Commercial Linux vs Free

What benefits do customers actually get from the commercial distros except for support and ability to run certain "certified" applications (such as Oracle)? Are they really more stable/reliable? For example, RedHat vs Debian. 350$/year vs 0$. What can you say?

Solution 1:

If you're a growing organization, sometimes the expensive option is actually the cheaper option. Having access to high-quality support and training courses is an advantage in many situations.

Another factor is how do you manage applications? If you're going to roll out an app, and just want to leave it alone and patch security vulnerabilities for 5 years, that's more viable with something like Suse or Red Hat.

Solution 2:

We use Ubuntu and Debian servers in my company. My system admin screamed for ages to install RedHat but the CEO wanted to keep the IT landscape the same.

In my company we have 6-7 people with A LOT of Debian/Ubuntu experience and we can basically get anything done very quickly.

When I came into the company I paid for 2 Ubuntu servers to have support for a year and I haven't even used it once.

So basically, what I am saying is, depending on the experience of the people of your organisation. If its not that great, then get support. If it is, then its not really needed.

Solution 3:

The question to ask yourself is: how many other people need what I need? If you're doing what everyone else is doing, it makes sense to outsource the bulk of the work to a specialized outfit who can hire kernel engineers, monitor bugs and security patches, etc. RHN can make managing servers much easier. I should note that Canonical offers a similar service though I haven't checked their prices.

However, there comes a point when your organization wanders past the run of the mill. At that point you need experts who can do all that with an eye to the breadth of possibilities. For example, the Sanger Institute originally ran their Human Genomics servers on a proprietary UNIX that could handle 64bit filesystems and so on. They later moved to Debian and build a custom Linux kernel that met their needs in the late 90s when their hardware support was discontinued and other unfortunate events. Two of their employees were Debian Developers capable of making all that happen.

It's tempting to price a community distro like Debian at zero and hire novices to handle things, but I believe it's more expensive in the long run when your novices failed to plan for things Redhat and Debian Developers are aware of. If you run Debian and don't give your admins the leeway to pursue Debian contribution, you're losing a lot of valuable free peer review.

In short, the price of Debian isn't 0, it's the price of hiring a DD or two.