Saturday, November 18, 2000

Why OpenBSD?

OpenBSDIn response to "Why OpenBSD?", Nick Holland wrote a response that I thought deserved repeating:

From: Nick Holland
Sent: Friday, November 17, 2000 11:43 AM
To: misc@
Subject: Re: Why OpenBSD?

Here's my list:
1) Security. While encryption is nice, what I like the most is the fact that OpenBSD is the closest thing to a set-and-forget Internet-ready OS I know of. Unlike a lot of people here, I work with very small clients (one to 100 employees, typically) -- companies that can't afford a full-time IT staff to babysit their computer systems and internet connectivity and services. I've spent the last eight or so years on my own supporting Novell Netware -- which I rather like due to the fact that I can set a server in place and hear from the client sometimes once a year or less, only when their needs change (or something breaks). I avoided Internet connectivity for a long time, as I just saw no good way to keep up my reputation for solid and stable solutions. I am not into routine updates for my clients -- if there are features they need, well, fine, update, but I really don't like the philosophy of "Here's a new version, you have to update (and pay me), uh, Just Because". While there have been some grumblings here about the dropping of support of older versions of OpenBSD, in roughly a year that I've been installing OpenBSD, I have had no reason to upgrade a client from the installed, CD version that was out at the time of the install. If a client changes needs, well, I might well slip an upgrade in (besides, I've learned a lot -- thanks, Misc@!), but the 2.5 and 2.6 were good platforms, no need to upgrade Just Because.

(and yes, from an economic standpoint, I would make a heck of a lot more money if I installed NT..on the other hand, I rather enjoy the militantly loyal clients).

Plus, having had a personal Linux system cracked and turned into a cracking shop got me VERY interested in choices where security was planned BEFORE the product shipped, not "we'll fix it when they find it". Doesn't matter how fast they write patches, if this means you have to be hovering over the security patches sites, it is bad.

2) Stability (operationally). OpenBSD just runs. I'll happily trade features I don't need for stability for server-like applications. O.k., OpenBSD doesn't support every bit of hardware out there. Fine, just have to design my systems instead of taking-as-markted, which is fine with me. I'd rather have one well-written driver than twenty (or two hundred) poorly written drivers.

3) Stability (of design). Being a BSD derivative, the basic plan has been laid down years ago, it is not being re-written every release. Pick up a book on Unix administration, it works with don't have to check the copyright date, or what version of what dialect it supports. I had looked at Linux as a platform to offer my clients, but despaired at the thought of supporting numerous versions, numerous distros, etc. If I were to constantly be updating my clients systems, or if I was the full-time administrator of a small number of systems, well, perhaps it would work out, but as I might end up supporting fifty or more very dissimilar systems, I don't think there is anyway I could keep up with Linux for my client base.

4) Support: While mail-list support may seem horribly unprofessional, I've found the quality of guidance offered on Misc@ is so far beyond what I could ever hope to get from a "single-point-of-contact" provider -- and faster. Look how many responses you have probably received to your post in the time that you would have sat on hold with most companies (granted, you asked a sales-oriented question, so you probably wouldn't have had to wait on hold very long). Instead of talking to one person, who for reasons of ego, will answer your question regardless of quality of response, people who know hardware can answer your hardware problems, people who know software can answer your software problems, experienced administrators can guide you on administrative strategy, etc., and for every reply, you get hundreds (thousands?) of people looking it over, and ready to pounce if they see anything wrong with it.

5) Documentation: Show me another modern OS (free or commercial) which keeps its documentation as up-to-date as OpenBSD. We all know documentation isn't nearly as fun to write as tossing new features in, but these people do it. Praise be to the OpenBSD documentation team!! Also, as indicated above, while the QUANTITY of books that cover OpenBSD is very small compared to say, Linux, the *BSD/Unix lineage means that any well-written book on Unix is applicable and useful. Sure, you may not get a key-stroke-by-key-stroke guide, but I don't find that style of book useful. You also get books written by someone who has something to say, not just books written by authors who feel that writing a book on the latest popular app or OS is a good way to put food on the table.

6) Design: Even though OpenBSD is definitely a team project, there is a clear Leader, which means decisions about issues don't end up being discussed endlessly by a number of people claiming to be equals. Even if I don't agree with every decision (that's a hypothetical -- I'm sure one of these days, a decision will be made I feel qualified to comment on 8-), I would rather say that I'd rather have a decision by a competent person than an endless series of discussion.

Welcome to OpenBSD, I'm sure you will be coming up with your own list soon. 8-)