How the Mobile Internet can be Realized through New Technologies & Applications

Introduction

Siavash Alamouti has stated that mobile broadband needs to evolve from its
current primitive state (a packet overlay of a cellular TDM network), to a
wireless broadband network that can accommodate much higher bandwidth per user
and overall traffic capacity.  A network optimized for mobile voice cannot
be "upgraded" to handle high numbers of high bandwidth mobile Internet users
that access rich multi-media content or are uploading/ downloading large video
and multi-media files. 

The following graphic depicts the mismatch between requirements of today’s voice
oriented mobile networks and tomorrow’s mobile broadband Internet.

To achieve acceptable service levels, more spectrum is needed along with higher
capacity backhaul and a different network architecture.  Alamouti believes
that the mobile Internet requires a technology revolution to accommodate
multi-Mbps subscriber connections from many simultaneous users.  Clearly,
the more bandwidth available per user, the more people benefit from the mobile
internet, assuming of course that the service is reasonably priced (which it’s
not on 3G networks).

To obtain a low cost per bit, a much higher level of spectrum efficiency is
needed than can be achieved by 3G or 3.5G networks.  All operators are now
in agreement that 4G networks will be characterized by OFDM, MIMO, and all IP
architecture.  Mobile WiMAX has these features now; LTE will once it’s
deployed.  Here’s an illustration of the timeline envisioned for the mobile
broadband Internet to be realized:

New 4G Applications & Services Attract More Subscribers

Here are some examples of new innovative mobile Internet applications and value
added services (mobile network providers collaborating with application service
providers):

- A smart camera uploads photos to FLICKR via a mobile broadband network just
after the photos are taken.  FLICKR then manages the on-line photo album
for designated friends or family members that were permitted access.  The
mobile network operator and FLICKR would share revenue equitably so that pricing
would be attractive to the subscriber. 

- Location Based Service (LBS) combined with GPS.  In this scenario, a
subscriber searches the Internet for a leather jacket.  The Mobile Internet
Device (MID) advertises its location to retail merchants that stock that type of
garment.  Retail merchants receiving the broadcasted message have
previously agreed to a revenue sharing arrangement with the mobile network
operator.  The merchants nearby that have the garment in stock are seen on
the MID’s display in order of closest distance. 

- A pager messenger service.  In this case, the network maintains a list of
the people permitted to send electronic messages to a mobile subscriber. 
When any of those people send a message to the subscriber, the network "wakes
up" the MID and displays the message title, providing a richer service than SMS
with more efficiency than contemporary Internet services using Yahoo Messenger
on a MID.

- Subscription based mobile video.  A user interested in a particular news
program or TV channel subscribes via the mobile operator, who guarantees QOS for
that channel or program.

These are only some obvious examples, with many other applications and value
added services that haven’t even been conceived.  According to Siavash,
once the mobile Internet is enabled, these applications will become pervasive.

Editors Note:  To read about more innovative and disruptive 4G
applications, please see

interview part II with Jose Puthenkulam of Intel.

Major Performance & Feature Upgrades Require New 4G Technologies

Siavash lists the following key performance characteristics and features for 4G
networks:

- Higher peak rates and average sector throughput
- Lower access and handover latency (<10 ms, <30 ms)
- Fair distribution of quality of service (> 4x better cell edge throughput)
- Higher mobility (up to 500 km/hr)
- Better coverage area with no signal drop- outs
- Larger VoIP capacity
- Enhanced broadcast and multicast services
- Enhanced location based services
- Deployment flexibility

Technology Enablers for the Mobile Internet include:

- Higher order single and multi user MIMO Techniques (4×4)
- Integrated Relay
- Interference management
- Standards-based techniques for multi-radio coexistence
- Multi-carrier (multi-channel) support
- Self organization and optimization, AKA as Self Organizing Networks (SON)

While most of these technology enablers are well beyond the scope and depth of
this article, we’ll briefly describe a few of them.

1.  Interference Problem

Mobile WiMAX and other mobile broadband technologies support single frequency
reuse throughout a given geographical area.  All cells/sectors operate on
one frequency channel to maximize spectrum utilization.  There is often
heavy interference in common frequency reuse, resulting in users at the cell
edge to suffer from low connection quality.  High power interferers
characterize the Downlink (DL), with a limited number of interferers, e.g. 
maximum of eight in the IEEE 802.16m draft standard.  Interference
estimation- using DL preambles – may be an effective control mechanism. 
There are more potential interferers on the Uplink (UL) – on the order of 100s. 
These are low power interferers, where interference estimation is harder and
must be dynamic. 

2.  Multi-radio co-existence

A wide variety of different radio types may exist in a given geographical area
and these must co-exist without interfering with one another.  This is
depicted in the graphic below:

3.  Multi-carrier support

Network operators have spectrum in different frequency bands with different
bandwidths.  Next generation broadband wireless systems should provide
flexibility to aggregate physically non-contiguous and non-uniform channels into
a single radio bearer channel.  This will aid in achieving efficient use of
spectrum and incremental scaling of the system bandwidth.  The resulting
system bandwidth would no longer be limited by the size of a single radio
channel.  The figure below illustrates what layers and protocols are needed
to achieve multi-carrier support:

4.  Self Organization and Optimization

In addition to the visible features and technology upgrades required for 4G,
Siavash sees several internal network control mechanisms that will be required. 
These will not likely be subject to standardization, but will have to be worked
out between the mobile operator and network infrastructure equipment vendors. 
He says 4G networks should support:

- Real plug and play installation of network nodes
- Self-configuration of the initial installation, including the update of
neighbor nodes and neighbor cells
- Fast reconfiguration and compensation in failure cases
- Support automated or autonomous optimization of network performance
- Self-optimization of service availability, QoS, network efficiency and
throughput

Is LTE the Holy Grail?

With the cellular industry now strongly backing LTE, Siavash asserts that LTE is
a positive development but the timing is too late.  He believes the
cellular industry is using LTE as a way of slowing down Mobile WiMAX and is not
committed to providing the capabilities of the technology near-term – even the
name itself is an indication.  How can we expect a technology called
"Long-Term" be available in short-term? Would you deposit your savings in a
long-term account when you know you need to access it today? Of course not! That
is why the Mobile WiMAX industry is not stopping its work and waiting for LTE. 
As a result, Siavash believes that LTE will be much "Longer Term" for deployment
than most pundits expect. 

Another important issue is that LTE is not an Evolution, but rather a "forklift
upgrade" – with a new RAN, new Base Stations, new backhaul, new packet core, new
network management equipment, and new spectrum all required for deployment. 
Mobile operators are spending billions of dollars on HSPA development today and
it is unlikely that they will start investing in new spectrum and infrastructure
to deploy LTE in the near future.  With the exception of Verizon in the US,
there are no firm dates given by mobile operators on availability of LTE
services – and Verizon’s end of 2010 date is for fixed LTE service- not for
mobile LTE deployment.

LTE modem pricing may also be an issue, if one extrapolates from already
expensive 3G modem pricing.  3G-HSPA modems are priced two to three times
higher than a Mobile WiMAX modems, but only deliver a peak rate of 14 M bit/sec
vs.  40 M bit/sec for WiMAX.  (HSPA’s average throughput per sector
was said to be 3-5 Mbit/sec by Alamouti).  With LTE offering much higher
peak rates than 3G, LTE modems will likely cost quite a bit more and therefore
will be unaffordable for many potential mobile Internet subscribers.

The cellular industry has always underestimated future data rate requirement. 
That’s why 3G only targeted 100′s of Kbits/sec data rate.  "It’s a stagnant
view of user demand for mobile broadband," says Siavash. 

Conclusions

Alamouti states, "As necessity is the mother of invention, WiMAX is the father
of LTE."  Mobile WiMAX is the first deployable wireless broadband
technology based on OFDMA and MIMO.  It also features a flat, all IP
networks that includes technology ingredients to enable a new service paradigms
and business models for truly open Internet.  He believes the cellular
industry has not delivered to the pent up consumer demand for mobile Internet. 
"We require a short-term revolution not a long-term evolution".  He
believes the battle between radio technologies need to end and the industry
needs to bring affordable Internet access to the consumer in order to stimulate
the economy globally. 

Siavash asks, "Mobile WiMAX is available today, why not use it"? He concludes,
"When LTE becomes available, Intel will definitely embrace it, but we will not
wait for that at the expense of WiMAX since this will only delay mobile Internet
further into the future."

Editors Note: Siavash Alamouti will be describing Intel Mobility Group
research project results at the IEEE ComSoc SCV meeting on October 14, 2009 in
Santa Clara, CA.  Details will be
posted next month on our web
site.  Previous Intel meeting presentations on WiMAX (Dec 05, Jan 08,
March 09) are archived there.

Alan J.  Weissberger
IEEE ComSoc SCV Program Chair and Vice Chair