Analyst says New BSD Releases won't win new customers

[nilsgecko]

Specifically talking about OpenBSD and Dragonfly BSD.

http://www.v3.co.uk/v3-uk/news/20467...alyst-suggests

[TerryP]

Somehow I think this guy is thinking more about Solaris and Mac OS X Server, because there really is not much Commercial BSD UNIX systems anymore. Or maybe he is still running BSD/OS out of a closet?

I just don't see how the article can draw a BSD-related conclusion.

[jggimi]

The article quotes Al Gillen, an IDC analyst who is looking at enterprise adoption rates of server systems. I can't find fault with his general statement on the Unix marketplace, wherein he states that with the exception of Linux on x86-based platforms, all Unix and Unix-like systems use is contracting in the marketplace....if you count physical platforms, of course. I expect virtualization in the data center is a major factor in this particular view of the marketplace, assuming the count is of physical platforms.

Unfortunately the article is short, and has only a small quote, taken out of its context from with a presentation or white paper, or perhaps from the midst of a telephone conversation. I would tend to think the complete picture of the OS marketplace from Mr. Gillen, in its full context, would be more readily accepted.

Disclaimer: Earlier in my career, I engaged with IDC analysts regularly. I both provided information for reports IDC published, and employed their analysts as consultants. I have never met Mr. Gillen, nor used his particular services.

[Oko]

I honestly think that the article is completely off base.

DragonFly has no enterprise user base that I am aware of (here on the South-East of U.S. nor on the South-West where I used to live until 3 years ago). Consequently, the user base can not shrink. Until 2.0 the project was not even usable beyond hobby/developer purposes. If Matt can further realize his ideas beyond Hammer I think that the sky is the limit.

OpenBSD is a network appliance and such has a very limited user base. Most very serious people in that business are fully aware of OpenBSD. Deployment is another thing. Very often at least here in U.S. business are very uneasy about deployment of the open source OS without a support (Linux i.e. RedHat is really commercial system just like Windows). In academia is different but even in academia administrative computing is often restricted by various state regulations (for instance we at the University System of Georgia have to follow various state regulations about commercial support and "approved' vendors) that mass deployment of OpenBSD is just not possible. Do we use it? Sure we do but do not on the mass scale.

If OpenBSD had more commercial support it could be widely used on the desktop and in particular on thin-clients. For serious scientific computing no. It is just not intended for that.
NetBSD for scientific computing yes but needs commercial support.

I am surprised that people didn't notice that OpenBSD amd64 now supports up to 64 cores with bsd.mp and in current virtually unlimited memory. It is not still as Solaris (I think supports up to 512 CPUs per mother board) and Linux which supports 256 per mother board but it is obvious that bsd.mp is the future.

By the way OpenBSD supports multiple processors on Alpha, hppa, macppc, sparc64, sgi, amd64, i386 now.

[rpindy]

I agree with the other posters. I know particularly with OpenBSD they aren't worried about attracting the masses but simply putting out a quality product. Ironically, if you look at the What's New for OpenBSD 4.9 there are a lot of major changes, such as the 64 cores in amd64 mentioned in the last post, multiprocessor support on hppa, and much more. In earlier releases you practically had to be a computer scientist to know what they were talking about in changelogs, but this is straightforward and the team did a great job.

http://www.openbsd.org/49.html

By the way, a commercial vendor with OpenBSD support would be nice.

[nilsgecko]

If OpenBSD had more commercial support it could be widely used on the desktop and in particular on thin-clients...in academia administrative computing is often restricted by various state regulations...'U.S. businesses are very uneasy about deployment of an open source OS without support '

I'm very curious about requirements companies or agencies tend to have which prevent these systems from being deployed. Are support contracts seen as a separate entity from the IT Staff who would otherwise be managing the infrastructure on a day-to-day basis? In other words, wouldn't a competent and qualified team be sufficient enough to have an organization deploy OpenBSD on a mass scale? Or are companies wary about deploying these systems because they don't feel there are enough qualified professionals to manage them?

Being as there are smaller vendors who provide commercial support, would they qualify as being an approved vendor or does vendor size matter? The Undeadly.org journal recently spotlighted a European-based OpenBSD solutions provider for example and here in the USA, iX Systems provide BSD Support as do various smaller consultancies/companies.

Also does this tie into what is called PCI compliance or auditing?
I'm just really curious about how this process works hence all the questions.

regards

[jggimi]

1) Some IT organizations require an outside support organization for all components in their infrastructures, by policy. It is perceived as liability coverage in the event of severe problems. The jargon for this is "having a throat to choke."

To me, this has nothing to do with infrastructure quality, or the mitigation of risk to business functions. Instead, this policy provides management with the ability to assign fault outside the IT organization, which mitigates risk to management careers.

2) Compliance to government regulation or industry standards, or infrastructure in support of a particular application may require specific and limited solutions, eliminating many options of free choice.

[Oko]

Originally Posted by jggimi

1) Some IT organizations require an outside support organization for all components in their infrastructures, by policy. It is perceived as liability coverage in the event of severe problems. The jargon for this is "having a throat to choke."

To me, this has nothing to do with infrastructure quality, or the mitigation of risk to business functions. Instead, this policy provides management with the ability to assign fault outside the IT organization, which mitigates risk to management careers.

2) Compliance to government regulation or industry standards, or infrastructure in support of a particular application may require specific and limited solutions, eliminating many options of free choice.

+1 for comment

Just to add one of my favorite phrases. In U.S. "we must have a legal entity that we can sue in the case something goes wrong". Whom you are going to sue if things go wrong with OpenBSD? Theo, Bob or OpenBSD foundation?