[maemo-users] Nokia device usage

From: John Holmblad jholmblad at acadiasecurenets.com
Date: Mon Mar 9 00:39:39 EET 2009

I have not read the HSDPA spec myself but my assumption is that it is 
"all packet, all the time"  As the "P" implies in "HSDPA" 

UMTS, as a superset of HSDPA incorporates  various  QOS  and other 
features (roaming, voice connection awareness, etc) that are needed in 
order to properly handle voice traffic as a part of the total packet 
stream on a UMTS network. As I understand it, the key difference between 
voice traffic on a 3g network vs voice traffic on a 4 g network such as 
one based on,  LTE, or  WIMAX  is that on a 4 g network the network core 
is packetized and runs IP (or at least an connectionless packetized core 
with a network protocol that looks like IP) with QOS enhancements for 
the voice packets. On the other hand, on a 3 g network with HSUPA or 
HSDPA, the voice traffic, although it may be statistically multiplexed, 
is not competing for the same bandwidth as the HSUPA/HSDPA packets. 
Perhaps a 3G expert on this list could clarify this somewhat subtle 
point. if not I will dig into the spec myself to figure this out.

I have always (professionally speaking) considered voice as  a 
particular case of data (bits if you will) with specific 
characteristics, i.e. connection oriented, duplex content flow, 
streaming, delay sensitive, and relatively narrow bandwidth (i.e. 
frequency limited) per "conversation"/connection.

My assumption is that most GSM mobile network service providers who are 
evolving to 3G are building HSDPA overlay networks  to first (and 
primarily) convey non-voice data (by my definition, above, not an 
oxymoron)  just as the CDMA based mobile network service providers have 
built EVDO overlays to do the same thing. In other words those GSM 
service providers are not attempting to move their voice traffic to UMTS 
right away.

A key difference  between HSDPA/UMTS on the one hand and EVDO/CDMA on 
the other (in terms of application of the underlying technology)  is 
that, in the U.S. at least the CDMA/EVDO providers (the largest being 
Verizon and Sprint) are not themselves using EVDO to convey voice 
although the end user of EVDO service could certainly do so (e.g. SKYPE 
or arrangement with a SIP trunking provider). They are using EVDO 
service to meet new and growing demand for broadband wireless Internet 

On the other hand mobile service providers who are evolving from GSM to 
3G/UMTS can, if they so choose, start to move their voice traffic over 
to their UMTS infrastructure (equipment and RF) and do so gradually by 
providing their customers with dual mode 2g/3g handsets. Although the 
UMTS standard, supports call handoff from GSM to UMTS, I have to wonder 
how much of that is actually going on right now since the user would 
have a dual mode (GSM + 3G) handset and the network would have to be 
engineered to implement such inter-technology (GSM<=>3G) handoff/roaming.

Best Regards,


John Holmblad


Acadia Secure Networks, LLC

* *

kenneth marken wrote:
> John Holmblad wrote:
>> Andrew,
>> yes, I am being overly presumptuous as to what kind of radio 
>> technology will and will not be in the next turn of the IT hardware. 
>> I must have read it somewhere that it was going to be HSDPA only.
> HSDPA do not result in data only, as HSDPA only builds on UMTS. and 
> UMTS carry voice just fine...

More information about the maemo-users mailing list