[Freedombox-discuss] Iomega iConnect lessons for FreedomBox
John Gilmore
gnu at toad.com
Thu Mar 24 02:46:32 UTC 2011
Consumer electronics maker Iomega is shipping an "iConnect" box with
similar electronics to the ShevaPlug. But unlike the GlobalScale
team, they seem to have done some marketing. We can learn from them:
http://go.iomega.com/en-us/products/network-storage-desktop/wireless-data-station/network-hard-drive-iconnect/
It has the same hardware inside it as the ShevaPlug, basically, but
they thought about what people might want to use it for, and aimed the
hw/sw design at a particular market: cheap NAS and printer sharing.
In every other product, NAS (network addressable storage) always seems
to cost a minimum of $1000 more than the disk drives; here it's $65
more. I bet they're selling more of them than the "plug computer"
guys. It's cheaper than the plugs, has WiFi-N, and doesn't overheat.
You can get it from Amazon for $64.29 with free shipping:
http://www.amazon.com/dp/B0035JCI6M/
There are 92 customer reviews in Amazon, averaging out to 3 stars out
of 5. But the reviews are instructive. They're quite skewed: 30
5-star reviews, 35 1-star reviews, and only 9 each for 2, 3, or 4
stars. It seems likely that:
* They had some early infant-mortality problems that would hang the
box with a blinking blue light. Later firmware, or perhaps
improved manufacturing QA, have reduced this.
* People who want to use it for wired disk sharing (NAS) tend to
love it, though they complain about a few bugs or limitations,
such as with multiple partitions on the attached drives. Its
speed is limited by the USB2 interface to the drives, and by its
CPU, which both seem to be slower than the GigE. Mac-formatted
drives tended to get corrupted. Time Machine worked fine for some,
never worked at all for others, worked initially for some and then
reported drive corruption within a week.
* People who used its WiFi tend to be disappointed. For many it
didn't work at all, wouldn't even connect to their network. For
others, they expected high speed and didn't get it.
* People who wanted to share a printer over it tend to lose. It seems
to only support a limited set of printers.
* Windows clients are better supported than Mac; Mac better than Linux.
Though all are supported, some are more equal than others, and the
documentation was extremely limited and poor. Blog entries from
other users provided better documentation than the official stuff.
* Iomega's phone/email tech support was not helpful to most
commenters, often leading them to give up and return it (to
Amazon for a refund), or throw it away.
Here's a thorough press review of the iConnect, by a NAS guy who also
knows the guts. Note he takes major points off for the UX of the
wireless (no LED, clunky config interface, especially compared to the
rest of the config). But, for him it WORKS with wireless, out of the
box, which many NAS's don't:
http://www.smallnetbuilder.com/nas/nas-reviews/31094-iomega-iconnect-wireless-data-station-reviewed
The Freedom Box project can learn tons from this:
* To avoid consumer frustration, aim it to do one thing well, in one
environment. Document that thing very well, up front, so that
people who try it for something else will be pre-warned that they
are unsupported experimenters. Provide detailed documentation
for that one thing in all supported environments, and evolve the
doc to cope well with frequent misunderstandings or failure modes
among users.
(When you've mastered that one thing, you can make the box do a
second useful thing. Before then, keep the team focusing on
making that one thing work flawlessly for normal people.)
* The box must be *useful* and *cheap* at doing that one function,
to attract adopters. A box that offers encrypted chat only to
others who have a box, isn't going to sell to any ordinary folks.
One that serves up disk drives on their network might. This
"first niche" is something we'll have to think hard about,
because it will make or break our box. Ideally it'll do
something that nobody else offers. (Perhaps secure wide-area
network filesystem support, using NFSv4.1 or the Andrew File
System?)
* The more things that can "plug in" to your box, the more
configurations are likely to fail. A box with a disk drive
inside it, with a known filesystem format, will be easier to make
reliable than a box that the user can plug any disk drive into,
with any PC, Mac, Linux, or USB-key filesystem format on it.
* Wired Ethernets are much easier to support than wireless (for
initial installation, ongoing debugging, and for product
performance). Ethernet has had 40 years of engineering that have
made it almost impossible for consumers to screw it up these
days. All the common failure methods have gradually been
engineered out, from shorts in sting-taps, to missing terminating
resistors, to loops in the network, to mismatched port speeds, to
cross-wired cables. You can see it work (via the link/activity
LEDs in the jacks) and when it fails, those physical indicators
usually let you swap cables or switches til you isolate the
problem. With WiFi, it's magic when it works, and magic when it
fails; there's no obvious debugging path.
* Solid tech support will be needed early, to avoid an early flood of
frustration and rejection among ordinary people trying to install it.
* If your hardware isn't rock-solid, or your firmware isn't rock-solid,
then don't ship it to ordinary people yet.
John
More information about the Freedombox-discuss
mailing list