asterisk users - Jun 2004 - Cost of IP Phones, or Isn't It Just Software?

I took a little foray into pricing out IP Phones for my home pbx
yesterday.  $75-$750 seems to be quite a range, so I took a closer look.

Cisco, for example, has different models such as the 7940 and 7960 which
seem to only differ in the software.  And buying a Cisco 7920 should
cost you $500 w/o a license, and $200 more with a license?  It seems to
me that these prices are substantially inflated.

Just to get an idea of hardware cost involved here:
- I can buy a 4-port router with built-in firewall, web-server and
email-client for $20-$30 RETAIL.  That would indicate a hardware cost of
$10 max.
- I can purchase a Sipura SPA-2000 for $100 -- actual hardware cost
should be $50-$75.
- I can also purchase a fully featured ADSI speakerphone for $80 retail,
with an expected hardware cost of $50-$60.

Combine all these pieces for hardware cost of $110-$145 -- I'd think
that synergies would push hardware cost under $100 -- and you have all
the hardware required to build at least a 2-line IP speakerphone with a
nice large display, webserver for config, and enough processing power to
run some advanced functionality.  You could even add one of the Yamaha
sound-chips for downloadable ringtones if you so desire.

The 2-line restriction would be purely theoretical, allowing for 4, 6,
or even 12 lines to be registered (who really needs more than 3 or 4?).
The device could be well documented, opening the door/port for
open-source software.

Does this sound to utopian, or do my fellow list-members think there is
an idea here?  How many more PBXs would you integrators sell if the cost
was down to $150/station for a business-class phone?  How many more
features could be implemented with an open-source UA?  (Menus, Visual
Voicemail, extended CallerID info, Call Delegation, Queue handling,
Email, Weather, Reminders, .... )

Could this be profitable?  
- You bet, I'd guess you could sell a few 100,000 of these devices in
the next three years.  The design would be open, but protected from
clones by virtue copyright law (bootloader and operating system could be
proteges).  A public company would own the design and contract with a
manufacturer.

Could this be financed?
- I don't think it would take that much -- maybe $10K - $20K to purchase
samples and development hardware and software.  Engineers could donate
time in exchange for revenue shares later.  A small investment would be
counted toward a purchase of a finished product, a large investment
would buy you a share in the company.  200 active members in this list
donating $100 each could get a handful of engineers on their way.

Where do we get the know-how?
- Partner with existing companies like Digium and Sipura.  Once the
project is complete, they would receive revenue shares and/or utilize
their manufacturing resources.

Am I dreaming?

>
>
>Am I dreaming?
>
Yes.

Community based development is too unreliable.

Just to refer to ongoing projects... Look at the farfon 
(www.farfon.com), It's an active project in the final stages of development.
It offers the benefits (modular, programming of your own features, 
quality components, low price, corporate backing, designed by skilled 
engineers...) without the negative sides of your plan.

> Just to get an idea of hardware cost involved here:
> - I can buy a 4-port router with built-in firewall, web-server and
> email-client for $20-$30 RETAIL.  That would indicate a hardware cost
> of $10 max.  
> - I can purchase a Sipura SPA-2000 for $100 -- actual hardware cost
> should be $50-$75. 
SPA 2000 hardware cost is much less then that!!!
> - I can also purchase a fully featured ADSI speakerphone for $80
> retail, with an expected hardware cost of $50-$60. 
> 
Same here...
> Combine all these pieces for hardware cost of $110-$145 -- I'd think
> that synergies would push hardware cost under $100 -- and you have
> all the hardware required to build at least a 2-line IP speakerphone
> with a nice large display, webserver for config, and enough
> processing power to run some advanced functionality.  You could even
> add one of the Yamaha sound-chips for downloadable ringtones if you
> so desire.      
>
Well.. It still depends of what cup, dsp etc are included..
 
> The 2-line restriction would be purely theoretical, allowing for 4,
> 6, or even 12 lines to be registered (who really needs more than 3 or
> 4?). The device could be well documented, opening the door/port for
> open-source software.   
> 
> Does this sound to utopian, or do my fellow list-members think there
> is an idea here?  How many more PBXs would you integrators sell if
> the cost was down to $150/station for a business-class phone? 
This is a certainly a question needing an answer! However, lets see if
anyone will actually do so?

 How
> many more features could be implemented with an open-source UA? 
> (Menus, Visual Voicemail, extended CallerID info, Call Delegation,
> Queue handling, Email, Weather, Reminders, .... )     
If it too complicated.. End users may be reluctant to buy it!
> Could this be profitable?
> - You bet, I'd guess you could sell a few 100,000 of these devices in
> the next three years.  The design would be open, but protected from
> clones by virtue copyright law (bootloader and operating system could
> be proteges).  A public company would own the design and contract
> with a manufacturer.    
Well.. This is REAL problem.. I am sure that a lot of people on this
list have similar ideas to yours.
It would be very nice if we all could get together to produce a solution
like this... However, time will tell!
 
> Could this be financed?
> - I don't think it would take that much -- maybe $10K - $20K to
> purchase samples and development hardware and software.  Engineers
> could donate time in exchange for revenue shares later.  A small
> investment would be counted toward a purchase of a finished product,
> a large investment would buy you a share in the company.  200 active
> members in this list donating $100 each could get a handful of
> engineers on their way.      
> 
Well, said... $100 each would take project to "somewhere".. However,
the
costs may spiral out of control!
> Where do we get the know-how?
> - Partner with existing companies like Digium and Sipura.  Once the
> project is complete, they would receive revenue shares and/or utilize
> their manufacturing resources.  
Yes, that would be good way... But what then? Who owns it? Who profits
from it...
A lot of questions needing an answer!
 
> Am I dreaming?
No, you are not dreaming!  It can be done! One just needs to find
correct way!
I personally would love to see a very low cost embedded board running *,
with FXS/FXO port on it!

At 1:18 PM -0500 on 6/16/04, Jay Milk wrote:
>I took a little foray into pricing out IP Phones for my home pbx
>yesterday.  $75-$750 seems to be quite a range, so I took a closer look.
>
>[snip]
>Am I dreaming?
>
Well, yes, you're sort-of dreaming.

The trick is not designing the hardware or the software - anyone with 
$100k (or much, much less) and the right engineers can get something 
working to the point where it is ready to be produced.

You will hit the wall with:
   - finding reliable suppliers of manufacturing technology
   - finding enough money for cash "float" during the sales cycle
   - finding enough money for marketing "float" during sales cycle
   - finding adequate sales channels
   - compensating your sales channels

The last two are the most important.  These are not engineering 
questions; they are business and political questions, which is why 
the hardware market looks so easy but is so littered with the dead 
husks of failed companies run by engineers.  Selling direct is a 
limited market; there are only so many Asterisk home users that you 
can advertise to via the mailing lists.  :-)

JT

On Jun 16, 2004, at 11:18 AM, Jay Milk wrote:
> Cisco, for example, has different models such as the 7940 and 7960 
> which
> seem to only differ in the software.
IIRC, the 7940 and 7960 run the same software, but differ slightly in 
hardware.  The 60 has 6 line appearance buttons, while the 40 has 2.  
I'm not sure quite how they manage to charge $25/button for the 
upgrade. :-)
> Does this sound to utopian, or do my fellow list-members think there is
> an idea here?  How many more PBXs would you integrators sell if the 
> cost
> was down to $150/station for a business-class phone?  How many more
> features could be implemented with an open-source UA?  (Menus, Visual
> Voicemail, extended CallerID info, Call Delegation, Queue handling,
> Email, Weather, Reminders, .... )
Well, the big problem with all of this is the development costs.  Your 
$30 router is probably the 5th or 6th nearly-identical product the 
company has made, and they've probably sold millions of them.  I doubt 
most VoIP phone vendors sell more then 100,000 of even their most 
popular models, and phone vendors are just starting to be able to share 
platform code between models.

Odds are, we'll see the phone that you're looking for, but it'll
take
another three years to show up, and it'll probably come from the same 
companies that make your cheap routers and cordless phones 
today--D-Link, Linksys, VTech, Panasonic, and whoever OEMs the home 
phones that AT&T and SBC sell by the boatload.  It'll take a while for 
commodity VoIP phones to appear.  I mean, even 12 months ago, it wasn't 
completely clear that SIP was going to win.  And you can't have 
commodity hardware without ubiquitous standards.
> Could this be financed?
> - I don't think it would take that much -- maybe $10K - $20K to 
> purchase
> samples and development hardware and software.  Engineers could donate
> time in exchange for revenue shares later.  A small investment would be
> counted toward a purchase of a finished product, a large investment
> would buy you a share in the company.  200 active members in this list
> donating $100 each could get a handful of engineers on their way.
My gut instincts say that you're low by at least an order of magnitude. 
  The hard part isn't really designing the basic code.  The hard part is 
building the user interface and doing all of the design work.  I mean, 
go take a look at really good consumer electronics devices (TiVo comes 
to mind).  Spend some time looking at how well they work, and then 
compare them to even the best OS interfaces (KDE?).  You can't really 
design good interfaces by committee.

On the other hand...  Go take a look at all of the ~$100 wireless 
router/firewall/print server/gateway boxes on the market, and you'll 
see one thing that almost all of them have in common: they all run 
Linux.  Most of them are even based on the same small number of tools; 
things like busybox and uclibc.  If you want to see cheap, powerful 
VoIP phones, think about what they really need in terms of software, 
and then set out to write it and license it so the phone companies can 
incorporate it into their products.  I'm kind of amazed that FXS ports 
aren't standard on medium-end home routers right now; they'd probably 
only add $5-10 to the cost of the router, *IF* they had the software 
and felt like the demand was there.


Scott

> -----Original Message-----
> From: asterisk-users-admin@lists.digium.com
[mailto:asterisk-users-
> admin@lists.digium.com] On Behalf Of Michael Sandee
> Sent: Wednesday, June 16, 2004 8:45 PM
> To: asterisk-users@lists.digium.com
> Subject: Re: [Asterisk-Users] Cost of IP Phones, or
Isn't It Just
> Software?
> 
> 
> 
> Am I dreaming?
> 
Firstly one would have to wonder if Digium will be
taking the next step to produce a handset based on
their, yet to be released, IAXy. The IAXy appears to
be a possible core just add the keypad and handset. 

Secondly IF a device could be built there MAY be
business models that wouldn't need centralised sales
and marketing budgets. For example many people on this
list are running VOIP businesses where it MAY make
sense to give CHEAP handsets away in order to gain
subscribers and then recoup the costs from call or
subscription fees.

Thirdly what sort of expertise would the group be able
to pull together? Software seems to be the main
competency and hardware is a different kettle of fish.
That being said these days a lot of consumer hardware
is going the way of reference designs (broadcom and
WiFi routers for example) with OEMs differentiating on
the software or even just the user interface. IF a
good hardware reference design was available for a
VOIP handset software would POSSIBLY become the
primary task and therefore POSSIBLY suit the expertise
of this group...

Regards,
Aaron





	
		
__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
http://promotions.yahoo.com/new_mail

Hi
> Just to get an idea of hardware cost involved here:
> - I can buy a 4-port router with built-in firewall, web-server and
> email-client for $20-$30 RETAIL.  That would indicate a hardware cost of
> $10 max.
> - I can purchase a Sipura SPA-2000 for $100 -- actual hardware cost
> should be $50-$75.
> - I can also purchase a fully featured ADSI speakerphone for $80 retail,
> with an expected hardware cost of $50-$60.
But you fail to discuss the most important fact.  Cisco charge USD 500 for a
phone simply because they can do so :)  If you look at the general pricing -
go back 2 years and you could hardly find a VoIP phone for less than USD
250.  Enter Grandstream and prices started dropping.  When Grandstream first
appeared they listed the price as USD 79.  Now they are generally selling at
around USD 50-60 or so (single pcs. end user pricing.  It's now possible to
get a lot of IP phones for less than USD 100 - which is about the price that
I and most people find reasonable to pay for them.

You see the same with the WiFi phones right now.  The first Cisco one was
what - 6-700 USD (I never even bothered to check the exact price).  The
second one - the ZyXEL/Pulver thingy - well - they generally seem to sell
now at around USD 200 or so.  The Senao one is going to hit the market in a
few month and they claim to target around USD 120 for the initial offering.
I'll bet you anything that shortly after the ZyXEL's will be around USD
100
and they'll both end up competing around that level.

Anyway - I'm babbling now :)  My point was only that the price of these
gadgets have next to nothing to do with the actual hardware cost - but is
only controlled by what the market is prepared to pay.

Regards,

	Lars...