suddenly now has to patch up the hole you just created. Exit a game by all means, but why shoot the remaining players?

Mods: checked the reports on this for reasoning but didn't find anything; if I missed it please feel free to mod this to oblivion but I would still like to know.

Re:Why

[mattdm]

Notice that they *haven't* stopped. It's just called Fedora now, and has a more open developement process.

Re:Why

[BadCable]

A product losing money!= company losing money.

It's common place for a company to produce many products and for some to lose money but the company to come out with a profit thanks to the other products.

By getting rid of the "losers" the company can raise profits by doing "less"

Re:Someone's already doing that

[Pros_n_Cons]

They might not encourage the Idea but I don't see how they would care. This new 'copy' would have to pickup the slack in support, bandwith. What business in thier right mind trusts thier business on software patched, updated and written by a guy referred to as "Someone"? Not me! And if your goal is a desktop for home users, who wants a slow stable (boring) OS at home? Not most of us. We want exciting new things consistantly. Everytime redhat had a new release routers were being burned out in some places of the world because of so much traffic (seriously) so that says we are eager for the latest and greates which is why Fedora is created. Its what we asked for.

IBM and RedHat

[darnok]

IBM has used RedHat's products at several of their sites - primarily US-based sites as far as I'm aware. IBM seems to focus primarily on using SuSE for their European-based customers.

Now that desktop RedHat has been replaced with Fedora, and the consequent loosening of control of the distribution by RedHat, has IBM indicated any change to their business relationship with RedHat?

IBM is now pushing to have more Linux desktop systems out there, and presumably either is or will be sending that message out through their field consultants. Now that you've dropped desktop RedHat, are you concerned that SuSE, or indeed any other Linux vendor, may step in and establish themselves as the dominant desktop Linux platform by riding on IBM's coattails, and that they may be able to leverage this strength to cut into your enterprise sales?

Where will your *server* software be field-tested?

[Olinator]

I work for a large university's CS department, as a network/server admin and resident linux guru. I've been a minor contributor to many RH bugzilla discussions, and may have helped to solve a few esoteric bugs here and there. RedHat's recent move leaves me in the cold; we can't move to RHEL because the cost would kill our IT budget, and I doubt very much that the Fedora model will work for us. Reading upthread, I see several posts whose authors tell much the same tale. For all of us, the Fedora description and proposed release schedule aren't something to which we'll feel able to trust our servers (and our professional reputations).

In those Bugzilla discussions I mentioned, I've seen a lot of highly-informed, to-the-point correspondence from people just like me -- we have no problems patching driver code, recompiling kernels, parsing debugger output &etc., and I'd hazard a guess that the issues we raise and the bugs we help troubleshoot in the field (the ones that make it past your inhouse QA folks) are part of the reason that RedHat has been so rock-solid. This translates directly to the stability of the server-class packages you sell to our cousins in the corporate trenches. The targetting of Fedora at bleeding-edge enthusiasts and hobbyist installations means that these bugs (remember, these are the ones that made it past the in-house RH team -- have a look at the tg3 driver issues in bug 69920 for one example) likely won't be caught before they bite your paying server customers. Do you foresee a decrease in stability for RHEL as a result? If not, where do you envision getting your field/beta testing done for the server components of your OS? Isn't it possible that, while taken by itself the "free" version of RH was an operating loss, when viewed in context of overall product line it was actually part of the reason you started operating "in the black"?

Red Hat's Financial Priorities

[Ilan Volow]

(not like this will ever make it to list. But wth)

Mr. Szulik, I keep seeing Red Hat put out one unusable, confusing and ambiguous interface design after another. Anaconda is a textbook case of what you shouldn't do when designing a user interface for non-geeks.

A Red Hat employee a while back who went to my campus' LUG told me that the reason why Red Hat software had such bad usability was they didn't have the money to fund HCI folks running a usability dept. at Red Hat.

Yet when we look at the financial history of Red Hat, we find that the company spent over $700,000,000 buying out other companies like Cygnus (purchased for $650M).

Mr. Szulik, how would you answer the charge that your company destroyed its chances to gain real home desktop marketshare by not investing a small sum of money that would make your software more accessible to the people you were trying to target, as well as substantially lower the costs you company would have incurred supporting that market?

--
Ergonomica Auctorita Illico!

Re:Education and Research Markets

[Confessed Geek]

I am in exactly the same boat. Our reseach group works closely with the both of the major midwest research centers and I'm pulling my hair out deciding what our right path for post-RH will be. The pricing of the Enterprise systems is just prohibitive for us, and for reseach fedora is too unstable to provide a fixed enironment that will last the length of an experiment. My perspective is obviously skewed but with the thousands of linux systems in use in just one of these major labs I would think that the Research/Academic sector is one of the biggest markets linux has.

If RH could provide some sort of option for non-profit/academic/research groups we could continue to use it. However the lack of this option is putting these communities in a serious quandry. Some thought is being given to repackaging the Enterprise version after removing the propriatary bits, but this seems like re-inventing the wheel to me. Overall this decision on RH's part is leaving a very bad taste in the mouth of a lot of influencial people (and some non-influencial folks like me).

Where I would like to see the Scientific/Research community go is to Debian. Easy upgrade, conservative release durations, excellent security /errata support, and a great package managment system which would make it easy and convenient for different research groups to package and distribute their home grown tools. The philosophy of Debian is also closely in line with the Research/Academic ideals and mindset. I have been told that such a big distro jump may be problematic due to the dificutly of porting some of the core software used, but while the individual cases may be a bit tricky, I would have to think that the overall labor would be less than rebuilding and distributing all of Enterprise. The other objection I've most often seen is that debian is to hard to install. With the likes of Knoppix, Anaconda being ported by the Progeny team, and the new Debian Installer I don't think this is really a valid excuse. Admittedly, I've got quite a bit of practice at it now, but even using the default command line install I find I can do a debian install in about a quarter of the time (including downloads) that it takes to do a RH install (including up2date) do to the ease of package management.

Hmm... to keep on topic:

Dear Mr. CEO,
How badly do you think it is going to hurt Redhat in the long run to alienate the scientific and academic sectors - two groups who are introducing linux to lots of smart professionals and students who will be the corporate decision makers 10 years from now?