Best Practices 3g 4g App Development PDF

© 2012 AT&T Intellectual Property Best Practices for 3G and 4G App Development
Best Practices for 3G and 4G

App Development
Whitepaper
Revision 1.0
Revision Date 7/31/2012
AT&T, the AT&T logo and all other AT&T marks contained herein are trademarks of AT&T Intellectual
Property and/or AT&T affiliated companies. All other marks, trademarks, and product names used in this
document are the property of their respective owners.
Legal Disclaimer
This document and the information contained herein (collectively, the "Information") is provided to
you (both the individual receiving this document and any legal entity on behalf of which such
individual is acting) ("You" and "Your") by AT&T, on behalf of itself and its affiliates ("AT&T") for
informational purposes only. AT&T is providing the Information to You because AT&T believes the
Information may be useful to You. The Information is provided to You solely on the basis that You
will be responsible for making Your own assessments of the Information and are advised to verify all
representations, statements and information before using or relying upon any of the Information.
Although AT&T has exercised reasonable care in providing the Information to You, AT&T does not
warrant the accuracy of the Information and is not responsible for any damages arising from Your
use of or reliance upon the Information. You further understand and agree that AT&T in no way
represents, and You in no way rely on a belief, that AT&T is providing the Information in accordance
with any standard or service (routine, customary or otherwise) related to the consulting, services,
hardware or software industries.
AT&T DOES NOT WARRANT THAT THE INFORMATION IS ERROR-FREE. AT&T IS PROVIDING THE
INFORMATION TO YOU "AS IS" AND "WITH ALL FAULTS." AT&T DOES NOT WARRANT, BY VIRTUE OF
THIS DOCUMENT, OR BY ANY COURSE OF PERFORMANCE, COURSE OF DEALING, USAGE OF TRADE
OR ANY COLLATERAL DOCUMENT HEREUNDER OR OTHERWISE, AND HEREBY EXPRESSLY DISCLAIMS,
ANY REPRESENTATION OR WARRANTY OF ANY KIND WITH RESPECT TO THE INFORMATION,
INCLUDING, WITHOUT LIMITATION, ANY REPRESENTATION OR WARRANTY OF DESIGN,
PERFORMANCE, MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE OR NON-
INFRINGEMENT, OR ANY REPRESENTATION OR WARRANTY THAT THE INFORMATION IS APPLICABLE
TO OR INTEROPERABLE WITH ANY SYSTEM, DATA, HARDWARE OR SOFTWARE OF ANY KIND. AT&T
DISCLAIMS AND IN NO EVENT SHALL BE LIABLE FOR ANY LOSSES OR DAMAGES OF ANY KIND,
WHETHER DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, PUNITIVE, SPECIAL OR EXEMPLARY,
INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS
INTERRUPTION, LOSS OF BUSINESS INFORMATION, LOSS OF GOODWILL, COVER, TORTIOUS
CONDUCT OR OTHER PECUNIARY LOSS, ARISING OUT OF OR IN ANY WAY RELATED TO THE
PROVISION, NON-PROVISION, USE OR NON-USE OF THE INFORMATION, EVEN IF AT&T HAS BEEN
ADVISED OF THE POSSIBILITY OF SUCH LOSSES OR DAMAGES.
Table of Contents
1. Overview .................................................................................................. 1
2. Scope ....................................................................................................... 2
3. Evolution of LTE ..................................................................................... 2
4. Benefits .................................................................................................... 3
a. Higher bandwidth ......................................................................... 3
b. Lower latency ............................................................................... 4
c. Improved security......................................................................... 4
d. Prioritized QoS ............................................................................. 4
5. LTE Enhanced Applications .................................................................. 5
6. Challenges ............................................................................................... 6
a. Channel quality ............................................................................ 6
b. Perceived network speed............................................................. 7
c. Limited battery life ........................................................................ 7
d. Handover...................................................................................... 8
e. Signaling traffic ............................................................................ 8
f. Third-party components ............................................................... 9
g. Device reachability ....................................................................... 9
h. CSFB (Circuit Switched Fallback) speed impact ......................... 9
i. IPv6 support ............................................................................... 10
7. Best practices ....................................................................................... 10
a) Data cap.............................................................................................. 10
i) Avoid large thumbnails............................................................ 11
ii) Provide configurable data transfer settings............................ 11
iii) Use Wi-Fi as much as possible ............................................. 12
b) Battery Life .......................................................................................... 12
i) Avoid constant polling ............................................................. 13
ii) Download in bursts rather than continuously streaming ........ 13
iii) Minimize aggressive behavior ............................................... 13
c) Speed .................................................................................................. 14
i) Decouple user transitions from data interactions .................. 14
ii) Cache more often .................................................................. 14

iii) Flush the cache regularly .......................................................15
iv) Create one session with multiple gets ...................................15
v) Pre-fetch ................................................................................. 15
d) Robustness ......................................................................................... 15
i) Plan for a mix of networks .......................................................15
ii) Use bigger buffers .................................................................. 16
iii) Support IPv6 .......................................................................... 16
e) Follow the latest mobile developments ...............................................16
f) Leverage existing tools and code samples ..........................................16
8. Topics to explore ................................................................................... 17
9. Conclusion ............................................................................................. 18
10. FAQ (Frequently Asked Questions) about AT&T 4G App
Development .......................................................................................... 19
1. Are all 4G flavors identical? ........................................................19
2. How do I test the speed of my handset’s 4G network? ..............19
3. How do I tell which cellular network (2G, 3G, 4G HSPA+ or 4G
LTE) my handset is connected to? .............................................19
11. Terms and Acronyms ............................................................................ 20
References .................................................................................................... 21
1. Overview
4
G LTE (Long Term Evolution) is the latest, fastest and most secure version of four
generations of mobile technologies: 1G, 2G, 3G and 4G. Prior to LTE, voice and
data were carried over two channels: A circuit switched channel for voice and an
IP-based packet core network for data. LTE is designed to provide a single,
simplified, all-IP channel to carry both voice and data. LTE is being deployed worldwide
to meet demands from the exponential growth of mobile data traffic and is one of the
three largest drivers of mobile data consumption 1:
• Faster networks: LTE, with higher speeds and significantly lower latencies,
enables the innovation of mobile apps with richer services, for example, video
sharing in high-definition on the uplink for video conferencing or playing latency-
sensitive games.
• Powerful devices: Computing power is bringing desktop quality computing to
mobile devices (smartphones and tablets). Declining price-points for increasingly
more powerful devices are resulting in the increased usage of Internet enabled
mobile devices. For example, over 56% of AT&T’s customers are already using
smartphones 2.
• Mobile apps: Apps are considered the major driver of mobile data consumption.
Video apps alone (e.g., Hulu, Netflix) consume 66% of all mobile data traffic 3.
Apps on powerful, faster mobile networks and mobile devices bring
opportunities to developers that understand both the device platforms and the
mobile networks (2G, 3G and 4G) in which they operate. Developers will be able
to build more efficient and powerful mobile apps when they understand how
networks function. This efficiency in turn makes networks more valuable so
customer adoption of apps accelerates.
This white paper emphasizes LTE technical considerations rather than legal or business
development issues. It covers the following main areas to help developers build LTE
efficient apps:
1. Evolution: A history of the evolution of mobile networks from 1G to 4G.
2. Benefits: The benefits of developing apps for LTE.
3. Apps: Examples of LTE specific apps.
4. Challenges: The LTE rollout is a gradual deployment to U.S. cities which
introduces issues that developers must consider while the LTE deployment is
ongoing.
5. Best practices: Core principles for developers that directly impact their users.
2. Scope
The focus of this whitepaper is to address the best practices for the four key areas of
application performance that customers perceive the most:
 Data cap: A significantly faster network hasn’t translated (yet) into larger
monthly data caps. Users are still dealing with 3 – 5 GB data caps.
 Battery-life: LTE handsets result in a heavier drain on battery life than earlier
generations. This heavier drain represents the biggest area of concern for
customers.
 Speed: Bandwidth and latency are the two metrics that users mean when they
say an app is “fast” or “slow.”
 Robustness: Apps must be robust in handling handover to slower 3G networks.
Unlike other carriers, AT&T’s LTE/HSPA+ combination will have a less jarring
speed drop than other carriers that downshift between an LTE/3G combination.
3. Evolution of LTE
The following chart shows the evolution of the capabilities of mobile networks from 1G
to 4G LTE.
1G 2G 2.5G 3G 4G HSPA+ 4G LTE
Year(s) 1980s 1990s 2001 2008 2010 2011
Voice calls      
Digital     
SMS     
Internet access     
Circuit Switched Data     
IP-Based Packet Core    
Industry Max Speed (Kbps) - 9.6 40 – 50 3,800 6,000 61,000
Icon displayed None None G or E 3G 4G or H+ 4G LTE
1G: In the 1980s, a 1G cellular network handled only voice calls (no data) in analog
waves at 9.6 Kbps. It had several limitations such as poor voice quality and a lack of
privacy (third parties could eavesdrop on calls on analog networks) due to the absence
of encryption.
2G: In the 1990s, second generation (2G) mobile phones and networks were introduced.
Due to modulation (voice converted to digital signals in the phone and then analog for
transmission), handsets overcame some 1G limitations such as a lack of privacy by
adding encryption. Unlike 1G phones which only supported voice calls, 2G also
supported data services such as SMS and WAP (Wireless Access Protocol – a lightweight
protocol for transferring data on handsets). Both voice & data were transmitted using a
single channel, CSD (Circuit Switched Data) technology.
2.5G: This was the interim technology to 3G. 2.5G was launched in two major flavors:
GPRS and EDGE. The letter “G” would be displayed on a handset’s screen next to the
bars if it was on GPRS (General Packet Radio Service) and “E” if it was on the faster EDGE
(Enhanced Data for GSM Evolution). EDGE is approximately 3x faster than GPRS and was
also known as 2.75G. GPRS had download speeds of 30-40 Kbps.
3G: By early 2008, the third generation, 3G, was launched with GSM UMTS (WCDMA,
HSPA, and HSDPA). Phones display “3G” on the screen when using a 3G network. With
3G, voice and data travel on two separate data paths: CSD for voice and a packet core
network for data. As a result, AT&T customers using 3G became the first users
nationwide, to be able to use voice and data simultaneously.
4G: The ITU (International Telecommunications Union) defines 4G as the ability to
achieve a theoretical peak download of 100 Mbps while mobile and 1 Gbps while
stationary. This UN entity also extended the classification to include slower 4G
forerunners which offer significant improvements over 3G such as HSPA+ (High Speed
Packet Access) and its faster successor LTE (Long Term Evolution). 4G LTE was rolled out
by AT&T in September, 2011. Phones typically display “4G” or H+ when on HSPA+ and
“4G LTE” when on LTE.
The next evolution after LTE is LTE Advanced (LTE-A) with theoretical download speeds
of 1 Gbps and 200 Mbps upload speed. However, deployment dates have not been
announced.
4. Benefits
4G LTE brings four key benefits to mobile devices:
a . H ig h e r ba n dw idth
Bandwidth is the rate at which data can be transferred via physical media such
as fiber, copper or in the case of smartphones, via radio waves. Depending on
device and network conditions, LTE is actually 7 to 15x faster 4 than the earlier
generation i.e., 4G HSPA+.
Note that LTE speeds shown in the following chart represent peak data rates
under ideal, lightly loaded network conditions. Actual speeds under real-world
conditions will be less than those at peak rates.
4G Type Definition Download (Mbps) Upload (Mbps)

HSPA+ High Speed Packet Access 6 1.2
LTE Long Term Evolution 5 100 50
b. L o w e r la te n c y
In LTE, the latency or lapsed time between a request and its response has been
reduced to Wi-Fi levels of less than 50 milliseconds (ms). The combined effect of
higher bandwidth and lower latency enables mobile developers to deliver better
experiences. That includes the use of data intensive technologies such as HD
video streaming, cloud-based speech recognition, augmented reality, and
interactive gaming.
Round-trip Latencies for cellular and wired environments
Environment Type Latency (ms)
Cellular (AT&T) 4G HSPA+ 100 - 200
Wi-Fi (phone) 802.11n 50
Cellular (AT&T) 4G LTE 40 - 49
Wired Ethernet 25
Note: Latencies in this table are for round-trip requests and

responses that occur entirely in the continental U.S.
c . I m pr o v e d s e cu r ity
Security improvements in LTE include the use of larger 128 bit network
authentication keys when data is sent on a secure channel. This allows the RAN
(Radio Access Network) to provide more secure communications between
network elements and terminal devices. IPSec is also provided for secure
tunneled IP communication if it is supported by the device.
d. P r io r itiz e d Q o S
LTE Quality of Service (QoS) introduces nine types of traffic with individual
processing priority. This means that latency-sensitive traffic can be prioritized
over latency tolerant traffic. For example, conversational voice and (real-time)
video with a higher QoS will not be delayed by lower QoS traffic such as email
and file-transfers. QoS prioritization also means that voice data can be
prevented from over-consuming bandwidth reserved for video broadcasts.
Quality of Service is defined by the carriers and can be customized at the
subscriber level. Some of the nine types of LTE traffic are conversational voice,
conversational video, streaming video, and real-time games6.
The policy and charging rules function (PCRF) is one of two dedicated network
elements that define, control, and enforce policies in addition to managing QoS
resources7. Currently, PCRF is controlled by carriers. Some examples of how

PCRF can be used by carriers are: Providing a data cap increase for certain
traffic types if bills are paid on time, and providing unlimited data usage after a
certain time, similar to how cell phone users have unlimited talk time on nights
and weekends.
Currently, there are no APIs provided to access PCRF. In the future, if PCRF APIs
are provided, developers will be able to build that functionality into their
services. For example, such a service could notify the carrier with an instruction
to zero-rate video traffic temporarily when a paid subscriber of Hulu Plus
watches HD video, while drawing down the data cap at the regular rate for an
unpaid user of Hulu Plus watching the same HD content.
5. LTE Enhanced Applications

The promise of an LTE network, with broadband type speed and latency, opens a world
of opportunity for the types of applications that developers can build. The following are
some examples of the kinds of applications that could be enhanced by LTE.
Video apps: High-definition (HD) extensions of existing 3G and 4G HSPA+ video and
video centric apps are obvious candidates for LTE enhancement. For example, a user
may want to do a live HD broadcast of an event, via a server, to the smart TVs and
phones of their friends, or conduct a private one-to-one broadcast to their TV. They may
even want to conduct live, one-to-many, HD broadcasts, via server, to broadcast to the
public.
Augmented reality: A new trend in mobile apps is to overlay information on a live view
of a physical real-world environment. An example of this would be a user trying to
identify if there’s a sale in the immediate vicinity of their location. The user would pan
the camera of a smartphone or tablet across a scene and tags would appear on the live
video view. The user would then be able to click on those tags and get rich data (audio,
video and text) specific to that store, such as coupons, menus, reviews, historical
factoids, or user generated content. Augmented reality works today on 3G. However,
where 4G LTE becomes vital is the spectral efficiency – fatter pipe – which allows larger
numbers of people to view bandwidth intensive video linked to tags.
Interactive games: There are over 135 million gamers in the U.S. 8, and with the rise of
MMOs (Massively Multiplayer Online games), gamers can now play latency sensitive and
bandwidth hungry games such as World of Warcraft that are streamed directly to their
LTE devices. In fact, in 2012, game streamers such as OnLive and Gaikai already stream
hardcore game titles such as Need for Speed directly to broadband devices such as
laptops and smart TVs. This is being extended to LTE devices as well which, because of
their reduced latency, provide a better gaming experience than prior generation mobile
devices. Note that streaming apps, as explained in the challenges section, are battery
and data hungry and should be tested and optimized with tools such as AT&T’s
Application Resource Optimizer 9 (ARO) to minimize power consumption.
M2M: Mobile networks used to be only for voice communication. But now, the trend is
to the Internet of Things or machine to machine (M2M) communication. Smart meters,
traffic lights, cars, household appliances, and other devices are able to communicate
with each other.
M2M is used in 2G and 3G 10. The benefit that LTE brings to M2M is increased spectral
efficiency. Spectral efficiency in 3G UMTS for example, is 3 bits per second (bps)/Hz. LTE,
however, increases that five-fold to 15 bps/Hz. This increased efficiency becomes critical
when forecasts show that the number of connected devices is forecasted to double to
12 billion by 2020 11. Thus, carriers are able to offload M2M data traffic from congested
3G to LTE.
One example of M2M is AT&T’s Digital Life (digitallifeservices.att.com) that provides a
remote monitoring and automation platform for connecting a user’s device to devices in
their home. This platform makes it possible to create apps for web-based home
automation, security, energy management, and healthcare services.
6. Challenges
Network changes don't happen overnight. It has taken almost ten years for each
wireless network generation to evolve and be deployed. The deployment of LTE
networks is well under way, but it is expected to coexist with 3G and even 2G networks
for years to come.
With all the claims around LTE and its blazing speed, it is important to understand that
there are many factors that may affect the perceived speed. Such factors include tower
proximity, packet fragmentation, concurrency, and signal blocking. The best practices
section in this white paper addresses how developers can leverage LTE best practices to
manage these challenges, and it introduces a software tool to help developers visualize
issues and evaluate how compliant their app is with these best practices.
a . C h a n n e l q u a lity
In wireless networks channel quality is measured by the amount of

successful packets delivered. Channel quality is generally affected by
the amount of concurrent subscribers on the same RF channel. LTE,
being a wireless packet network, confronts challenges similar to that
of a wired network where packet collisions affect performance.
It is foreseeable that densely populated areas may experience lower quality at peak
usage times with a noticeable impact on performance. Channel quality is also
impacted by unnecessary pings and keep-alive signals.
b. P e r c e iv e d n e tw o r k s pe e d
On occasion, LTE effective data rates may fall below users’

expectation. By adhering to mobile development best practices,
application developers can help to minimize these performance
problems.
The following factors will impact the actual speeds an end user will experience:
• Distance from cell towers.
• Network congestion. As a shared medium, cellular networks typically slow
down as they are loaded with users connected at the same time.
• RF interference caused by weather, terrain, buildings, and walls.
• Device specs such as OS type and processors.
• Application design.
c . L im ite d ba tte r y lif e
Batteries may have improved in capacity, but have not kept pace
with power demands. LTE phones consume 5% - 20% more than
prior generation phones, depending on the apps used 12. Handset
radios are the single biggest source of power drain in any device
apart from the touch screen. Unlike the display, the radio is always
on and is draining the battery. Apps also drain devices by not managing network
connections. And with LTE devices, the radio is doing a lot more than in previous
generations of devices. For example:
• Multiple rabbit ears: All LTE devices use MIMO (Multiple Inputs, Multiple
Outputs), which essentially means a handset no longer uses a single antenna
to send or receive signals. Rather, it uses two parallel transmissions. Each
antenna requires its own power amplifier for transmission. So, at a basic
level, two parallel transmissions is (almost) analogous to running two phones
off a single battery. To compound the challenge, future devices will use even
more than two antennas.
• Scanning further: LTE deployment is still underway, and even within an LTE
market, cell towers may not be densely located. With cells spaced much
further apart, devices have to boost transmission power to reach further.
Since there will also be coverage gaps, smartphones will be dropping in and
out of LTE coverage taxing an already over-taxed battery. So apps, especially
streaming apps, need to be robust enough to run on 3G and 4G HSPA+
networks and conserve battery.
• Packed airwaves: Faster networks essentially mean that more bits are
packed into a radio wave. But as waveform complexity increases, so too does
the power needed to modulate and demodulate the radio wave.
d. H a ndover
As new technologies are introduced into a carrier’s network, developers need to

build apps that are robust enough to manage the handover delays between a mix of
2G, 3G, 4G HSPA+, and 4G LTE environments. For instance, the handover time in a
switch from an LTE environment to an HSPA+ environment is 2.4 seconds 13.
e . S ig n a lin g tr a f f ic
Signaling storms caused by applications have brought down LTE networks around
the world. Signaling is the communication overhead that enables user
authentication, authorization, charging, and billing 14. Signaling costs occur when
phones change their state, send data, and move from one area of coverage to
another.
The addition of LTE networks increases the amount of signaling traffic. Unlike 3G’s
three state energy model, LTE has two states - RRC_CONNECTED and RRC_IDLE - as
shown in the following diagram. Details of these LTE states are provided in the
reference section of this white paper 15.
But at a high level, understanding the RRC (Radio Resource Controller) state machine
for LTE will help developers better understand how to maximize the battery life.
Assume for example, an LTE phone has a battery capacity of 5 Wh. An app that is in
the RRC_CONNECTED state consumes 1,000 – 3,500 mW but when it is in the
RRC_IDLE state it consumes less than 15 mW.
Continuous Timer expiration

Reception
DRX
Data transfer
T1
Short T3
Long
DRX DRX
T2
RRC_CONNECTED RRC_IDLE
(1,000 – 3,500 mW) (< 15 mW)
Even before LTE was added to the environment of connected devices and 3G
networks, mobile operators faced significant costs in handling signaling, as opposed
to the costs of handling data traffic alone 16.
Applications that disrupt a phone’s radio-state management through poor
networking optimization greatly increase the amount of signaling, irrespective of
data size. In fact, by ignoring the implications of polling and chattiness, apps have
been known to cause signaling storms and even localized network outages.
f. T h ir d- pa r ty c o m po n e n ts
A common aspect of modern mobile app development is the use of drop-in

components like advertising SDKs, analytics SDKS, and map controls. Banner ad
controls have been observed to have a significant effect on uncontrolled network
traffic; analytics engines and maps may be communicating with their respective
back-end services outside of the direct control of the developer. Therefore, it can be
useful to analyze and understand the types of traffic and the patterns of data use
that these controls introduce into your app. A tool like the ARO 17 can help assess the
impact of these and controls.
g . D e v ic e r e a c h a bility
Users and developers both expect that devices are, in essence, permanently
connected to the internet, regardless of the physical mechanisms by which this
connection occurs. Platform vendors are addressing the problem of device
reachability through the provision of push notifications: Apple, Google, Microsoft
and RIM all provide some variation of this free service. However, push notifications
have server-side limitations for developers wanting to ensure their apps can be kept
synchronized and always reachable. One example of this limitation is that the push
notifications are strictly one-way: there is no return path for the phone to
acknowledge receipt of the message.
h . C S F B ( C ir cu it S w itc h e d F a llba c k ) s pe e d im pa c t
LTE, being an all-IP network, also has the ability to transmit voice over LTE (VoLTE)
rather than using the 3G circuit switched network. AT&T has not yet deployed
VoLTE, so the default method for voice is currently CSFB. However, VoLTE is the
long-term goal for delivering voice, and it brings some challenges like the following:
 Enhanced 911 (E911) U.S. regulations mandate that the physical location for
users dialing 911 be known to the 911 service. While that’s possible for
circuit-switched networks, a physical location can’t be identified in an all IP
network.
 LTE coverage is not available in all markets and IMS (IP Multimedia Services)
network elements, a prerequisite for VoLTE, are not fully deployed.
 There are longer handover times with VoLTE. The legacy CSD has handover
times of a few milliseconds, while VoLTE takes 2-3 seconds.
For these reasons and others, an LTE handset falls back to 3G circuit-switched voice
and SMS services in two situations: When it is out of LTE coverage, and when a non-
VoLTE voice call is made or received while in LTE coverage. This means that users will
see a drop in data speeds from LTE to HSPA+ during these two scenarios. The same
situation occurs when users are playing a game on LTE and a call is received.
i. I P v 6 s u ppor t
The capacity of IPv4, the Internet’s main addressing protocol with over four billion
addresses, was exhausted in most parts of the world by 2011. Because of this, IPv6,
with 3.4 x 1038 addresses, replaced IPv4. IPv6 is available on non-LTE devices, but it
is worth noting that while the goal is for LTE devices to support IPv6, not all handset
manufacturers support NAT64 (Network Address Translation from IPv6 to IPv4). This
means that developers should continue to ensure their apps are IPv6 and IPv4
compliant.
7. Best practices
This section discusses the best practices for addressing the top issues 18 concerning radio
resources.
Data cap Battery life Speed Robustness
a ) Da ta ca p
Due to the limited spectrum, data caps are still present in LTE
data plans and users are charged for overages. Note that data
caps count data uploads AND downloads. However, using tools
such as AT&T ARO, developers can identify and work to
eliminate duplicate content that can comprise as much as 20%
of HTTP traffic.
As a comparison, a YouTube SD (standard definition) video on a
tablet size screen (10 in) will consume 60 MB per hour while HD (high definition)
video like that on YouTube HD or Netflix will consume 150 MB per hour on the same
screen. So, it’s clear that users can quickly exceed the cap of the most popular data
plans (which are typically a few GB per month) with 50 hours of SD video or 20 hours
of HD video, and thus incur overage charges.
Data consumption will obviously vary by screen size and by the types of content that
users consume. To estimate daily and monthly usage for an end user, developers can
use AT&T’s data calculator 19.
i) Avoid large thumbnails

A 3G best practice that still applies with 4G is minimizing thumbnail sizes. A
thumbnail is a miniature version of an image or a static image of a video. And the
thumbnail file size is a function of four attributes:
Image file size = f (width, height, bit depth, compression)
Apps sometimes download high-resolution images for thumbnails, unaware that
many devices can’t display these images at that quality. This causes users to
unnecessarily drain their data buckets and lose time waiting for a high-resolution
thumbnail image to download, while a lower-resolution image would have saved
both time and impact to their data cap.
For that reason, developers should programmatically check the device resolution
and download the appropriate thumbnail for that resolution.
For example, assume a user has a download speed of 10 Mbps and downloads
two, uncompressed images. Assume also, the bit depth for both images is 24
bits/pixel which is common for JPG photos.
• Image 1
o Size = 3,456 pixels x 4,608 pixels x 24 bits/pixel = 382 Mb (45.6 MB)
o Download time = 382 Mb / 10 Mbps = 36 seconds
• Image 2
o Size = 269 pixels x 202 pixels x 24 bits/pixel = 1.3 Mb (0.16 MB)
o Download time = 1.3 Mb / 10 Mbps = 0.1 second
Key lesson: Send high-resolution thumbnails to high resolution devices and low-
resolution thumbnails to low resolution devices.
ii) Provide configurable data transfer settings

Always provide configuration settings for high volume data transfer scenarios such
as defaulting to downloading or uploading large amounts of data over Wi-Fi. A
partial list of configurable options is:
 Force Wi-Fi for specific apps or features.
 Switch to Wi-Fi rather than 3G to conserve battery.

 Alert on long downloads: It’s critical to allow users the option to either
configure data transfer options (Wi-Fi vs. LTE) or to be prompted before a
large amount of data is transferred in either direction.
 Alert when near data limits.
 Limit or cap data quality.
 Limit or cap audio quality.
 Limit or cap overall upload and download rates.
 Limit data cache size.
 Set limits on audio or video streaming time. For example, a sleep timer for
audio streaming
iii) Use Wi-Fi as much as possible 20

Real-world LTE latencies are significantly lower than HSPA+ and in fact rival Wi-Fi
speeds. From a usability perspective, users tend to feel as though a web page or
video loads instantaneously when the latency is 50 ms or less.
Round-trip latency is the time spent by a user’s device waiting to get a response
after sending a request. So when latency matters, such as when playing lag-
sensitive online games, the app can stay on LTE rather than switching to Wi-Fi,
since LTE latency is around 1 - 10 ms faster than even 802.11n Wi-Fi.
However, when this slight increase in latency will not impact user experience, the
app should switch to Wi-Fi (if available) since Wi-Fi avoids reducing a user’s
already limited data bucket. As always, tools such as ARO will help the user analyze
whether their app is meeting their latency requirements.
b) B a tte r y L ife
Dual and quad-core devices with multiple antennas (MIMO) on

LTE will consume power faster than legacy bands such as HSPA+
and HSPA. In fact, the first releases of LTE-ready mobile devices
in 2012 and 2013 will consume even more battery power for
reasons such as limited LTE coverage while LTE networks are still
being rolled out.
Due to the higher amount of power used in the
RRC_CONNECTED state to achieve the faster transfer speeds in LTE, battery life
should represent the biggest concern for developers, and should lead them to adopt
a policy of using the minimum amount of battery power needed 21.
i) Avoid constant polling 22

Reducing polling conserves battery power. Note that frequently moving between
LTE, HSPA+ and 3G networks could cause apps relying on keepalives or frequent
updates to timeout during network handovers. Key lesson: Poll as infrequently as
possible rather than leaving a channel open all the time.
Use the push notification mechanisms provided by Apple (APNs), Google (C2DM),
Blackberry (Push Service) and Microsoft (MPNS) to ensure your app can refresh
when necessary without relying on polling, but use push notifications with
caution since depending on your app, push can actually use more battery power
than polling.
For example, imagine that your email server pushes 25 emails to you in an hour.
That translates into 25 push connections. However, polling every ten minutes
translates into only six connections an hour. So in this scenario, polling uses less
battery power than push.
Translating this into battery savings, assume that an AT&T ARO analysis shows
that your app consumes 10-13 Joules (J) for each poll or email pushed to the
phone.
• Poll: 13 J x 6 connections/hr = 78 J/h
• Push: 13 J x 25 connections/hr = 325 J/h
• A Samsung Captivate has a 5 Wh battery capacity
• 3,600 J/Wh x 5 Wh = 19,800 J in the battery
• Polling usage = 78 J per hr / 19,800 J x 100 = 0.4% of the battery/hr
• Push usage = 325 J per hr / 19,800 J x 100 = 1.6% of the battery/hr
ii) Download in bursts rather than continuously streaming 23

Continuous streaming drains batteries faster than downloads do. In areas with
slower networks, batteries are depleted faster because the device boosts its
network reception in order to overcome noise.
When an app is providing a streaming service, it should warn users of the power
and data costs. Key lesson: Avoid battery drain by streaming or downloading in
short bursts. Using the ARO tool will graphically show how data downloads in
bursts save energy.
iii) Minimize aggressive behavior

If packet loss occurs on a crowded wireless network, a device needs to back off or
wait for a response. If your app is continuously attempting to regain lost packets,
it can exacerbate the original failure conditions, degrading or denying service to
other devices and congesting traffic. This overly aggressive behavior also drains
device batteries faster. Key Lesson: An app that keeps trying to pull down signals
from a congested network rapidly sucks the life from batteries.
c) S peed
There are three metrics that quantify what users mean when
they say an app “feels” fast: bandwidth, latency and app
responsiveness. This section focuses on principles that
developers can follow to further enhance the user’s feeling of
using a fast application.
i) Decouple user transitions from data interactions

A big component of "feeling fast" is also communication to the user on actions: an
app can feel sluggish if a tap or swipe by the user requires a data connection and
response before the next UX (user experience) display. This is because users
waited for a data response, even a fast data response, before the UX responded.
Decoupling, whenever possible, the UX transitions from the data interactions, will
allow for a smoother UX flow and "faster" app perception to the user.
ii) Cache more often 24

Mobile applications that allow their main operations to work even when no
network is available are the most valuable. As in any distributed environment,
mobile applications can benefit in various ways by maintaining a local copy of
data that is unlikely to change in the short-term. By doing this, mobile
applications achieve several important objectives:
 They significantly reduce repeated retrieval of duplicate data.
 They improve the app’s responsiveness by accessing data locally.
 They enable limited functionality even when the app is off-line.
 They reduce battery drain by reducing overall network activity.
Reference data is a prime candidate for caching, but any kind of data that occurs
within a regular time frame is also a good bet (e.g., news articles).
Caching extensively also helps eliminate spikes in network requests that can
potentially cause user perceivable performance problems. Spikes in network
activity typically occur at predictable times. If you understand how and when your
app needs data, you’ll be able to cache in advance of periods of high network
activity.
iii) Flush the cache regularly

When implementing caching, also implement a sensible flushing schedule to
prevent rarely-used files from consuming too much storage space.
iv) Create one session with multiple gets

Previous network generations such as 3G, due to network restrictions, were
designed to handle small requests frequently. With LTE’s higher bandwidth and
the capability to handle longer sessions, the preference is to group requests into
one or as few sessions as possible. Key lesson: Maximize bandwidth in fewer calls
and group multiple requests concurrently.
v) Pre-fetch 25
Leverage LTE’s larger bandwidth and lower latency by anticipating and pre-
loading data that the user wants. This requires experimenting with algorithms
that work best for your app. Data retrieval during hours of expected high traffic
can put your user and the network under unnecessary pressure. Consider when
your application or its user may require some particular data and schedule its
retrieval ahead of time.
This technique, although imperfect, provides for a better user experience. An
example of where pre-fetching becomes a problem is when the app pre-fetches
so much data that it is never used. This overuse drains customer’s data caps
rapidly in addition to using excessive memory.
d) R obu s tnes s
Apps will have to deal with multiple real-world scenarios such as

users crossing dead zones, handovers between networks, and
varying network speeds. Some techniques for developers to
manage these scenarios is to periodically check for network type
and to use bigger buffers.
i) Plan for a mix of networks
2.5G EDGE 4G HSPA+
3G HSPA 4G LTE
Networks differ in bandwidth, latency, throughput, reliability, and security. Even

on LTE, data transfer rates will not remain constant since many factors such as
network load, distance from cell towers, and terrain, impact transfer rates. This
variability creates a challenge: Is the user on a fast 4G HSPA+ network or a slow
LTE network? Apps will still have to make intelligent choices based on network
availability (2G, 3G, 4G HSPA+, 4G LTE).
ii) Use bigger buffers

LTE is significantly faster than 4G HSPA+ and 3G, so it makes sense to use bigger
buffers. Large buffers work better even when switching between networks.
Buffering, unlike caching, refers to pre-loading data into a reserved area of
memory. In streaming audio or video, buffering refers to downloading or
streaming a certain amount of data before the user has consumed the data e.g.,
downloading some video before playing a video clip. Caching however, refers to
maintaining the data, audio clip, emails, or movie, after the user has consumed
the data.
iii) Support IPv6

As explained in the challenges section not all handset manufacturers support
NAT64. So developers should continue to ensure their apps are IPv6 and IPv4
compliant. One way to do this is to ensure that apps do not use a fixed string
length for IP addresses26.
e ) F o llo w th e la te s t m o bile d e v e lo pm e n ts
The mobile development community has a number of useful resources online
where questions and intelligence are pooled to derive best practices. Although
the generic principles provided in this white paper will mostly apply to any mobile
platform, in most cases you can find specific solutions related to individual OS’s
and LTE standards on a number of sites like the following.
• 3rd Generation Partnership Project (3GPP) which has become the focal point
for mobile systems beyond 3G (www.3gpp.org).
• AT&T Labs research (www.research.att.com).
• AT&T’s Developer Program (http://developer.att.com).

f ) L e v e r a g e e x is tin g to o ls a n d c o d e s a m ple s
A number of recent initiatives have been directed at helping developers write

more network-friendly apps.
• The Application Resource Optimizer (ARO) from AT&T is a diagnostic tool for
analyzing mobile web application performance. AT&T ARO automatically
profiles your application based on suggested best practices, and provides
recommendations that allow you to optimize performance, make battery

usage more efficient, and reduce network impact. Download it here.
o http://github.com/attdevsupport/ARO
o The github.com site contains open source repositories with multiple
projects related to mobile development, including ready-to-use
libraries for common programming tasks.
o This particular branch contains the repository for Open Source ARO
which lets you integrate the ARO tool and its best practice concepts
into your testing and development processes.
• The Smarter Applications initiative, from the GSMA. As well as offering a
prize for smart apps, these guidelines contain specific code samples based
around best practices.
8. Topics to explore
There are several more best practices to explore in addition to the ones described in this
white paper:
 Legal: The legal and ethical issues around collecting, managing and transferring
end users’ personal information has not changed with LTE. Developers are
expected to have experts in privacy law and policy to guide them in the
management of end users’ information. In the U.S., the FTC (Federal Trade
Commission) has brought several enforcement actions against app developers
accused of misusing user data.
 Biz Dev: This white paper also does not cover business models like in-app
advertising, in-app purchases, or freemium models. Nor does it cover which
category of apps – games, utilities, social, news, etc. – are the most popular to
develop. If users implement advertising in their apps, however, this paper
provides the background and tools needed to optimize apps since much of the
signaling in apps comes from advertising.
 Hardware platform: Mobile devices now not only include laptops and
smartphones but tablets, e-readers, and potentially other consumer devices.
This white paper is device agnostic. While this paper references primarily LTE
smartphones and to a lesser extent tablets, readers should assume that
statements about these devices could be extended to other consumer devices.
 Software platform: Developers are also assumed to be familiar with the best
practices for building apps for their mobile OS such as iOS, Android, Windows
Phone, and BlackBerry OS.
 Non AT&T carriers: This paper takes a US-centric, AT&T LTE approach. However,
best practices should apply equally well to LTE or WiMax app development for
other carriers like Verizon, T-Mobile, Sprint, and Clearwire.
9. Conclusion
LTE, the fourth generation mobile network, is currently being deployed nationwide. This
4G network has a significantly higher bandwidth, lower latency, better security, and
more types of QoS than earlier generations. All these benefits offer the opportunity for
developers to build enhanced applications not possible with 3G. Some possible
applications are live, HD broadcasts from smartphones or the ability to play latency-
sensitive, interactive games.
However, this network is only one of three elements in an ecosystem that has the
potential for new revenue opportunities for carriers and developers. Networks, devices,
and apps have to work seamlessly together to drive consumer adoption and thus
revenue.
That’s why this white paper focuses on technical best practices for developers to follow
when building LTE optimized apps. For the same reason, non-technical topics such as
business development or licensing are not addressed. These technical best practices are
identical to 3G best practices for areas that customers can perceive: data caps, battery
life, speed, and robustness. Developers only need to learn a few, slight modifications
around LTE buffering and radio states that differ from 3G best practices. Tools such as
AT&T ARO and websites such as 3GPP continue to be vital to ensure apps incorporate
the best practices for app development.
Due to the vast amount of detail that could be written around each best practice, this
paper intentionally stays at a high-level. However, developers are encouraged to take a
deeper dive into these topics using the references section.
10. FAQ (Frequently Asked Questions) about AT&T 4G App Development

1 . A r e a ll 4 G f la v o r s id e n tic a l?
No. There are several different flavors marketed as 4G such as:

4G Type Definition Download (Mbps) Upload (Mbps)
HSPA+ High Speed Packet Access 3.7 1.2
WiMax Worldwide Interoperability for 3.9 0.8
Microwave Access
LTE Long Term Evolution 23 – 61 15 - 23
Regardless of the underlying technology and marketing terms used, the key
differentiators between 3G and 4G are speed and energy consumption. According to
ITU (International Telecommunication Union) standards, a 4G network requires peak
data speeds of 100 Mb/sec.
2 . H ow do I te s t th e s pe e d of m y h a n ds e t’s 4 G n e tw or k ?
For a manual test, download a speed test app from your phone’s app store (e.g.
www. speedtest.net). Make sure you turn-off Wi-Fi before running your speed test
since speed test apps measure Wi-Fi speeds if enabled.
3 . H o w d o I te ll w h ic h c e llu la r n e tw o r k ( 2 G , 3 G , 4 G H S P A + o r 4 G
L T E ) m y h a n ds e t is c o n n e c te d to ?
Your cellular network is usually, but not always, identified by the icon next to the
bars or signal indicators on your smart phone’s screen.
For example, in the U.S. on AT&T’s HSPA+ network with an iPhone, you may get 4G
displayed. But outside the U.S., on a similar HSPA+ network, 3G may be displayed
instead. Some U.S. only examples are:
AT&T Network Icon Displayed
1G 1G or none
2G 2G or none
2.5G G or GPRS
2.75G E or EDGE
3G 3G
4G HSPA+ 4G or H+
4G LTE 4G LTE
11. Terms and Acronyms

Acronym Definition
ST nd rd th
1G, 2G, 3G, 4G 1 , 2 , 3 and 4 generations of mobile networks
3GPP 3rd Generation Partnership Project
API Application Programming Interface
ARO Application Resource Optimizer
bps Bits per second
CSD Circuit Switched Data
CSFB Circuit Switched Fallback
EDGE Enhanced Data for GSM Evolution
E911 Enhanced 911
FTC Federal Trade Commission
GB Gigabytes
GPRS General Packet Radio Service
GSM Global System for Mobile Communications
GSMA GSM Association
HD High Definition
HSDPA High Speed Download Packet Access (3G)
HSPA High Speed Packet Access (3G)
HSPA+ High Speed Packet Access Plus is the faster version of HSPA
IMS IP Multimedia Services
IP Internet Protocol
IPv4 Internet Protocol with four octets (xxx.xxx.xxx.xxx)
IPv6 Internet Protocol with eight chunks
(xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx)
ITU International Telecommunication Union
LTE Long Term Evolution
LTE-A LTE Advanced
M2M Machine to Machine
MIMO Multiple Input, Multiple Output
MMO Massively Multiplayer Online games
ms millisecond
OS Operating System
PCRF Policy & Charging Rules Function
QoS Quality of Service
RF Radio Frequency
SDK Software Development Kit
SMS Short Message Service
UX User Experience
VoLTE Voice over LTE
WAP Wireless Access Protocol
References
1
Mobile apps one of three biggest drivers. http://goldengekko.blogspot.com/2010/11/mobile-apps-one-of-3-
biggest-drivers.html
2
Data Deluge. http://gigaom.com/broadband/despite-critics-cisco-stands-by-its-data-
deluge/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+jkOnTheRun+%28GigaOM%
3A+Mobile%29
3
Global Mobile Data Traffic Growth.
http://ibmtelconewsletter.files.wordpress.com/2011/02/wp_closing_mobile_data_revenue_gap_a4.pdf
4
AT&T LTE speeds & latencies: http://www.lightreading.com/document.asp?doc_id=212403
5
LTE speeds: https://developer.att.com/developer/forward.jsp?passedItemId=2400016
6
QoS Services: http://developer.att.com/home/community/conference/LTE_Implications_of_Services.pdf
7
PCRF examples: http://whitelassiblog.wordpress.com/2011/02/16/policy-control-and-qos-awareness-for-ims-
and-lte-an-evolving-trend/
8
U.S. rapidly becoming a nation of gamers: http://www.digitaltrends.com/gaming/study-us-rapidly-becoming-a-
nation-of-gamers/
9
AT&T’s Application Resource Optimization Optimizer (ARO) is downloadable here
10
M2M needs LTE: http://m2m.tmcnet.com/topics/m2mevolution/articles/251313-why-m2m-needs-lte-vice-
versa.htm
11
M2M needs LTE: http://m2m.tmcnet.com/topics/m2mevolution/articles/251313-why-m2m-needs-lte-vice-
versa.htm
12
Why LTE sucks (your battery): http://gigaom.com/mobile/why-lte-sucks-your-battery-that-
is/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+OmMalik+%28GigaOM:+Tech%29
13
Handovers from LTE to HSPA+:
http://www.accuver.com/storage/downloads/AccuverResource_attdrivetestresultsandreportpreview_1326383911.pdf
14
More Fuel for the Signaling Storm: http://lteconference.wordpress.com/2012/04/20/new-ipad-more-fuel-for-the-
signalling-storm/
15
Want to give your apps a boost? http://developer.att.com/home/community/conference/Boost_App.pdf
16
Avoid Network Meltdowns. http://www.ospmag.com/issue/article/Avoid-Network-Meltdown
17
AT&T’s Application Resource Optimization Optimizer (ARO) is downloadable here
18
Top wireless issues and their solutions.
http://developer.att.com/home/develop/referencesandtutorials/networkapibestpractices/Top_Radio_Resource_Issues
_in_Mobile_Application_Development.pdf
19
AT&T data calculator: http://www.att.com/standalone/data-calculator/index.html
20
Use Wi-Fi as much as possible. http://developer.att.com/developer/forward.jsp?passedItemId=9700190
21
A Call for More Energy Efficient Apps: Minimize battery drain & reduce latencies
http://www.research.att.com/articles/featured_stories/2011_03/201102_Energy_efficient?fbid=bFz29dRdhkG
22
Reduce inefficient connections with periodic transfers.
http://developer.att.com/developer/forward.jsp?passedItemId=9700174
23
Handle multiple simultaneous TCP connections.
http://developer.att.com/developer/forward.jsp?passedItemId=9700172
24
Handle duplicate content and caching. http://developer.att.com/developer/forward.jsp?passedItemId=9700172
25
Utilize pre-fetching. http://developer.att.com/developer/forward.jsp?passedItemId=9700172
26
Transition to IPv6: https://developer.att.com/developer/forward.jsp?passedItemId=3100042
Note: Some developer.att.com URLs require developer account credentials in order to login and access the content.

Best Practices 3g 4g App Development PDF

Uploaded by

Copyright:

Available Formats

Best Practices 3g 4g App Development PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Best Practices 3g 4g App Development PDF

Uploaded by

Copyright:

Available Formats

What are some of the benefits of 4G networks compared to 3G?

What are some of the benefits of 4G networks compared to 3G?

How can developers optimize apps for better performance on mobile networks?

How can developers optimize apps for better performance on mobile networks?

© 2012 AT&T Intellectual Property Best Practices for 3G and 4G App Development

Best Practices for 3G and 4G

ii) Cache more often .................................................................. 14

4G Type Definition Download (Mbps) Upload (Mbps)

LTE Long Term Evolution 5 100 50

Environment Type Latency (ms)

Cellular (AT&T) 4G HSPA+ 100 - 200

Wi-Fi (phone) 802.11n 50

Cellular (AT&T) 4G LTE 40 - 49

Note: Latencies in this table are for round-trip requests and

resources7. Currently, PCRF is controlled by carriers. Some examples of how

5. LTE Enhanced Applications

In wireless networks channel quality is measured by the amount of

On occasion, LTE effective data rates may fall below users’

As new technologies are introduced into a carrier’s network, developers need to

Continuous Timer expiration

A common aspect of modern mobile app development is the use of drop-in

Data cap Battery life Speed Robustness

i) Avoid large thumbnails

ii) Provide configurable data transfer settings

 Switch to Wi-Fi rather than 3G to conserve battery.

iii) Use Wi-Fi as much as possible 20

Dual and quad-core devices with multiple antennas (MIMO) on

i) Avoid constant polling 22

ii) Download in bursts rather than continuously streaming 23

iii) Minimize aggressive behavior

i) Decouple user transitions from data interactions

ii) Cache more often 24

iii) Flush the cache regularly

iv) Create one session with multiple gets

Apps will have to deal with multiple real-world scenarios such as

i) Plan for a mix of networks

2.5G EDGE 4G HSPA+

Networks differ in bandwidth, latency, throughput, reliability, and security. Even

ii) Use bigger buffers

iii) Support IPv6

• AT&T Labs research (www.research.att.com).

• AT&T’s Developer Program (http://developer.att.com).

A number of recent initiatives have been directed at helping developers write

recommendations that allow you to optimize performance, make battery

10. FAQ (Frequently Asked Questions) about AT&T 4G App Development

No. There are several different flavors marketed as 4G such as:

11. Terms and Acronyms

You might also like