[Discuss] different perspectives on open source hardware

Matt Maier blueback09 at gmail.com
Fri Jan 31 23:31:04 UTC 2014

The Open Compute Project is a good example of something I ran across while
working on my masters, which is that "open source" is a framework that
sometimes allows proprietary entities to cooperate better. That should be
contrasted against the typical community point of view that open source is
a way for people who believe in sharing first, or at least in the selfish
benefits of metered sharing, to take each other's work to new heights.

"In general, standards appear in technology when it has matured to the
point that the competing vendors can no longer differentiate on the basis
of features or performance. At that point the only source of
differentiation left is price, and the vendors end up locked into a "race
to the bottom." The technology becomes a commodity that provides razor thin
margins, and only the lowest cost providers can survive...So the question
is, can standardization on one or more open source architectures that
benefit from development by a large community rather than the much smaller
groups of developers that individual companies can summon, provide more
benefit than those companies can realize from proprietary products. The
huge and growing market for white box servers, which by definition have no
proprietary identity, says the answer is yes."

In the world of million dollar development budgets, cooperating is often an
attractive idea because everyone wants to save their money. It's not all
that unusual for competitors to try out some kind of cost-sharing
arrangement where they combine their development forces. But those
arrangements typically work out poorly. In the software domain, open source
code has proven to be an excellent substitute for cost-sharing arrangements
in many cases. There are fewer project-ending problems.

The Open Compute Project is using "open source" in that sense. It's a
big-business perspective. They see that some of their line-items are pretty
much already standardized, so the differentiation is becoming irrelevant,
so it makes sense to pick one and move on to more valuable activities. If
they succeed, their industry will gain the ability to order the same
hardware (&software) from several suppliers. Then they can focus on areas
where differentiation is still possible.

This is "open source" in the literal sense because they're using, or at
least intend to use, "open source" licenses. Although at the moment their
primary license seems to be CC Attribution. It is not "open source" in the
community sense. At least, if it is, then it applies to a very select
community of people who manage and fund huge data centers. It is unlikely
that the community will grow beyond that because the project is less about
innovation and more about optimization...for an extremely demanding

But the Open Compute Project is still adding to the PUBLIC pool of
knowledge/education, which is awesome and should be encouraged. If they can
commoditize high-performance/low-cost data center technology then maybe in
a few years we can have mom-n-pop ISPs cutting up the cable monopolies. Or
maybe the "personal cloud" will actually start to make sense. Probably
something I haven't thought of.

Most of y'all can stop reading now. The post script is speculation about
how the subject relates to export control.


PS. An interesting (to me) angle on this situation is that I wonder whether
or not the government is going to try to add data center and/or server
technology to the USML. Not right now, obviously, but there are trends that
could move in that direction.

The military has already stood up "cyber" as a totally different domain of
warfare that demands its own approach. There is also a never-ending battle
over who has the best encryption, which largely boils down to who has the
best computers. Back before interneting (yeah I made that up just now,
because I'm on the internet) was called "surfing the web" the government
considered cryptography a national security domain because it was so
important during the previous couple wars; so it went onto the USML early.
Nobody was thinking along the lines of data centers or super computers
being anything like cheap enough to proliferate. We haven't had a cyber-war
yet, but there have been enough skirmishes that a lot of people are
worried. When we do eventually get into a large scale conflict it will
inevitably involve the cyber domain in a way that no previous war ever did.
The technology being developed in the Open Compute Project would arguably
be equivalent to jet engines, or radar, or mechanized transportation in its
ability to give one side an advantage.

I'm working with Mach 30 on the subject of space technology, which is
already on the USML, because of ICBMs and space-based nukes and whatnot. So
trying to do open source work in the space domain is already a challenge
because that technology already demonstrated its military potential before
the USML was written (the same can be said for cryptography and firearms).
But data center technology matured after the list was written, and by the
time it gets a chance to demonstrate its military potential some of it,
possibly a large percentage, will already be open source. If the government
decided to add it to the list of things that deserve export control, what
would that mean? The same question could probably be asked of drones, which
are in a similar position, although they are further along in demonstrating
military value. Open source biotechnology (bacteria, viruses, etc) is less
far along, but it's getting there.

How would someone adapt export control, or the government's interest in
maintaining a military advantage in general, to a world where open source
is a widely successful development methodology? Arguably there are some
technologies that are too dangerous to let just anybody have access to. And
some of those technologies haven't been invented yet.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.oshwa.org/pipermail/discuss/attachments/20140131/0c1eec6b/attachment.html>

More information about the discuss mailing list