[wsfii-discuss] OpenHardware Meeting - Paris Dec.13th 2007

Thu Nov 15 11:57:06 UTC 2007

I wonder if we understand what this would really be able to achieve. 
First of all there is an existing reference design from Infineon for 
client units with one or two ports for WiFi cards, and the 
Intel/Marvell reference designs for four ports are doing well in the 
market already. What really comes into play is the volume of units 
built - we do short runs and so buy our boards rather than build 
them, but would build them ourselves if we had scale in that specific 
product. For instance, we would need about 30K units to even begin to 
make production worthwhile versus importing the boards from the 
factory that does them for us now.

In other words, generic boards exist now and we are using them and 
selling them to supported customer groups to "roll their own" or to 
kit together with firmware we provide.

We are meeting some executive level people from Senao at the end of 
this month and would bring up with them the entire concept, but I 
have the overall impression here that many Asian, and most in 
general, manufacturers only care about volume.

Another condition for success is the total parts for a board - some 
people will want an extended operating temperature range while others 
will not want to have the additional expense. If everyone simply 
would build their own boards - a la desktop fabrication - then the 
parts cost will be extremely expensive.

If the goal is to have large OEM outfits build this board, you can 
only expect one company will want to do it - as soon as there is a 
manufacturer with a lower cost base, it is not possible to compete, 
because there is no ability for better pricing by typical product 
differentiation.

At this point, I would be very interested to see a concept which 
would use the Sun open source design CPU, since that eliminates a 
massive cost driver. The other issue would be having Wi-Lan on board 
since they are suing everyone right now for the underlying enabling 
intellectual property. Unless someone can develop a new media access 
controller - which is hard on a small budget - and not use any third 
party technologies and then tie it to a good RF front end, then you 
have to use a chip set from Atheros, my favorite, Broadcom, or 
someone else. We priced the cost for delivering our own RF chip, and 
it would only be justified for commercial products where high 
performance would be met with a higher value in pricing as well. What 
would be available otherwise for WiFi chips? You need someone like 
Atheros or Qualcomm with its new universal platform for multiple 
systems or something like that. Developer kits are expensive and I 
have not seen any company offer up their chips and tools for less 
than what any given quantity can "justify".

After all, what would be the benefit for Atheros to open the kimono 
for an open platform? They are in so many products now, and also 
making money from that, so there is no advantage for them to truly 
participate at any advanced level.

Like I said, if this would allow us to use a "free" and open CPU like 
Sun offers then at least we would be able to create a platform and 
simply have folks plug in whatever card - WiFi, WiMAX, CDMA, etc. - 
they want to use.

Daniel

At 11:30 07.11.2007, you wrote:

>Hi all!
>
>We (Vic Hayes, Xavier Carcelle, Marek Linder, Florian Fainelli and 
>Juergen Neumann=me) are currently organizing an OpenHardware 
>get-together in Paris - http://www.openpattern.org . Our first goal 
>is to achieve cheap FOSS compatible open embedded wireless hardware. 
>As we all know the world needs cheep FOSS compliant wireless devices 
>(simple nodes), which come off the factory with e.g. OpenWRT or 
>similar OS. Second we need a modular more comprehensive wireless 
>router (super node) for backhaul infrastructure 
>(http://wiki.freifunk.net/OpenHardware and more IEEE conform: 
>http://openpattern.org/drupal/IEEE_HN1_Conf_OpenPatternProject_Article_071007.pdf)
>
>The meeting in Paris, which is a co-conference to IEEE home 
>networking 2007 http://home-networking2007.org/, shall be a first 
>official get-together of FOSS developers, OEM industry, wireless 
>activists, business people and FSF to have talks and workshops on 
>various issues on the road towards OpenHardware. It shall target the 
>following issues:
>
>1. What is open hardware in the sense of a licence?
>
>We have a very first draft on this topic: The term OpenHardware 
>refers to hardware, which is fully documented, with both 
>specifications and documentation freely available, including the 
>drivers and firmware required to make the hardware operational for 
>its principle task. The source code of the drivers and the firmware 
>as well as all items of the MIB should be directly available to the 
>software developers. The product with the drivers and firmware are 
>type approved globally and the purchasers are indemnified from any 
>Intellectual Property claims. It must be possible to write free and 
>open source drivers for this hardware without any restrictions or 
>limitations. Furthermore, the hardware design shall be available so 
>that anyone can access, modify and/or reproduce any of the hardware 
>components. A licence is needed to protect "OpenHardware"
>
>
>2. Why is there a global need for cheap open source hardware 
>(especially wireless routers)
>
>Here we want to point industry to the potential markets in Afrika, 
>Asia and even in USA an Europe. Over the last 5 years I have been 
>working with wireless people from all over the world. There are 
>wireless communities in almost every country and places on this 
>planet. They all seek for well designed cheap open embedded 
>hardware. FOSS compliant hardware would prevent us from all the time 
>and resources consuming reengeneering process a.s.o. There are 
>millions of potential customers for cheap wireless mesh devices like 
>accesspoints or wifi-phones, etc ...
>
>
>3. What are the benefits of OpenHardware?
>
>Today's hardware development is split into separate hardware 
>divisions across the industrial landscape resulting in individual 
>development branches and only little innovation. OpenHardware will 
>bring about dialogs between all parties involved -  the producers, 
>the customers and the consumers. By directly communicating with 
>customers / communities / free developers, producers can access 
>innovative input from outside the company itself on advices to 
>improve their products thereby working closer on the consumers needs 
>and ultimately widen their market. The customer could express needs 
>for adaptation of old products or even lead the way to the creation 
>of a totally new product altogether. These synergy effects 
>drastically reduce the research & development costs, increase the 
>number of potential customers and most importantly form the 
>fundamental basis for a level of customer satisfaction demanded in 
>the 21st century.
>
>
>4. Why is this interesting for OEM producers?
>
>Traditional OEM producers depend on external purchasers which 
>commission a product to sell it by themselves via their established 
>distribution channels. After finishing the production the result is 
>handed over and the OEM has to find a new purchaser. OpenHardware 
>designs would fill the gap between the producer and the actual 
>market, by providing the necessary designs which can be produced and 
>shipped without an external purchaser. The combination of modern 
>marketing through Open Source Communities and customer networks 
>reached over the Internet are the key to selling these products 
>around the globe.
>The OpenHardware concept would allow for one product being adapted 
>for several thousand completely different needs, as it could be 
>shipped in parts and later be assembled at its final location, after 
>being thoroughly customized to the situation at hand. OEMs could use 
>the feedback and the contributed efforts to further improve their 
>products and claim a strong share of the market. Furthermore, the 
>community formed around an OpenHardware product could serve as a 
>pool for the recruitment of capable and willing developers/testers 
>or distributors who already know the product, thereby saving a lot 
>of time in the initial and crucial phase of employment. Also would 
>the customer bonding to the company increase dramatically, as the 
>product is no longer developed for the consumer but with and most 
>importantly BY the consumer.
>
>
>5. New channels and ways of distribution
>
>When you buy a Linksys WRT54GL in India or Afrika it is about twice 
>as expensive as if you buy it in the US or Europe. Though produced 
>in Asia for less than 10 Dollars, it is shipped and brandend into 
>the western world, before it can be redistributed to Asia or Afrika. 
>In the west it's at about 50 EUR and back in Asia it's about 100 
>EUR. That is stupid nonsense, but has it's roots in the trading 
>tradtions formed over the last centuries. New e-business plattforms 
>(webshops) and global logistics offer much more inexpensive and 
>direct distribution channels to almost every place in the world. 
>Community needs on a product are not so much about well designed 
>boxing and accessoires, but much more about bulk boards packed in 
>charges of 20, 50, 100 or even larger. Following these ideas it 
>should be possible to sell wireless equipment for about 10 to 15 
>Dollars all over the world, but especialy in the developing 
>countries. That's what we want to achieve.
>
>
>Of course I know that we cannot reach all these goals in just one 
>meeting, but I am very sure that Paris has the potential to be a 
>first important step in this direction. Also there will not be 
>OpenHardware until we have open chips. So we need to see it as a 
>phased model. Step one is cheap FOSS compliant wireless hardware off 
>the factory.
>
>All help and support is welcome! Please spread the word and give 
>me/us your feedback to this first drafted outline.
>
>Kindly,
>
>Juergen
>
>
>_______________________________________________
>wsfii-discuss mailing list
>wsfii-discuss at lists.okfn.org
>http://lists.okfn.org/mailman/listinfo/wsfii-discuss