Comparing WiMAX and HSPA — a guide to
the technology
N Johnston and H Aghvami
Intel has referred to WiMAX as ‘the most important thing since the Internet itself’. So is this true and if so what is so special
about WiMAX? This paper provides a basic overview of WiMAX in terms of what it is and how it works, its throughput, coverage
and security features as well as the services it could support. The paper also looks at high-speed packet access, another
broadband wireless access technology, and compares it with WiMAX.
1.
WiMAX — introduction
WiMAX stands for worldwide interoperability for microwave
access. It is a trade name for the IEEE802.16 international
standards [1]. WiMAX aims to provide a metropolitan access
network which will provide higher bandwidth and larger
coverage than are currently available with existing wireless
technologies such as Wi-Fi and 3G. Its longer reach will allow
hot zones to be developed as opposed to hot spots which are
currently available with Wi-Fi. WiMAX gets its improved
performance through its physical layer technology which is
orthogonal frequency division multiplexing (OFDM) and
through the use of smart antennas. WiMAX also provides
quality of service and improved security.
WiMAX was originally developed for fixed wireless
broadband access under the IEEE802.16a or IEEE802.162004 standard. It was believed that it could act as a possible
complement to fixed line DSL (digital subscriber line) or a
low-cost backhaul solution for local area access technologies
such as Wi-Fi. It was then developed to support mobility
under the IEEE802.16e standard which was finalised in
December 2006. This standard provided support for fast
mobility, handovers and roaming.
Over 330 companies back the standard and certification
work through the WiMAX Forum [2]. These backers include
equipment vendors such as Intel, Alcatel, Nokia, Nortel, and
Samsung, as well as network providers such as BT, Deutsche
Telekom, BellSouth, and Sprint.
1.1
Orthogonal frequency division multiplexing
WiMAX uses OFDM as the physical layer. OFDM allows large
amounts of digital data to be transmitted over a chunk of
spectrum with greater efficiency than existing wireless
technologies. OFDM works by splitting the radio signal into
multiple smaller signals that are then transmitted
simultaneously at different frequencies to the receiver. An
OFDM-based system is able to squeeze a 72 Mbit/s data rate
out of 20 MHz of channel spectrum under ideal circumstances. The key to OFDM is that the different frequencies
can be transmitted and received completely independently
of each other (this is the orthogonal property).
Typically in a wireless system the radio waves travel from
the transmitter to the receiver in a similar manner to light
rays — some rays might go straight from the transmitter to
the receiver and other will bounce off trees, buildings and
cars [3]. These various multipaths, as they are known, can
interfere constructively or destructively which causes
varying signal power at the receiver. If the data rate of the
channel is low, compared to the time difference between the
various multipath components, then a fade (a reduction in
the received signal strength) can result — deep fades of up
to 1000 times (30 dB) are possible in such systems. If,
however, a channel is transmitting high rates of data then
multipath propagation results in a frequency-selective
fading, i.e. the channel is distorted in phase and amplitude.
Complex equalisers are often employed to measure the
channel distortion and correct it in real time.
The key advance in OFDM is to divide up the spectrum
into a large number of overlapping channels that have a slow
enough data rate to not require any form of equalisation —
they are what is termed ‘flat’ fading channels (see Fig 1).
Fixed WiMAX serves different users by sharing all the
channels in time — time division multiple access (TDMA) —
whereby one gets access to the channel, then another and
another and so on. Fixed WiMAX also uses a different
frequency for uplink, and downlink transmissions, so that
they can take place independently — this is known as
frequency division duplex (FDD).
BT Technology Journal • Vol 25 No 2 • April 2007
191
Comparing WiMAX and HSPA — a guide to the technology
5
6 bit/s/Hz
64 QAM
0
response, dB
-5
4 bit/s/Hz
16 QAM
-10
2 bit/s/Hz
QPSK
-15
-20
2× throughput
-25
0 bit/s/Hz
-30
64 QAM
16 QAM
QPSK
-5
Fig 1
-3
-4
0
-1
1
frequency, MHz
-2
3
2
5
4
OFDM divides up the spectrum into channels and uses an appropriate data rate for the conditions on that channel.
1.2
Mobile WiMAX goes further and uses a technique known
as orthogonal frequency division multiple access (OFDMA)
— this allows multiplexing in both time and frequency,
whereby the channels are divided into groups and users are
allocated slots in time for one or more of these groups of
channels, as shown in Fig 2.
The other ‘big thing’ that WiMAX introduces is smart
antennas — OFDM is very well suited to smart antenna
techniques, much more so than existing 3G technology.
Smart antennas could increase the coverage, range or
throughput of a base-station. In ideal conditions
throughput can be increased by a factor of 2 or more.
In addition, mobile WiMAX uses a single piece of
spectrum for both uplink and downlink and alternates uplink
and downlink frames in a technique known as time division
duplex (TDD).
1.2.1
user B
Beam forming and null forming
Typical cellular base-stations are split into different sectors
with each sector having its own antenna and operating on a
different frequency in GSM, for example (see Fig 3).
It is expected that OFDM will become dominant in all
wireless technologies, including cellular, due to its inherent
efficiency, as well as its ability to effectively utilise smart
antennas (as described in the next section). OFDM is already
used for advanced Wi-Fi systems (including IEEE802.11a
and IEEE802.11g) and is the chosen technology for the next
generation of cellular networks — known as long-term
evolution (LTE). It was even considered as a technology for
GSM and for 3G, but the electronics complexity was such
that a low-cost solution was not possible until recently.
user A
Smart antennas
With normal antennas you can adjust the power and the
tilt of the antenna but you cannot adjust the beam itself.
Smart antennas work differently — they consist of multiple
transmit elements and the key to their operation is to adjust
the amplitude and phase of each element so as to create
constructive interference where there is a user, or group of
users, and destructive interference where there are no users,
as shown in Fig 4. This results in more of the overall transmit
power being received, reduces multipath and can avoid
user C
user D
user E
sub-channel
time
A
A
A
A
A
B
B
B
C
C
C
C
C
C
C
C
D
D
D
D
D
E
E
E
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
D
D
D
D
D
E
E
E
G
G
G
G
G
G
G
G
G
G
G
H
H
H
H
H
H
H
H
H
H
E
E
E
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
E
E
E
time
Fig 2
192
Time division multiplexing with OFDM as used in fixed WiMAX (top) and OFDMA as used in mobile WiMAX (bottom).
In OFDMA, sub-channels are allocated to different users simultaneously.
BT Technology Journal • Vol 25 No 2 • April 2007
Comparing WiMAX and HSPA — a guide to the technology
Fig 3
Fig 4
Typical GSM 3-sector base-station.
receive
chain 1
transmit
chain N
receive
chain N
Beam forming.
other sources of interference (such as neighbouring cells reusing the same frequencies).
If interference becomes a problem a null beam is sent
out to reduce the interference. The technique of beam
forming should increase the coverage of a WiMAX radio
access station.
1.2.2
transmit
chain 1
Spatial multiplexing
Multipath signals need not be bad news for radio systems —
it is actually possible to send and receive different data
streams on each path. If there are N transmitter and N
receiver elements it is possible to beam form effectively at
both the transmitter and receiver, and align one beam at
both ends to a single multipath. N elements at a receiver or
transmitter are able to form N beams. If there are also N
independent multipaths, then the capacity can, in theory, be
N times that of a system with a normal antenna. In practice,
however, some multipaths are weak and can only be used
with low data rates; multipaths also change rapidly with
terminal movement and they are only really present in urban
environments. Such systems are typically nown as MIMO
(multiple in multiple out — as shown in Fig 5).
Fig 5
1.2.3
Multiple in multiple out (MIMO).
Combining beam forming and spatial multiplexing
MIMO and beam forming are not ‘either/or’ options —
although they can be deployed separately — with WiMAX
supplier Navini believing that MIMO needs beam forming to
reach maximum range, coverage and data rates [4]. Beam
forming, it is claimed, is the key to rolling out WiMAX with an
economic number of base-stations and good indoor
coverage. Navini currently has beam forming deployed in its
pre-standards WiMAX kit deployed with Unwired of
Australia and Irish Broadband [5].
There is a trade-off with using smart antennas, with
beam forming and spatial multiplexing, between the benefit
gained and the signalling overhead. Therefore it is difficult
to say at this stage how much of the potential benefits, in
terms of coverage and throughput, will be realised. Many
companies are in the process of developing smart antennas
including Samsung and UK-based company Arraycom. It is
expected further products will be announced in 2007 as
WiMAX deployments ramp up world-wide.
BT Technology Journal • Vol 25 No 2 • April 2007
193
Comparing WiMAX and HSPA — a guide to the technology
1.3
WiMAX spectrum
There are three frequency bands defined in the standard, the
2.5 GHz and 3.5 GHz licensed spectrum, and the unlicensed
5 GHz spectrum. The mobile standard (IEEE802.16e) only
supports the licensed frequency bands. It also supports
channel sizes between 1.5 MHz and 20 MHz which is an
important feature of WiMAX — allowing it to operate in
small segments of spectrum that might become available,
although this greatly reduces the efficiency of the system.
Ideally, for a mobile WiMAX system, 10—20 MHz of
spectrum would be needed, since mobile WiMAX uses only a
single segment of spectrum for both uplinks and downlinks.
The frequency band used also affects the spacing of basestations.
As a general rule, the cell radius is proportional to the
carrier frequency used. Thus a WiMAX network at 3.5 GHz
could require roughly 60—80% more base-stations than a
2.1 GHz 3G network.
The licensed frequencies available for broadband
wireless access vary from country to country. However, it is
thought that operators are also looking especially at lower
frequency spectrum for potential WiMAX early deployment
due to the larger cells and attendant lower start-up costs.
There is very little harmonisation across the continents and
even when common spectrum exists, individual countries
often apply different regulations on how that spectrum can
be used. For example, the 3.5 GHz licences offered in the UK
in 2003 prohibited mobile services to be offered. This
therefore prevents technical features, such as handover,
which would be needed in order to support mobility. The
WiMAX Forum is appealing to regulators across the world to
release frequency allocations in the 3.5 GHz and 2.5 GHz
bands as soon as possible. This problem is demonstrated in
the map shown in Fig 6.
In Europe most interest centres on the so-called ‘UMTS
extension band’ (2500 MHz to 2690 MHz) — which has
been declared ‘technology neutral’ and will be auctioned in
the next year or so in the UK. After this, the next major
opportunity for spectrum will be when analogue TV
transmissions cease in 2010—12 in the UK. Regulations on
the use of the unlicensed 5 GHz band should be harmonised
across the EU shortly, following a decision by the European
Commission. All member states will have to comply with the
EU regulations on the licence-free use of the 5150—
5350 MHz and 5470—5725 MHz frequency bands.
1.4
WiMAX throughput and coverage
WiMAX coverage and throughput has been the subject of
considerable debate — with a throughput of 70 Mbit/s with
a coverage area of 50 km being claimed at one point. In fact,
more realistic simulations and trials, such as those run by
AT&T in the USA and trials of WiBro1 in Korea, indicate a
more prosaic performance. WiBro offers an aggregate data
throughput of 20 to 30 Mbit/s from a base-station. This
gives a sector throughput of 18 Mbit/s on the downlink and
6 Mbit/s on the uplink giving each user a downlink
throughput of 3 Mbit/s and an uplink throughput of 1 Mbit/s
[6]. The base-station coverage radius was 1—5 km and yet
the throughput deteriorated quite rapidly at the cell edge
with the rate down to around 500—300 kbit/s. The trials in
Korea were run using the 2.3 GHz frequency bandwidth. The
1 WiBro is a pre-standard version of mobile WiMAX (IEEE802.16e) used by
Korea Telecom. It will be harmonised with the standard in the near future.
Europe*
2.5 GHz†
3.5 GHz
5.8 GHz‡
Canada
2.3 GHz
2.5 GHz
3.5 GHz
5.8 GHz
US
2.3 GHz
2.5 GHz
5.8 GHz
Central and
South America*
2.5 GHz
3.5 GHz
5.8 GHz
Middle East
and Africa*
3.5 GHz
5.8 GHz
Russia
2.5 GHz†
3.5 GHz
5.8 GHz
Asia Pacific*
2.3 GHz
2.5 GHz
3.3 GHz
3.5 GHz
5.8 GHz
*The spectrum is currently very fragmented, with little to no co-ordination within any given region
†These frequencies are currently licensed for specific IMT-2000 technologies; the licence terms for IMT-2000 would need revision in order for WiMAX to use them
‡5.8 GHz is not currently available in most European countries
Fig 6
194
BT Technology Journal • Vol 25 No 2 • April 2007
Possible WiMAX spectrum. [Source: Forrester Research]
Comparing WiMAX and HSPA — a guide to the technology
coverage and throughput is likely to improve with the
introduction of smart antennas.
•
One of the main uses of WiMAX will be as a fill-in or an
alternative to cable and DSL. Fixed WiMAX offers a
good solution for broadband access in greenfield sites.
At least 15% of the US market, and huge portions of the
rest of the world, do not have a broadband
infrastructure. WiMAX is therefore a good solution as it
can be rolled out quickly with less expense than a wired
network.
Fixed WiMAX has been deployed extensively for
broadband infill as an alternative for DSL — a good example
being WiMAX Telecom [7] which has fixed WiMAX operations
in Austria, Slovakia and Croatia with over 6000 subscribers.
1.5
WiMAX quality of service
WiMAX, in contrast to Wi-Fi, has an architecture with a
base-station that is responsible for processing requests to
send or receive data from terminals, performing access
control and allocating the required radio resources to meet
the requests that are accepted. In this respect it is very
similar to 2G or 3G cellular systems. As a result it is possible
to offer guaranteed quality of service as well as best effort
and other service types, as outlined in Table 1.
1.6
WiMAX security
The issues regarding the shortcomings of Wi-Fi’s security
have been well documented. The Wi-Fi Alliance has reacted
to the security problems with wired equivalent privacy (WEP)
by developing Wi-Fi protected access (WPA). WPA is not yet
widely deployed and it requires an upgrade to the access
point as well as the network adapter. The perceived lack of
security in Wi-Fi is one of the key drivers behind using
WiMAX.
WiMAX provides industrial-strength measures for
privacy and encryption. Authentication is achieved with
X.509 certificates and extensible authentication protocol
(EAP). EAP can be used to provide standard SIM
authentication as well as support for soft SIMs. For
encryption on the link layer (from handset to base-station),
WiMAX supports the Advanced Encryption Standard (AES)
and triple DES, where DES is the Data Encryption Standard
(3DES), which is the recognised ‘strong’ encryption
standard. WiMAX also has built-in virtual private network
(VPN) support, which provides protection for data that is
being transmitted by different users on the same basestation.
1.7
WiMAX applications
There are four main fixed WiMAX applications, as described
below.
Table 1
Digital subscriber line (DSL) alternative or fill-in
•
Mobile backhaul
The cost of backhaul for celluar and Wi-Fi networks
represents a significant portion of their recurring costs.
WiMAX can provide point-to-point links of up to
30 miles and can provide data rates capable of
supporting multiple E1s (standard 2 Mbit/s links).
Operators can therefore use WiMAX equipment to
backhaul base-station traffic to their network.
•
Temporary broadband
Due to the ease of deploying WiMAX, the technology
could be used for temporary situations such as sporting
events, construction sites and trade shows. Examples of
where this has been done include the Tour de France,
the Winter Olympics in Turin, the Ironman Triathlon
World Championship and the Sundance Film Festival.
•
Public safety
Police departments are considering equipping patrol
cars with computers with wireless access. WiMAX offers
a good solution for this due to its more robust security.
WiMAX could also provide video surveillance cameras
with broadband connectivity to control centres and
support vehicles enabling officers to view situations
and make decisions in real time.
The mobile WiMAX applications can be split into two
categories (see Table 2).
•
4G network
Sprint Nextel has announced that they have chosen
WiMAX for their next generation — 4G — mobile
network build. Korean Telecom has also been
undertaking trials of WiBro as a 4G technology [8].
WiMAX quality of service classes.
Service types
Description
Unsolicited grant service (UGS)
UGS is designed to support real-time data streams, consisting of fixed-size data packets issued at periodic intervals,
such as backhaul and voice over IP.
Real-time polling service (rtPS)
rtPS is designed to support real-time data streams consisting of variable-sized data packets that are issued at
periodic intervals, such as MPEG video.
Non-real-time polling service (nrtPS) nrtPS is designed to support delay-tolerant data streams consisting of variable-sized data packets for which a
minimum data rate is required, such as FTP.
Best effort (BE)
BE service is designed to support data streams for which no minimum service level is required and which can be
handled on a space-available basis.
BT Technology Journal • Vol 25 No 2 • April 2007
195
Comparing WiMAX and HSPA — a guide to the technology
Table 2
Class description
Mobile WiMAX services.
Real time?
Application type
Interactive gaming
Yes
Interactive gaming
VoIP, video conference
Yes
VoIP
Bandwidth
40—85 kbit/s
4—64 kbit/s
Video phone
Streaming media
Yes
Information technology
No
Media content download (store and forward)
No
32—384 kbit/s
Music/speech
5—128 kbit/s
Video clips
20—384 kbit/s
Movies streaming
> 2 Mbit/s
Instant messaging
< 250 byte messages
Web browsing
> 500 kbit/s
E-mail (with attachments)
> 500 kbit/s
Bulk data, movie download
Peer-to-peer
Urban mobile broadband access
Given the high cost of a nationwide network build, it is
likely that mobile WiMAX would be deployed firstly in
urban areas as a means of offering broadband-like data
rates on the move (500 kbit/s+).
Table 2 shows the types of service that mobile WiMAX could
support, while Fig 7 shows the challenge for WiMAX mobile
services — plotting the revenue per MB versus the required
download speed of a number of current mobile services.
It is clear that, to compete with existing delivery
mechanisms, either WiMAX has to offer a lower cost per MB
than existing technologies, or new services (not possible on
current systems) must be introduced. What is needed is a
service that both requires high data rates, which only WiMAX
can provide in the medium term, and generates a high
revenue per MB to make the service profitable. Mobile TV or
mobile HDTV are potential candidates for this class of service.
2.
High speed packet access (HSPA)
Packet-based data services in current 3G systems are
technically quite limited — with long set-up times for a data
2G/2.5G
3G
4G
GSM/GPRS (0.1/MB)
text messaging
revenue per MB, US$
1000
100
10
nomadic
mobile BB
0.1
0.01
fixed BB high
10
Fig 7
196
fixed BB low
1000
10 000
100
throughput/performance requirement, kbit/s
Revenue per MB versus required speed for a range of mobile
services. [Source: Analysis]
BT Technology Journal • Vol 25 No 2 • April 2007
3G non-HSPA
300
250
200
150
100
50
0
2006
2007
2008
2009
2010
2011
2012
Fig 8
3G subscribers.
[Source: UMTS Forum Report 39]
OFDM (0.02/MB)
music
3G HSPA
350
HSDPA (0.03/MB)
video clips
fixed voice
substitution
1
400
WCDMA (0.06/MB)
downloads
MMS
mobile
voice
connection, high latency (200—300 ms), and low data rates
(64—128 kbit/s). HSPA attempts to address these issues by
enhancing 3G networks. HSPA is, in fact, a series of upgrades
to both the base-stations and the receivers and will be rolled
out to many 3G networks in the next few years with basic
HSPA already launched on a number of UK networks
including ‘3’ and Vodafone. HSPA is split into two upgrades
— high-speed downlink packet access (HSDPA) for the
download, from base-station to mobile, and high-speed
uplink packet access (HSUPA) for the upload, from the
mobile to the base-station. For brevity, this paper will mostly
describe HSDPA. HSPA terminals will eventually replace nonHSPA capable terminals with more than 50% penetration
predicted by 2020 (see Fig 8).
subscribers at year end, million
•
> 1 Mbit/s
> 500 kbit/s
2.1
HSDPA — basic principles
HSDPA introduces a number of technical innovations that
aim to increase the efficiency (i.e. to squeeze more users and
more data into a given chunk of spectrum), reduce the
latency to nearer 50 ms, and increase the maximum data
rate for users to over 2 Mbit/s [9].
The first innovation of HSDPA is to use high-speed
shared channels with a very short transmission interval. In
existing 3G, each user has a dedicated data channel and
Comparing WiMAX and HSPA — a guide to the technology
transmissions are scheduled at 20 ms intervals. This is
inefficient as users move rapidly from good to poor radio
conditions and the data rate on the dedicated channel is
then biased towards the poorest conditions to ensure a
continuous connection. In HSDPA a number of shared
channels are used and the scheduling time frame is reduced
to 2 ms. This means that users in good radio contact with the
base-station can receive at high data rates and, when
conditions are less favourable, the base-station stops
transmission and selects another user (with a higher signal
level), as shown in Fig 9. In practice, each user would be
guaranteed a minimum data rate — whatever the radio
conditions. The net result is that the average data rate for all
users is much higher and the spectrum is used more
efficiently — this is known as user diversity.
2.3
user 1
user 2
user 1 user 2 user 1 user 2
time
Fig 9
uplink
3000
2000
WCDMA
HSPA basic
HSPA enhanced
user 1
User diversity.
•
HSDPA introduces higher order modulation, i.e. more
data can be sent on a particular radio channel — this
requires a better signal-to-noise ratio and would not be
suitable for the dedicated channels of existing 3G, but,
with the fast scheduling and user diversity already
described, can be used effectively,
•
fast link adaptation is used to adjust the amount of
error coding used on the radio channel — this can take
up to 75% of the raw data rate and is adjusted in
conjunction with fast scheduling to improve efficiency,
2.2
downlink
4000
0
Three other innovations in HSDPA also contribute to
increasing the efficiency, data rate and lowering latency:
•
5000
1000
low data rate
user 2
HSPA data rates and evolution
Figure 10 shows the likely uplink and downlink capacity for
basic 3G (WCDMA), HSPA as it is currently available, and an
enhanced version which will be rolled out on many networks
within the next year or two. The enhanced version features
better receivers and interference cancellation at the basestation. Typical end-user data rates will increase from today’s
3G 128 kbit/s to 512 kbit/s and even 1 Mbit/s during this time.
kbit/s
signal quality
high data rate
of 10 MHz (one for the uplink and one for the downlink)
centred around 2 GHz in the UK. In addition mobile
operators are interested in ‘re-farming’ their GSM spectrum
for 3G use — in the UK this is either located at 1800 MHz or
900 MHz. New spectrum for 3G, including HSPA, may also
be available at 2.5 GHz in the near future as well as at lower
frequencies after the analogue TV switch-off in the UK.
fast hybrid automatic repeat request is able to combine
two frames containing errors into one, error-free,
frame, whereas in the past both frames would have
been thrown away and the frame repeated until it was
received error-free — the error detection and repeat
mechanism is also moved to the base-station, from the
remote radio controller, to reduce latency.
HSPA spectrum
HSPA and existing 3G systems can share exactly the same
spectrum — allowing the operator to choose the capacity
mix. Initially HSPA will be introduced by existing 3G operators on their current 3G spectrum — typically two chunks
Fig 10
Capacity evolution with HSPA [10].
HSPA — as currently available — offers a measured
throughput on the TCP layer (which is typically half the rate
quoted in any statistics — including Fig 10 —due to IP and
TCP overheads) of approximate 184 kbits downlink with a
(one-way) delay of 75 ms and 148 kbit/s uplink (with a
delay of 85 ms) [11]. At the same test location (UK rural) the
figures for 3G were 64 kbit/s downlink and 41 kbit/s uplink
with a delay of 200—300 ms. GPRS managed 31 kbit/s
uplink and 19 kbit/s downlink. These figures illustrate the
relative data capacity of the three technologies but all would
offer higher data rates in urban areas where the cells are
smaller and signal-to-noise ratios higher.
There are a number of further initiatives to increase the
capacity and performance of HSDPA which are being considered within standards bodies at the present time. These include higher maximum data rate terminals (up to 9.6 Mbit/s),
reduced latency (less than 50 ms), faster service set-up
(from 2.3 sec to 0.8 sec) and support for voice over IP (VoIP).
VoIP enhancements include the specific support for VoIP
data rates, seamless mobility and the ability to stay idle but
resume data transmission within 50 ms. There is also work on
introducing MIMO techniques into HSPA but early tests and
simulations show that the gains are much less than with WiMAX
(estimates range from 15—50% improvement in efficiency).
2.4
HSPA quality of service
HSPA will offer four quality of service (QoS) classes, as shown
in Table 3, ranging from a guaranteed data rate to a bestBT Technology Journal • Vol 25 No 2 • April 2007
197
Comparing WiMAX and HSPA — a guide to the technology
Table 3
Traffic class
Conversational class
conversational real time
HSPA QoS classes.
Streaming class
streaming real time
<< 1 sec
< 10 sec
approx 1 sec
> 10 sec
Example: error
tolerant
Conversational voice and
video
Streaming audio and video
Voice messaging
fax
Example: error
intolerant
Telnet, interactive games
FTP, still image, paging
eCommerce, www browsing e-mail arrival
notification
Fundamental
characteristics
Preserve time (variation)
between information
entities of the stream
Conversational pattern
(stringent and low delay)
Preserve time (variation)
between information
entities of the stream
Request response pattern
Preserve payload content
HSPA security
In terms of security, HSPA will not add anything to the
already strong security systems in use in 3G. These are very
comprehensive, based on the SIM card, offering:
•
two-way challenge for attaching terminals,
•
highly encrypted signalling traffic,
•
strong encryption on the radio link from the basestation to the terminal,
•
two-way authentication of the 3G network to the
terminal (to ensure that it cannot be a fake network),
•
support for secure applications on the SIM card — such
as automatic VPN (virtual private network) set-up,
•
support for IP-layer encryption (such as IPsec) with
header compression.
Table 4
3.
Destination is not
expecting the data
within a certain time
Preserve payload
content
Conclusions — comparing HSDPA and WiMAX
HSDPA forms the next stage in the mobile phone evolution;
it will follow up the 384 kbit/s speeds provided with UMTS
with the promise of 3 Mbit/s. From WiMAX’s point of view,
HSDPA is a form of competition (see Table 4 for a comparison of the key differences). HSDPA is not quite as fast as
WiMAX but it has better mobility. It is expected that
eventually the two technologies will become almost
identical as HSDPA gets faster and WiMAX’s mobility
improves. At this point in time both technologies are
targeting a slightly different sector. At the start, HSDPA will
be about mobility and data and voice from a cellular
platform, whereas WiMAX will be about broadband data to
the enterprise and coverage infill.
One key difference between the two technologies is the
fact that HSDPA does not require any new infrastructure as it
only requires a software upgrade. WiMAX, on the other hand
needs a whole new infrastructure to be built which will be
time consuming and very expensive. Also HSDPA already has
a head start on WiMAX. O2’s Isle of Man subsidiary Manx
Telecom launched HSDPA in November 2006. The other UK
mobile operators are also in the process of trialling HSDPA
and also expect to launch in 2007. It is unlikely that
companies will begin deploying WiMAX until around 2010.
Comparison of WiMAX and HSPA (data rates from Rysavy [12]).
Mobile WiMAX IEEE802.11e
198
Background
background best
effort
Delay
effort service. In this sense they will not differ appreciably
from those provided by WiMAX. In real, mixed standard 3G
and HSPA networks it would be expected that most constant
bit rate traffic — such as voice and video — would be carried
over the existing 3G circuit-based data channels and that
HSPA would be reserved for bursty, Internet-type traffic.
2.5
Interactive class
interactive best effort
HSPA (R5 to R7)
Date of introduction
2008—9
2006—7
Initial downlink data rate (max)
23 Mbit/s
14.4 Mbit/s
Initial uplink data rate (max)
4 Mbit/s
384 kbit/s
Evolved downlink data rate (max)
46 Mbit/s
28 Mbit/s
Evolved uplink data rate (max)
4 Mbit/s
11 Mbit/s
Latency
50 ms
100 ms
Spectrum
2.3 GHz (WiBro) new spectrum at 2.5 GHz or analogue TV Current 3G spectrum (2 GHz) new spectrum at 2.5 GHz,
reuse (112 MHz spread over 400 and 600 MHz UHF
reframed GSM (900/1800 MHz) or analogue TV
bands)
QoS
Four QoS classes; supports video and VoIP
Four QoS classes; supports video and VoIP
Mobility support
Limited in first deployment (wave 1) and seamless for
real-time services in wave 2; mobile IP used for mobility
management
Almost seamless on day 1. Totally seamless in enhanced
version. Cellular mobility protocols and soft-handover
used.
Security
Based on certificates or EAP
SIM-based security
Service set-up time
50 ms
2 sec reducing to 0.6 sec
BT Technology Journal • Vol 25 No 2 • April 2007
Comparing WiMAX and HSPA — a guide to the technology
The mobile operators appear not to be concerned by
WiMAX for these reasons. In fact Hanaro Telecom has
recently handed back its licence for WiBro (South Korea’s
version of WiMAX) to the government, partly because the
operator said it was not sure WiBro could compete against
HSDPA.
References
1 IEEE802.16 working group — http://www.IEEE802.org/16/ and
IEEE802.16e task group — http://www.IEEE802.org/16/tge/
2 WiMAX Forum — http://www.wimaxforum.org/home/
3 Hodgkinson T: ‘Wireless communications — the fundamentals’, BT
Technol J, 25, No 2, pp 11—26 (April 2007).
4 WiMAX-Vision.com: ‘Why MIMO needs beamforming’, (September
2006) — http://www.wimax-vision.com/newt/l/wimaxvision/view
article.html?artid=20017410725
5 Unwired Australia — http://www.unwired.com.au/
6 Yoon S Y: ‘Introduction to WiBro Technology’, Telecom R&D Center
Samsung Electronics Co Ltd (September 2004) — www.itu.int/ITU-D/
imt-2000/documents/Busan/Session3_Yoon.pdf
7 WiMAX Telecom — http://www.wimax-telecom.net/en/index.php
8 Wisely D: ‘Cellular mobile — the generation game’, BT Technol J, 25,
No 2, pp 27—41 (April 2007).
9 3GPP standards for HSPA — http://www.3gpp.org/ftp/Specs/archive/
10 Holma H and Toskala A (Eds): ‘HSDPA/HSUPA for UMTS: High Speed
Radio Access for Mobile Communications’, Wiley (2006).
11 Ormston C: ‘Measurements made by BT at Martlesham Suffolk’, Private
Communication (to be published May 2007).
12 Rysavy P: ‘Mobile Broadband: EDGE, HSPA and LTE’, (September 2006)
— http://www.rysavy.com/
Nicola Johnston started work at BT in 2000 as
a graduate after achieving a first class BSc in
IT. During her five years working at BT she
worked as an OSS designer and was involved
in OSS design to support BT’s 21st Century
Network. She also provided consultancy to
external telecoms companies world-wide.
She now works as a business analyst for a
major OSS provider and has worked on
projects for major telecoms companies both in
Europe and Australasia. She studied for a
masters degree in Telecomms Engineering at
the University of London and for her
dissertation she conducted an extensive study
of WiMAX. This paper covers the major findings from the study.
Hamid Aghvami joined the academic staff at
King’s College, London, in 1984. In 1989 he
was promoted to Reader and in 1993 was
promoted Professor in Telecommunications
Engineering. He is presently the Director of
the Centre for Telecommunications Research
at King’s. He carries out consulting work on
Digital Radio Communications Systems for
both British and International companies. He
has published over 400 technical papers and
given invited talks all over the world on
various aspects of Personal and Mobile Radio
Communications as well as giving courses on
the subject world- wide. He was Visiting
Professor at NTT Radio Communication Systems Laboratories in 1990 and
Senior Research Fellow at BT Laboratories in 1998-1999. He was an
Executive Advisor to Wireless Facilities Inc, USA, in 1996-2002. He is the
Managing Director of Wireless Multimedia Communications Ltd (his own
consultancy company).
He leads an active research team working on numerous mobile and personal
communications projects for third and fourth generation systems, these
projects are supported both by the government and industry. He was a
member of the Board of Governors of the IEEE Communications Society in
2001-2003. He is a distinguished lecturer of the IEEE Communications
Society, and has been member, Chairman, and Vice-Chairman of the
technical programme and organising committees of a large number of
international conferences. He is also founder of the International Conference
on Personal Indoor and Mobile Radio Communications (PIMRC). He is a
Fellow of the Royal Academy of Engineering, Fellow of the IET, and Fellow of
the IEEE.
BT Technology Journal • Vol 25 No 2 • April 2007
Unless otherwise stated, copyright of the papers appearing in the BT Technology Journal is reserved by British Telecommunications plc. The views of the contributors are not necessarily those
of the Editorial Board nor of the BT Group. Mention of products and services available from suppliers outside the BT Group does not imply an endorsement.
The papers in this Journal describe processes, products and services that may be the subject of patents or patent applications. The Journal is indexed/abstracted in ABI Inform.
199