Do you really like this stuff?

(a provocative post for my friend Andrea)

Do you?

 

We know UMTS CS uses almost the same codecs as GSM. On the air interface voice codecs streams use data rates between 12–28 kbit/s, depending on the spreading factor.

Since 3G launch, in order to switch towards the full packet switched world, several techniques based on complicated “real-time rate adaptive dynamic switch etc.. etc.. codecs ” have been studied and released for voice coding.

They are all meant to allow operators to reduce the impact on network capacity.

4th generation codec rate adaptation  for VoIP over LTE has been clearly defined in the 3GPP Rel-9 specifications. The use of VADs and CNG has contributed to reduce bandwidth requirements as well.

To make it short, air interface bandwidth required by new VoIP codecs is really low; AMR WB (which is like “HiFI voice call”) can reach 6,6 kbps only.

Not so much considering the throughput evolution curve (see figure below from AT&T, taken from http://www.atis.org).

 

I believe we are not that far to say that VoIP on wireless can be optimized and shrunk enough so that full packet switched networks can think to manage it in terms of capacity in the next years (providing making some CAPEX investments on backhauling).

The fact is that we shouldn’t simply think of a block migration of voice from one technology to the other.

Let’s consider a couple of other aspects, not strictly technology related.

Are packet users more hungry of conversations using mouth or using fingers?

Look at this interesting Nielsen report.

Now, supposing packet based wireless networks have sufficient bandwidth to carry IP data also for mobile IP voice in indoor coverage, if you had to choose between using the terminal for IP voice over fixed lines and you mobile, which terminal would you use? As for myself, if the QoS is good enough, I’d definitely go for the mobile wireless.

So, from one side we can think of voice over mobile decreasing, both for social behaviours and for very low prices of VoIP over fixed lines. (I see young people not so interested as in the past in talking on the phone. They can sit next to each other and digitally chat. Just spend a few hours in a park close to secondary schools and you’ll see that yourself).

On the contrary, people will increase working on the move (even in offices, see people running up and down in open spaces), so if you could make cheap VoIP also on your mobile, why should you sit and pick up a fixed line phone?

The decrease of voice from teens can somehow compensate the increase voice on mobility of business?

If I were a teen I’d  probably have at least 5 different free VoIP mobile apps on my device , each implementing a set of sophisticated coding.

If I were a business man, in the future era of tiered plans,  I’d join a “platinum” data plan with fantastic voice quality “network allowed”.

Are these fixed line stuff on desk still going to survive for a long time?

 

4G 4Everyone

For those who wanted to dig into LTE I would recommend some good blogs:

1) http://ltewatch.blogspot.com/

This is for who wants to stay up to date with the roadmap and with what major player are releasing on LTE.

2) http://3g4g.blogspot.com/

I like this one as it often refers to other blogs which cover most specific part of the subject.

 

Finally, this one gets deeper into protocol and specs but sometimes it could be better to read specs directly as it seems to paste part of them quite often: http://www.3g4gtech.com/

Note: For those who can understand swedish there’s the chance for some drive test results too: http://4g-patrullen.se/

 

Pipes strike back

Verizon wireless announced they will be introducing tiered data plan packages within the next few months.

In order to introduce a tiered dataplan, you should be able to effectively differentiate and measure time spent downloading under specific QoS conditions and permit the user to choose among different radio access available over the same place.

It actually won’t be the case, and you’ll probably be consuming your LTE fast line bucket when available.

This differentiation will introduce a complexity grade in a scenario simplified by unlimited prepaid bundles.

In Italy, in fact, operators are often offering full time/volume access providing you pay appropriate fee.

Limits are just to discourage peer to peer over mobile.

This is a way to encourage the use of mobile internet, a way to make people used to a certain use of the internet. If you know you have still 10 hours to use a day, you won’t probably complain too much if on a specific area you are forced to reconnect several times or you are downloading very slowly.

Tiered packages instead are a sort of customized service for specific high quality needs.

But will your device or application be able to discriminate among the need to use the proper tier for its connectivity.

Will you be asked to agree for wearing off you LTE plan connectivity package by your social network application, leaving the “gold reserve” to the app for managing the download of that urgent huge document?

Flat rates for heavy usage customers are not a good deal for operators. The applications business “governed” by “smart pipe” agreements did not brought the money expected. ARPU is dropping too.

The good thing for operators is that smartphones have almost fully replaced mid and high tier devices, teenagers want applications and data connectivity and managers increased the amount of data used by the mail service improving the use of attachments.

This means the market scenario has changed, but services really bandwidth demanding are not yet commonly used.

A service potentially making customers hungry of high tier connectivity packages is mobile TV.

Until mobile streaming won’t permeate people habits, I see hard to find attractive the chance to subscribe a tiered data plan.

The fact is that mobile data and tethering needs are different.

The “tiering” of bundles makes sense only for the latter. Are you, operator, still so smart to tell about the two connections?

Unlimited bundles avoid operators very expensive and difficult ways to constantly advise user she’s running out of data available on subscribed plan, with tiered data plan this mechanism should work properly and on time.

Probably the only way to make a high tier plan attractive will consist of making it  … unlimited.