Abstract :
The International
Data Corporation (IDC) indicates that the most popular activities of mobile Internet
users are search engine querying, reading news information, downloading multimedia,
and using email and instant messaging clients. Over the next four years, IDC predicts
the fastest- growing activities will be purchasing, social networking,
and blogging.1 In essence,
most services available through desktops
will soon be accessed mainly through a large quantity (and variety) of mobile devices.
Exciting systems and networking innovation is happening on these devices,
and it is a good time to be a mobile computing researcher. At the operating system
level, we have seen the rise of the Linux-based Android platform gaining market
share at the expense of the closed-source iPhone. At the same time, the App Store
launched by Apple contains more than 200,000 applications that users can download
for their phone. Google has followed suit with an equivalent (although less controlled)
store for Android phones. Regardless of the authority model — on which the jury
is still out — the application store model is here to stay. While previous application
models involved service providers and phone producers only, now the service market
is open to third- party developers who might even be the users themselves.
Clearly, this field is burgeoning and developing in many directions,
driven by users’ interests in interacting and “staying connected.” With this in
mind, it is worthwhile to stop and reflect on users’ future needs and expectations
as well as the challenges that this reality brings and the opportunities offered.
Upcoming Challenges
One interesting
observable mismatch is that users expect the same level of service and performance
on their phone as they experience on their desktops. For example, low latency and
high responsiveness, ease of interaction, and always-on connectivity are the implicit
user requirements. It is also, however, true that current devices’ limitations are
impacting the way individuals use devices (skimming through email rather than reading
carefully, and producing very short answers). This, in turn, also induces new behavior
in users. For instance, consumers who bought the iPad sued Apple because the device
was shutting down after overheating, leaving
users disappointed but still determined to own a device they can use to read the
news from a deck chair on the beach.
There are a
number of other challenges, as well, for mobile application and network developers
that need attention. Each of these challenges must be solved before users’ expectations
can be met.
Energy :
Energy seems
to be a permanent issue with these devices. Screens are getting bigger and more
colorful, functionality is increasing, as is the number of network interfaces and
sensors available and used by applications. Several of the most recent smart phones
last no more than 12 hours without recharging, when continuously connecting to the
network. While some of these issues will be solved at the hardware level by use
of more sophisticated batteries or energy-scavenging solutions, there is also a
need for improved software that works effectively (for example, through a smart
sensor and with efficient use of network connectivity).
Bandwidth :
Downloading
and streaming video and audio content seem popular. Unfortunately, even today,
the current cellular network infrastructure is put under considerable pressure,
and it seems bound to become worse. Research is happening on alternative techniques
for smarter content caching or opportunistic use of local connectivity links (such
as downloading interesting content from your neighbors).
Intelligent switching between interfaces, which would let users seamlessly exploit
networks with more bandwidth when available (such as Wi-Fi), is also a promising
idea.
System :
Given the availability
of multiple applications developed by different sources, there is a need to engineer
the device system in ways that allow for responsiveness and concurrent execution
while considering the limited resources and different user interfaces available.
At the same time, the high number of mobile devices and high frequency of use (because
of the device’s constant availability) imposes a load that burdens not just the
network but also the back-end servers. Hence, we need customized techniques that
take advantage of the knowledge of device location, access patterns, and user behavior.
User Interface
:
The way people
use services and devices cannot be forecast with precision. The success of short
message service (SMS) was an unpredicted event. It is apparent that the interaction
with systems happening from mobile devices is different from the one on desktop
machines. Work on voice interfaces is promising but not yet ripe, although it has
the potential to considerably change the kinds of services
and ways in
which services can be offered to users. The abundance
of challenges for hardware and software research
in this area are therefore appealing, and virtually all services offered now on
the Internet need to consider the likelihood that users will soon be accessing them
through a mobile device.
Security
and Privacy :
While mobile
malware hasn’t captured too many headlines yet (with a few recent exceptions), it
is clear that with more applications being deployed, more businesses placing phones
at the core of their interest, and more data being collected through these devices,
the interest of criminals is bound to increase.
Moreover, mobile devices can be stolen and tampered with while the owner is not
present in easier ways than with desktops and servers.
Integration
and Standardization :
We are at a
stage where various forces are at work to gain a monopoly on the mobile services
market. Although it is unclear how this will develop, as part of the academic community,
I have the hope that some sort of open market solution will dominate. However,
with open solutions come the challenges of standardization, heterogeneity, and interworking
of different technologies and hardware: Will desktop computer history be able
to teach us something?
Download :
Download :