Panels
Oct 18, 4:00pm - 5:00pm: Can WiFi hotspots do without
QoS?
Oct 19, 4:00pm - 5:00pm
Oct 20, 10:30am - 11:30am
Oct 18,
4:00pm - 5:00pm:
Can WiFi hotspots do without QoS?
Panelists:
Vojislav B Misic (University of Manitoba, Canada) - Moderator
Magda El Zarki (University of California, Irvine, USA)
Puneet Sharma (Hewlett-Packard, USA)
Xi Zhang (Texas A&M University, USA)
With increasing deployments of WiFi hotspots, wireless data access is becoming a
reality. Driven by high bandwidth, low cost, and inbuilt WiFi capability in a
variety of hand-held devices, the WiFi market is rapidly expanding. With
increased demands for bandwidth, mechanisms for QoS are needed for proper
sharing of critical resources. In addition applications such as VoIP, HDTV,
multimedia streaming, interactive video, and gaming, demand better QoS support
from the network.
However, today's WiFi standards such as IEEE 802.11a and 802.11g, can provide
raw data rates of up to 54 Mbps, and even higher bandwidths are possible using
other proprietary solutions. Due to low hardware costs of access-points, it is
feasible to deploy WiFi using a large number of small cells with very high
throughput. If raw bandwidths can be very high and frequency re-use is feasible
with smaller cells, is there a real need for supporting QoS in WiFi networks?
The panelists will present their opinions and arguments for research on QoS for
hotspots. Based on recent advances in the ever-emerging 802.11e standard for QoS,
the panelists will give insights on what needs to be researched in this area.
Oct 19,
4:00pm - 5:00pm:
QoS in the Internet: How
strong is the need?
Panelists:
Balakrishnan Prabhakaran (University of Texas at Dallas, USA) - Moderator
Jun-Hong Cui (University of Connecticut, USA)
Matt Mutka (Michigan State University, USA)
Sudhir Aggarwal (Florida State University, USA)
The Internet is evolving to support Diffserv to meet the QoS demands of
tomorrow's applications. While it is a scalable and deployable approach, will it
survive the networking demands over the next decade and beyond? Will the
applications of tomorrow make us revisit the architecture for supporting QoS in
the Internet? Would we have to succumb to an Intserv like approach or something
in between?
What are the immediate and long-term areas for research on QoS in the Internet?
Which applications will fail over the short-term and over the long-term, should
the Internet not support any form of QoS? Will support for QoS in the Internet
enable new applications or shall we wait for new applications to evolve before
designing solutions to meet their requirements? These and other open questions
will be addressed by the panelists to provide direction for future research on
QoS in the Internet in a lively and controversial discussion.
Oct 20,
10:30am - 11:30am:
QoS for sensor networks: Is
it an overkill?
Panelists:
H. Mouftah (University of Ottawa, Canada) - Moderator
Kalyan Basu (University of Texas at Arlington, USA)
Arjan Durresi (Louisiana State University, USA)
Numerous applications of sensor networks have been identified in diverse fields
including industry, defense, national security, and space. These networks are
characterized by nodes with little compute power, small memory, and limited
energy reserves. In such resource constrained environments, providing QoS for
proper allocation of resources is critical to meet the demands of the
application.
However, QoS enhancements increase the complexity of protocols, add significant
overheads, and effect critical metrics such as latency and power consumption. It
is against the sensor network philosophy of keeping the design "simple and
robust". This raises a question -- is designing systems and protocols for QoS
support in sensor networks an overkill?
The panel will share expert opinions on the value of research on QoS for sensor
networks, and provide guidance and direction to such research. The panelists
will also also provide insights to other related questions such as -- What are
the significant QoS metrics in sensor networks? What are the different types of
QoS that can be provided in sensor networks? Are the metrics application
specific? Which network layers are most suited for QoS enhancement?