Panel: Towards an All IP Next-Gen Mobile Networks

• Sessions as fast as 30 seconds on mobil
  • On the move info must be short and easy and fast to fetch
• Remember
  • wireline bandwidth is per-user
  • wireless bandwidth is per base station (density matters)
• Mark/Willingness-To-Pay
• Market & Technology Competition
• Technical Challenges
• QualComm
  • Moving towards a common core network
  • Note work on WiFi+Cellular integration
  • Highspeed short-range device to device connectivity!
    • Not for same reasons as we want, but still good to know that it's an

interest!

• Verizon Wireless
  • No matter what, next gen is higher bandwidth and lower latency
  • Frustration: speed of technological progress
  • Not all bad from a business perspective
    • Incremental improvements are better from an economic standpoint
    • Still seeing significant customer-base growth
    • New growth is coming more from data
    • Data service in Verizon is paying for itself!
  • Not all bad from a customer perspective
    • The business success at one generation funds the deployment of the next
  • Many problems solved by device agility and network agility
    • Multi-client multi-band multi-technology
  • Major challenge: battery life
• T-Mobile
  • Fundamental question: why do we *need* next generation?
  • Users are not an unlimited source of revenue!
  • In some ways, 3G is not fulling expectations? (What about Verizon's

comments?)

• Call to researchers: let's build a platform for innovations

• Cisco
  • No point in having fast radios is the base stations are a bottleneck!
  • There is room for technical and business model innovations for

differentiated service pricing

• Intel
  • Hey, what about WiMax?
  • Key necessary: flat-IP core network
  • Standard convergence is a pipe dream, but technological convergence is a

possibility

• Ericsson
  • There is a lot to learn from fixed side as far as making things easy-to-use

and well-integrated

• Last-mile is shortening
• Challenge: QoS and Security
• Challenge: Open interfaces for access networks (for research)

• Cingular
  • 11.6% users use data
  • Feels the voice is still the driver
  • Blah blah marketing blah
• AT&T
  • Data is going from everyone to everyone now
  • Users will invent ways to use capability
  • Intelligence will be forced to network edges
  • Users will decide now
• Sprint
  • Sprint's strategy is to achieve "Internet Everywhere"
    • Visually-rich content and bandwidth-intensive apps
  • Chosen WiMax as next-gen technology!
    • Performance and Cost
    • Time to Market
    • Business Model
  • Samsung, Moto, Intel are partners in WiMax decision
• Followups
  • We can do a lot with applications to more efficiently use wireless bandwidth
  • Politics are a holdup at lower frequency, despite tech advances
  • Three laws of cognitive radio (like robots)

Session - New Topics

Fast and reliable estimation Schemes in RFID Systems

• Counting - Find number of tags in the system in the shortest time possible
• Model
  • Large number of wireless end devices: RFID i.e.
  • Probing device only intrested in counting
  • Devices capable of picking random number
  • Use time-slotted ALOHA
  • Time sync provided by reader
  • Reader can sense idle slot/collision

Low-cost Communication for Rural Internet Kiosks Using Mechanical Backhaul

• Bridging the digital divide in developing regions
  • Farmers: market prices
  • Tele-medicine
  • E-government
  • Grassroots media
• Kiosks
  • Dial up, low and slow and flaky
  • VSAT: Expensive
  • Cellular: Low penetration
  • WiFi/WiMax: Expensive, planning needed
  • Short range WiFi: equipment security
• Questions
  • What kind of applications?

Coverage and Connectivity in Three-Dimensional Networks

• Common assumption: 2D plane.
  • Not exactly accurate, but usually reasonable assumption for terrestrial

nets.

• What about underwater/space networks?

• 2D networks: Hexagonal cells.
  • What is shape for 3D?
• Assumptions
  • Deployment anywhere in space
  • Omnidirectional
  • Sensing radius is smaller than network (ignore boundary effect
• Goals
  • full coverage with minimal nodes
• Volumetric quotient
  • Ratio of volume to circumsphere
  • Find space-filling structure with highest volumetric quotient
• Questions
  • This is done already, for packing and covering (ouch)

Hot Topic: Physical-Layer Network Coding

• Turn interface into an advantage
  • Interference is good
• Effect network coding at EM-wave level
• Combinations of two bits at EM-wave level
• 3-2 timeslot mapping
• Synchronization issues
  • Phase
    • No carrier-phase sync = SNR penalty <1dB
  • Symbol alignment
    • Uniformly distributed misalignment = 1.5dB penalty
• Paper-generating: just plug in new constraints :-)