| My area of interests
are Computer Networks and Artificial Intelligence. My projects
in college include Tamil Character Recognition, Analyzing
a Robotic Football field, Patient Monitoring and Offline
Mail Client. My final year project was on Wireless Network
Management and Efficient Hand-offs. Our work was published
in the 12th IEEE conference on HIPC 2005. I interned with
the MSN Connector team in Microsoft IDC after my 2nd year
on building a generic Outlook Connector. I was one of
the 8 undergraduates from India to be selected last year
to do summer internship at Microsoft Research Redmond.
I worked with the Networking Research group there for
12 weeks on the DAIR project. Our work has been published
in ACM Hotnets 2005 and in Mobisys 2006. We have also
filed for two patents. |
| |
| My span of interests
extend far beyond academics and research. I headed the
Computer Society of Anna University, the largest technical
student organization in the campus, which aims at bringing
the latest in the field of computers to students of all
disciplines. In the Asia Regional Finals of the prestigious
ACM Contest my team finished at 5th place. |
| |
| My hobbies include
playing cricket, writing poems, composing computerized
music and doing lot of strange things. You can know more
about me @ http://www.geocities.com/leninravindranaths |
| |
| In Code4Bill, I work
with Alec Wolman, Jitu Padhye and Ranveer Chandra
of the Networking Research team in Redmond on the WiFiAds
project. |
| |
| I intend to pursue
a career in research in the future. After Code4Bill I
will continue to work at MSRI as Assistant Researcher
for a year or two before I go for a PhD. |
| |
| |
A Novel Approach
for Delivering Location-Sensitive Advertisements Using
Wi-Fi Networks
The tremendous growth of Wi-Fi networks in recent years
provides a unique opportunity for delivering location-sensitive
advertisements (and other location-specific information)
to users. Several techniques for determining a user’s
location are currently in use. For example, MSN has introduced
the “Locate Me” feature, which determines
the approximate location of the user based on which Wi-Fi
networks can be overhead by the user’s Wi-Fi device. |
| |
| The existing approaches
to delivering location-sensitive advertisements require
two capabilities: the user must have a reasonable quality
connection to the Internet, and there must be an automatic
method for indicating the user’s location to the
ad delivery service. Both these requirements can be onerous
in many environments. For example, it may be very difficult
to maintain an Internet connection in many of the locations
where one wants to receive ads, for example while going
for a drive.. |
| |
| We propose a new scheme
for distributing location-sensitive advertisements to
Wi-Fi devices. Our approach relies on making clever use
of certain features of the Wi-Fi protocol. It offers three
main advantages over existing approaches for delivering
location-sensitive ads. |
| |
| First, our approach
does not require the client device to reveal any information
in order to receive location-sensitive ads. Second, our
approach does not require the client to have Internet
connectivity. In fact, we can deliver ads even when the
client is connected to the Internet via a competitor’s
Wi-Fi network. Third, our approach can allow the advertisers
to supply dynamic information to consumers in real-time.
For example, using our approach, a popular restaurant
can continuously advertise the expected wait time to all
wireless clients in its vicinity. |
| |
Broadcasting
Ads over Wi-Fi
Our approach is based on a “push model” of
ad delivery. Our key idea is to overload IEEE 802.11 beacons
to carry ad messages. Beacon packets are normally used
just to announce the presence of a Wi-Fi network. As a
result, a Wi-Fi client receives the beacons sent from
all nearby APs, regardless of whether the client is connected
to any Wi-Fi network. In fact, even when the client is
connected to a specific AP, it periodically scans all
the channels to receive beacons from other nearby APs
to keep track of other networks in its vicinity. The client
doesn't need to transmit anything to receive the beacons;
it merely has to listen. |
| |
| Our basic idea is to
modify certain fields of the beacon packet to carry ad
messages. We treat the advertisement as an arbitrary string
of bytes. In most cases, we expect the ad message to be
a text phrase. However, our techniques could also be used
to deliver short audio jingles or video clips. |
| |
| We are currently implementing
three different approaches to embedding ad messages into
beacons: |
| |
| 1. |
SSID Concatenation: The SSID field in the beacon
identifies the wireless network, such as MSFTWLAN.
Its maximum length is 32 bytes. |
| 2. |
BSSID Concatenation: BSSIDs are 6 byte unique identifiers
of an AP, and can be set to any value. |
| 3. |
BIE
Concatenation: The 802.11 standard allows AP vendors
to add up to 253 bytes of vendor specific information
in the Beacon Information Element (BIE) field of
the beacon. We use this feature to add a Wi-Fi Ads
BIE for sending ads. |
|
| |
Preventing
fraud and attacks
We need to prevent two types of attacks on our Wi-Fi Ads
system. The first is forging. For example, attackers can
set up APs and send fake ads. For example, someone could
set up an AP next to Starbucks, which advertises “Starbucks
is closed for the day”. The second threat against
our system is replay attacks. In the same scenario as
above, someone could replay an ad that was sent by Starbucks
at an earlier date. We use standard cryptographic techniques
to counter both these attacks. The ad message and its
time of validity can be encrypted by MSN with its private
key and the encrypted message is broadcasted by the APs.
Wi-Fi clients then use MSN’s public key to decrypt
the ad, and display the ad along with the time of validity
to the user. |
