Wi-Fi

• Group Memebers
– Muhammad Tariq
– Muhammad Sohail
– Sajjad Hussain Shah
– Wajid Ali
 Introduction
• Short for wireless fidelity and is meant to
be used generically when referring to any
type of 802.11 network, whether 802.11b,
802.11a, 802.11g, dual-band etc
• Wi-fi is a wireless technology that uses
radio frequency to transmit data through
the air
 Brief History
• IEEE (Institute of Electrical and Electronics
Engineers) established the 802.11 Group in
1990. Specifications for standard ratified in 1997
• Initial speeds were 1 and 2 Mbps
• IEEE modified the standard in 1999 to include
802.11 a and b
• 802.11g was added in 2003
• 802.11b equipment first available, then a,
followed by g
• IEEE create standard but Wireless Ethernet
Compatibility Alliance certifies products.
 Research Paper-1
Paul S. Henry, Hui Luo, “WiFi: What’s Next”,
IEEE communication Magazine, December
2002.pp. 66-72.
 Abstract
• WiFi, also known as 802.11b, has become
the preferred technology for wireless local
area networking in both business and
home environments.
• WiFi design and Deployment
• Challenges for future market opportunities
– Technical
– Business related
 Introduction
• WiFi is now available as a built-in feature in a
wide range of user devices.
• This article give a clear picture of inexpensive
WiFi hotspots offering easy, and often free,
broadband Internet access to anyone equipped
with a WiFi-capable laptop or PDA.
• Extension of WiFi from the office environment to
wide-area coverage opens new areas for WiFi
technology
 Cont…
• To make WiFi networking an everyday tool for
road warriors, as commonplace as their cell
phones, major enhancements to existing WiFi
technology are required.
• The challenges ahead for WiFi engineers can be
grouped into four broad categories:
– Ease of use
– Security
– Mobility
– Network management.
 Ease of use
• Simplicity of operation has been an important
concern for WLAN.
• Windows XP offers “automatic wireless network
configuration,” for automatic connection
establishment with nearby WiFi access points
(APs). But Initial setup is not simple.
• In a typical academic environment, 18 steps are
required for initial configuration.
• So expected more errors and user frustration.
 Cont…
• In the corporate environment, due to data security
concern, this is more burdensome.
• Road warriors want to use a commercial WiFi
hotspot.
• There must be Access Control Mechanism for
hotspot operators to serve the following purpose
– Authorizing existing subscribers
– First-time users to sign up and one-time users to pay
on site.
 Browser-based Authentication
• Serves both these ends, is convenient and user-
friendly
• Vulnerable to theft-of-service attacks
• 802.1x standard protects against this and
several other attacks
– This requires a WiFi prearranged service account.
• Until provision of user-friendly setup and secure
hotspot sign-on, it will be unable to live up to its
full potential.
 Security
• Both on corporate premises and off, WLANs represent
potentially serious security vulnerabilities.
• Common scenario involves a hacker sitting in a parking
lot listening to the WiFi communications of a nearby
retail establishment (May capture credit card number)
• WEP (wired equivalent privacy) encryption key can also
be recovered
• So no bullet-proof security even WEP is on or off.
• Two approaches
– Native Security
– Virtual Private Network
 Native Security
• Mutual authentication b/w Network & User
• It relies on a database of authorized users
• Permits WLAN access only to those who
properly authenticate themselves.
• 802.1x support for key exchange between the
endpoints of a WiFi air link using Temporal Key
Integrity Protocol (TKIP).
• Two challenges to 802.1x/802.11i approach.
– Time to market
– Convincing doubtful System administrators
 Virtual Private Network
• Assumes that the
WLAN is insecure
regardless of whether
WEP is on or off.
• Separate security
mechanism IPsec
tunnel provides end-to-
end protection.
 Cont…
• VPN technology provide strong security, but the cost is
significant, especially for corporate WLANs.
• Since the (Access Point) APs and associated wiring are
assumed to be insecure, they must be kept physically
separate from the existing premises wired LAN.
• That is, a distinct, independent wired network must be
installed and maintained.
• All WiFi communications, even traffic confined within the
corporate premises, must be processed by a VPN
gateway, scaling to large numbers of users is difficult.
 Cont…
• Enhanced WEP, has the advantage of fitting
naturally and simply into the WiFi infrastructure
• But market acceptance may be disappointing due to
standardization delays and failure of original WEP
security could be repeated.
• VPN is an established and trustworthy security
solution, but implementation complexities, especially
on corporate premises are unacceptable.
• Another WiFi concern issue is “rogue APs”
(Discussed in Network Management)
 Mobility
• Extending it to public hotspots with secure
always-on connectivity is far more difficult.
• To support the road warrior similar to
global roaming capability like in cellular
system.
• Requires
– Technology
– Service Providers
 Technology
• Device-level multivendor interoperability.
• Reliable interoperation achieved through the WiFi
initiative under the Wireless Ethernet Compatibility
Alliance (WECA), is not guaranteed by IEEE 802.11b
standard.
• Another WECA initiative known as Wireless Internet
Service Provider roaming (WISPr) provide a uniform
mechanism for handling the authentication,
authorization, and accounting (AAA).
• The missing component is always-on mobility, secure
wireless connection with “close and go, open and
resume” operation.
 Mobile IP (MIP)
• Always-on mobility can be implemented via Mobile IP
(MIP).
• Together, the manager and client create the MIP routes
• Sessions do not get broken when the user suspends
operation and handing off from one subnet to another.
• Mobile IP, together with VPN, provides a conceptually
satisfying approach to the always-on connectivity.
• This capability is not currently available in operating
systems (OSs) such as Windows or Mac because both
MIP and VPN are layer 3 approaches.
 Service Provider
• Entrepreneurs must start small and arrange to grow
quickly to profitability, so cellular operators would see the
value in owning a large-scale WiFi network and buy them
out.
• Many WiFi service entrepreneurs are attempting to look
up prime sites for hotspots (airports, hotels, etc.) to
maximize their future revenue potential.
• Three major approaches
– The Franchisor
– WiFi Carrier
– The Aggregator
 The Franchisor
• The franchisor provides software and back
office operations to allow the franchisee to
offer paid public access to the network
• A franchisor may strikes an agreement
with an individual who already has WiFi
installed for its own business.
 The WiFi carrier
• Some small carriers, like SurfAndSip, are
“pure play” WiFi operators whose sole
product is WiFi service.
• Larger carriers, like Wayport, tend to offer
a variety of services, like wired broadband
access.
• Other WiFi carriers are for parent
communication company.
 The Aggregator
• The aggregator strikes wholesale
partnerships with WiFi operators and
resells their services.
• If WiFi public access shows signs of
economic viability, cellular operators will
likely be interested in participating.
• But at this early stage it is not clear if
building an independent infrastructure for
WiFi is economically viable.
 Network Management
• Managing the network to ensure even
minimal physical-layer performance delivery
of adequate signal strength with an
acceptable level of unwanted interference is
a major problem.
• Three major issues:
– Signal Strength
– Interference Management
– The Rogue AP
 Signal Strength
• Change in the location of a user can cause
signal strength variations of 30 dB and more.
• For a small network, consisting of a dozen or so
APs, signal strength issues can be managed by
a hands-on administrator.
• For large networks consisting of hundreds or
even thousands of APs, personal familiarity with
network environment is practically impossible.
 Interference Management
• WiFi networks use unlicensed spectrum.
• WiFi networks are also vulnerable to other sources such
as Microwave oven and other devices operating in
nearby frequency band networks which are not under the
WiFi operator control.
• It will become increasingly severe as WiFi proliferates.
• Management tools must be developed to enable system
administrators to monitor interference.
 Rogue AP
• An unauthorized AP attached to the corporate intranet is a
major security vulnerability.
• Regardless of WEP or VPN, a rogue AP can expose internal
corporate communications to the outside world.
• Detecting rogues on corporate premises through sniffing and
pinging techniques may probably manageable burden for
network managers.
• More difficult will be the problem of detecting rogues in
employee residences.
• Tools and techniques must be developed for security is not
being compromised by rogues, whether on premises or off.
 Summary
• In this article we explore what will be required to make
ubiquitous broadband mobility vision a reality.
• We focus on WiFi service to support the travelling
professional.
• Presented key challenges to WiFi evolution.
• Technical problems, although by no means trivial, seem
manageable, and the proposed approaches to solution
are promising.
• Much important and troubling, are the business issues.
• The economic viability of the public hotspot market is still
unclear.
Samuel H.Russ, Member , IEEE and Sasan
Haghani, Member IEEE, “802.11g Packet-
Loss Behaviour at High Sustained Bit Rates
in the Home”, IEEE Transaction on
Consumer Electronics, vol.55, No.2, May
2009.pp.788-791.
 Abstract
• 802.11g is a popular way to transmit data throughout
homes wirelessly. However, common broadcast
quality video bit rates, up to 20 Megabits per second,
represent difficult challenges for wireless networks in
terms of latency, packet loss, and packet jitter.
• An in-home network was sampled between 0.5 and 2
Megabits per second (Mbps) and analyzed with
respect to packet loss. Our experimental results
indicate that the distribution of the number of
consecutive error bursts is a combination of the classic
GE model and A long tail model.
 Introduction
• IEEE 802.11g as successful method for
transmitting data on wirelessly in home.
• The purpose of this paper is modification
in-home video distribution and measure
the range to which the 802.11g channel is
suitable for it.
• Assumption in the paper
– Bandwidth-reservation mechanism
– One Directional traffic flow
 Cont…
• Major Sections in this paper
• A survey of previous research results
relevant to the field of packet loss in 802.11
networks is presented.
• Documents the assumptions and the
experimental setup.
• Presents the results of the statistical analyses
• Conclusions and Suggestions for future work
are presented in Section
 SURVEY OF PREVIOUS RESEARCH

• Gilbert-Elliot (GE) model.

• The classic channel model for the number
of consecutive arrived and lost packets is
the Gilbert-Elliot (GE) model.

P[ X = n] = α n −1 (1 − α )
 Cont…
• In [2] and [3] measurements were made of 2-way real-
time traffic over real 802.11g connections using
artificially generated data streams, two-way traffic, and
UDP connections.
• The collected data was relatively low bit rate then the
required broad cast video.
• First, the papers demonstrate a maximum bit rate of 256
kilobits per second (kbps), far below what is needed for
broadcast-quality video (e.g., MPEG2 HD video requires
up to 20 Megabits per second (Mbps), Second, the
method in [2] considers two-way traffic, and this paper
only considers one-way, time-reserved traffic as it is
more reflective of an engineered video network.
 ASSUMPTIONS AND EXPERIMENTAL
PROCEDURE
• A platform was designed to emulate an in-home
video distribution system by routing the “video”
from the access point to a station.
• Insure no other traffic on the network.
• Distance b/w sender & receiver was 35 feet
away.
• Utilities used for transmission
– Iperf
– Windump
– Tcpdump
 Cont…
• The experiments was conducted on different
data rates of
- 0.5, 1 and 2 mbps.
- Time frame for 1 and 2 mbps
one hr for each
- Time frame for 0.5 mbps
4 hrs
Final transmission results
- 1.5 million packets
 Measurement and analyses of
result
• It has been noted that wireless channels
tend to have bursts of packet errors, and
an important factor in modeling channels
is the ability to estimate the number of
packets that are dropped when a burst
occurs. After analyzing the data, the
results are summarized below, in terms of
error-burst probability versus size
Cont…
 Cont…
• The relative occurrence of error bursts of
different sizes. The horizontal axis is the
number of packets dropped in a single
burst and the vertical axis is the probability
that a burst of that size occurred. Note the
similarities of the streams at 1 to 3 packets
dropped
 Cont…
• Figure 2 shows the probability of a burst of
n consecutive bad packets, P[X=n], as a
function of n, the number of dropped
packets for bitrates of 0.5, 1 and 2 Mbps.
For example, Figure 2 shows that at 2
Mbps, 98.2% of all error bursts had a
length of 1 packet.
Cont…..
 Cont….
• Figure shows the number of packets
dropped in each second of the 2 Mbps
test. The four large packet-loss bursts, as
shown in the figure, occurred at 592, 1115,
1691, and 2330 seconds into the test,
suggesting that these four bursts occurred
at random intervals.
 Cont…
• Figures 2 and 3 suggest that the
distribution of burst sizes is actually the
confluence of two different effects. For n≤3,
P[X=n] follows the GE-model distribution.
For n>3, P[X=n] is a distribution with a
much longer tail. These results suggest
that the best channel model for 802.11b/g
networks is a hybrid model, combining the
GE model with a long-tail model of packet
loss.
 Cont…
• Long tail physical possible explanation.
- Router itself
- Outside interference
Access points senses high data rate counter.
• The experimental results were analyzed with
respect to packet inter-arrival time. For each
packet, the interval of time between its arrival
and the arrival of the previous packet was
computed, and the intervals were sorted by time.
In Figure, the Y axis shows how often an inter-arrival gap of the
indicated number of milliseconds occurred during a 1- hour test.
Small gaps in time are due to the data rate in use. They have
been deleted from the histogram (the X axis starts at 50 ms) to
provide better resolution
A magnified version of the histogram is shown below in figure
 Cont…
• There are four possible explanations for the
paired delays, namely outside interference, the
sending laptop (and/or the iperf software), the
router, and the receiving laptop.
• Paired delay did not result in loss of packet.
• Paired delay are caused by outside interference
• Paired delay are not an explanation for the long
tail distribution of error burst.
• If the paired delay were a function of brand of
router.
In order to rule out the sending laptop and the iperf
tool, the receiving laptop was set to play a video
from You Tube via the same wireless network
 Cont…
• The figure shows that the paired delay cannot be
caused by the sending laptop or the iperf tool.
• One remaining possible explanation for the paired
delays is the receiving laptop.
• The packet logs did not show any traffic from the
receiving laptop (e.g. ARP packets) during the
paired-delay events. In addition, traffic was logged
when the laptop was idle, and there was no traffic
occurring at 10- second intervals. These
observations rule out the receiving laptop as the
cause of the paired delays.
 Cont…
• Our results indicate that the paired delays
are caused by a process in the router, and
that multiple vendors’ routers exhibit the
same behavior. The effect must a
consequence of some aspect of the
802.11 standard.
 CONCLUSIONS AND FUTURE WORK

• Our results show an unexpected behavior of

802.11 networks that cause loss of the channel
for several hundred milliseconds every 10
seconds. This behavior is unacceptable for high-
bitrate video transmission.
• This work can be extended in two important
directions. First, additional testing is needed to
determine a statistical model for the long-tail
distribution. Second, tests at even higher
bitrates, closer to that of high-definition video,
are needed to validate the model.
 REFERENCES CITED
• [1] L. Wilhelmsson and L. Milstein, “On the effect of imperfect interleaving
for the Gilbert-Elliott channel”, IEEE Transactions on.Communications., Vol.
47, No. 5, pp. 681-688, May 1999.
• [2] L. Carvalho, J. Angeja, and A. Navarro, “A new packet loss model of the
IEEE 802.11g wireless network for multimedia communications,” IEEE
Transactions on Consumer Electronics, Volume: 51 Issue: 3, Page(s): 809-
814, Aug. 2005.
• [3] L. Carvalho, J. Angeja, and A. Navarro, “802.11G WLAN modeling for
real time packet communication,” 15th IEEE International Symposium on
Personal, Indoor and Mobile Radio Communications, 2004. (PIMRC 2004)
5-8 Sept. 2004, Volume 4, pp. 2958- 2962.
• [4] A. Bazzi, M. Diolaiti, and G. Pasolini,, “Measured performance of real
time traffic over IEEE 802.11 b/g infrastructured networks”, 2005 IEEE 61st
Vehicular Technology Conference, 2005. VTC 2005-Spring, 30 May-1 June
2005, Vol. 5, pp. 2885-2889.
• [5] http://dast.nlanr.net/Projects/Iperf/iperfdocs_1.7.0.php
• [6] http://www.winpcap.org/windump/default.htm
• [7] Private communication with Access Point Vendor, 2001.
Thanks