Neither iPhone or OpenMoko are revolutionary
thomas.cooksey at bt.com
thomas.cooksey at bt.com
Thu Jan 18 17:53:16 CET 2007
I work for BT, who have been trying to get WIFI VoIP working for _YEARS_. It's very, very difficult and has lots of problems:
1) As several other people have mentioned, the first problem is coverage. A typical access point will only have 50m - 100m of coverage.
2) There's no guaranteed RF bandwidth (On GSM you get a dedicated chunk of bandwidth just for you which no-one else can use). On Wifi, everyone shares the available RF bandwidth, which is usually <11Mbs.
3) The backhaul network (i.e. the internet) is fairly unreliable as far as VoIP is concerned. GSM networks usually have a circuit switched backhaul like SDH, so again, guaranteed bandwidth. There are IP QoS solutions out there, but there difficult to implement well. Having said all that, it is far, far better than it was.
4) Most access points are usually connected using ADSL. The 'A' bit of ADSL is the problem here. VoIP requires a lot of upload bandwidth, which ADSL doesn't provide.
5) Most access points have firewalls meaning VoIP has to tunnel through them using HTTP, confusing any backhaul QoS (How can the network tell the difference between your time-critical VoIP call and someone downloading a 4GB DVD image?)
6) Again as people have pointed out, a roaming Wifi contract is very, VERY expensive (e.g. £40 per month)
7) Power. Even the low power wifi chipsets are very power hungry. I guess you could argue that during a call, GSM draws a fair bit of power but it's nowhere near that of a wifi chipset. GSM actively changes its power output so it only uses what it needs to reach the base station. Although the more advanced Wifi chipsets can also change their power output, it's not done very well and typically only have 3 power settings (including on and off!). I use Marvel's 88W8385 chipset which is designed specifically for mobile applications and have been very disappointed with its power consumption. I guess with a regular Lithium Ion/Polymer phone battery you'd be lucky to get 30 minutes of talk time. When it comes to standby however, the power requirements of wifi don't drop too much whereas GSM drops to almost nothing. GSM Idle really is amazing. All a hand unit needs to do is transmit a short keep-alive every few minutes to let the last station know it's still there.
Like I say, BT has been working on this for ages and DO have a solution: BT Fusion (http://www.bt.com/btfusion/). It uses your own home/work access point if your in range, then switches to BT OpenZone access points, then drops to GSM if your not covered by WiFi. I think there's also some magic to move you onto GSM if you move out of range of Wifi DURING a call.
BTW: I have absolutely nothing to do with Fusion development and am probably wrong about all of its features. This information is my own personal understanding of the technology and is not supported by BT whatsoever.
From: community-bounces at lists.openmoko.org [mailto:community-bounces at lists.openmoko.org] On Behalf Of Andreas Kostyrka
Sent: 18 January 2007 16:05
To: Renaissance Man
Cc: community at lists.openmoko.org
Subject: Re: Neither iPhone or OpenMoko are revolutionary
* Renaissance Man <renaissanceman at macmail.com> [070118 12:36]:
> On 18 Jan 2007, at 10:22 am, Attila Csipa wrote:
> >On Thursday 18 January 2007 10:01, Renaissance Man wrote:
> >>The problem with the Nokia E Series, N80s, and Windows smartphones is that they're either very expensive and/or they don't actually make VoIP via WiFi easy.
> >Why should they risk? They are selling millions of handsets through carriers, and they sure don't want to lose those contracts.
> And, as is often the case, someone else's risk is another's opportunity.
> All your arguments against WiFi on the Neo seem a little moot, as it's pretty clear from what people are saying that it will have WiFi; it's just a matter of time.
> I just wish it had been on the first model because I would have had all my needs fulfilled. As it stands I'm in the market for something else now, and may even end up with an iPhone if
> Apple includes VoIP via WiFi before OpenMoko. I wish this not just for my own selfish reasons but because I'd like to see an open product like OpenMoko bet out a closed product like
I don't think so. You seem to have missed the detail, that iPhone will
be offered this year on one (1!!!!) US network this year. They don't
even say if it's GSM or not. (Probably.) They have not yet told with
which networks they intend to cooperate in Europe. Even if it's GSM,
there is no telling if they will work on other GSM networks:
*) frequencies are different in the US as in Europe.
*) the data part of the phone might depend upon network details. E.g.
a Sidekick won't work on a different network. Well the voice part
will, because they tend to be not-simlocked, but all the advanced data
services won't, as the exact configuration is burned into the
firmware, which is carrier specific.
So the chances are good that OpenMoko phones will have WiFi before the
iPhone will be available in the UK are pretty good *g*
OpenMoko community mailing list
community at lists.openmoko.org
More information about the community