ARM as a primary architecture

[Chris Tyler]

On Thu, 2012-03-22 at 16:12 +0200, Nikos Roussos wrote:

> 
> Despite the fact that this is pretty wild guess, I don't think that
> Fedora should only care about what's happening on the developed
> countries and ignore the rest of the world.

I don't think this is necessarily a developed-vs-developing thing. I did
a talk on this at FSOSS 2011, suggesting that we're heading towards
three physical tiers:

(A) Content-consuming platforms - tablets/MIDs/smartphones. Almost all
of these are taking a tablet-like form factor (counting most smartphones
as 'tiny tablets'). These are great for consuming content - reading,
surfing, watching, playing - but they are awful for creating content --
they don't have the horsepower, storage I/O, screen real estate, or
long-use-comfortable UI devices for writing long documents, editing
video, or composing multitrack sound. Most of these devices run ARM for
low energy consumption, small size, and low cost, and many run some kind
of a Linux system (even if Android).

(B) Content-creation platforms - desktop/laptop. These devices aren't
going away anytime soon, because there are a lot of people who create
content or capture information -- students, bookkeepers, bloggers, video
creators, developers, documentation writers, animators, graphics
artists, engineers, secretaries, and so forth. A lot of the Fedora
community falls into one or more of these categories, so any
navel-gazing sample we take will be skewed toward this category. To be
frank, Intel x86, proprietary operating systems (Windows and Mac OS/X),
and proprietary apps still reign in this space. However, most creators
consume a lot more than they create, so we will increasingly find that
people who have devices in this category also have a device in category
(A). This category is waning as many people realize that they are almost
entirely content consumers and re-equip themselves with devices in
category (A).

(C) The Cloud - whether private cloud or public cloud, the backend lives
here, and this connects the content creators to the content consumers.
Space, heat, and power costs and scaling issues are driving cloud
operators to look for new solutions. ARM is looking like a pretty good
option but has not yet been proven in this space. Projects that have a
legacy component often use proprietary software here, but the majority
of new cloud projects use open source for a significant part of the
stack.

So of the three categories, two are on the rise, and one is on the
decline. The ones on the rise are categories where ARM competes well and
open source software thrives. If we want to remain relevant as a
community and project, we need to adapt and move our focus to these
emerging/rising categories -- hence the focus on Fedora ARM.

-Chris