WiFi Direct

Usability and Security of Wi-Fi Direct and Hotspots

Adhoc networking was initially designed for
military application area. But adhoc networks have been
found also appealing for autonomous computing. The
adhoc mode of IEEE 802.11 (Independent Basic Service
Set (IBSS) has not been succesful due to several reasons.
Let us explore and compare two alternatives for
adhoc network formation in heterogenous environments:
Wi-Fi P2P also known as Wi-Fi Direct, and Wi-Fi Hotspot.
The comparison shows that there are usability, security
and performance reasons to favour Hotspot for application

In spite of the growing number of nomadic users,
IEEE 802.11 adhoc has not turned into a successful
technology. It has experienced a very discrete deploy-
ment in services and frameworks, and it has a very
reduced impact in the popular applications we use in
our computers and smartphones. We can cite three main
reasons for this lack of success: insufficient user support,
security and energy consumption.
1) Insufficient user support. Adhoc networking is an
alternative to infrastructures, and requires users to
configure adhoc networks by their own means.
This is not an easy process, mostly because of the
reduced support available from vendors software,
operating systems and manufacturers. Besides dif-
ferent users have to cooperate to set up the adhoc
network, each one facing her own operating system
and configuration tools. This is a challenging task
for normal users, and this lack of usability is
perceived by application developers. Application
developers do not address adhoc since it would
require them an extra effort in usability, reduced
at the operating system. This would be the only
way to meet standard usability requirements.
2) Insufficient security. Adhoc networks tend to be
set up with little or no security. The previously
explained reason (insufficient user support) makes
it difficult enough for users, which tend to leave the
network open or use insecure settings like WEP.
Even more, adhoc networks have been used by
attackers to bridge to an enterprise wired network.
Users may not even notice of the wireless adhoc
being up, while it is used to further scan and access
the rest of the enterprise network.
3) Energy consumption. This is the most severe issue
in 802.11 adhoc mode. It was not designed with
power efficiency and it rapidly drains the battery
of adhoc connected devices. It does not offer the
power saving schemes of managed modes.

Wi-Fi Allaince standardized Wi-Fi P2P v1.5 in 2014.
Also known as Wi-Fi Direct, it allows for dynamic
discovery of peers and services as required by nomadic
users. It also offers the possibility of two power saving protocols:
Opportunistic Power Save and Notice of
WiFi P2P consists of several phases: discovery, group
formation, and group operation. It starts with an exchange of
probe request/response frames conveying the
so called Information Elements (IE) which allow to
discover the responding peers. An optional service dis-
covery phase follows where different service discovery
protocols (UpnP, Bonjour, WS-Discovery, etc.) may be
used according to the Generic Advertisement Service
(GAS) protocol/frame exchange as defined in IEEE Std
If the users are interested in the discovered services
and peers, they can go for the next phase, where invitation
requests and responses are exchanged to achieve the
group formation. The formation of the group requires
a negotiation and a provisioning phase. The first is
the negotiation of which peer will become the Group
Owner (GO). This is achieved exchanging GO negotia-
tion requests, responses and confirmation frames. The
provisioning phase uses Wi-Fi Simple Configuration
(WSC) to set up the new wireless network with the
desired WPA2 encryption between the peers.
If the user discovers an existent group, she can ask to
be invited using an identical procedure as the previous
paragraph, excluding the GO negotiation. So the peer
request has to be approved by the GO and follow the
provisioning phase according to the WSC.
Once the group is formed, the description of the group
includes the Information Elements of the nodes and of
the group itself and its related capabilities. In particular
the possibility of direct communication between the
peers (indicated by the Intra-BSS Distribution field).

Most of the smartphones offer the possibility of acting
as AP, routing the traffic between the wifi interface and
other network interface (3G, 4G, etc.). The most com-
mon use case is offering Internet connectivity to other
devices which are dynamically configured by DHCP
server running at the smartphone. Most modern computer
operating systems also offer this possibility, though they
typically route between the wifi and a ethernet interface.
This is a common scenario for nomadic users, though
it may require explicitly starting aditional discovery
servers and clients for other services besides networking.
A device may associate to a Hotspot using either the
WPA2 Passphrase, or WSC.


After lots of lab hours we were able to set up
direct connections between linux and android devices (smartphones and tablets).
Many tests and experimentations lead us to appreciate
the superior usability and increased security of the Hotspot scenario.
This scenario is easier for application developers to dynamically
provide with services in ubiquitous scenarios.
From the usability point of view, the Hotspot required
a very minimal configuration, something logical since
manufacturers have included this use case in the smart
devices software support and configuration.
From the security point of view, though the Wi-Fi P2P
can be configured not to use PBC in Linux, this is not the
case for android devices, though applications may use
more secure configurations (PIN or NFC tags), which
can also be used for the Hotpot.
The service discovery facility of Wi-Fi P2P also
showed very difficult to configure and we left it out of
the experimentation. On the other hand, specific service
discovery can be run in Hotspot and they only require
being triggered by the user.

About Andres Marin Lopez

Associate Professor in the Telematics Department at the Carlos III University of Madrid. In 1997, he received a PhD in Telematics Engineering from the Polythecnical University of Madrid, Spain. His research interests include pervasive computing, trust and security in next generation networks. He is a member of the Pervasive Computing Laboratory (pervasive.it.uc3m.es) at the GAST group (www.gast.it.uc3m.es). Contact him at amarin@it.uc3m.es.

Leave a comment

Your email address will not be published. Required fields are marked *