Wireless Communications Research Overview - PowerPoint by 2ee6rq


									     EE360: Lecture 11 Outline
   Annoucements
       One homework, due May 30
       Remarks on Paper Summaries
   Wireless LAN Design
   Wireless LAN Predictions and Reality
   Wireless LAN Protocols
       “Comparison of 802.11a and HyperLan” Tom Bruns.
   Rate Adaptive Techniques
       “A rate adaptive MAC protocol” Jie Weng
        Paper Summary Remarks
   Should convey the main ideas, techniques, results,
    contributions, and possible extensions of the papers
       Write for an audience that has not read the paper
       Do not write a “review” of the paper
       Analysis should be briefly described

   Should compare and contrast results in all papers, not
    just separately summarize each one
       Give an overview of the general topic these papers address
       Give the “big picture” and how these papers fit in
   Collaborative summaries must be collaborative, not just
    a division of labor
       All collaborators must read and discuss all papers
       The writeup should be collaborative and read/critiqued by all
       Wireless LAN Overview
   Wireless LANs provide high-speed wireless
    access over short distances (10s-100s of feet)

   Typically operate in the “free” bands
     Interference   rejection/avoidance needed

   Geared for low-mobility users
     Slow channel variations
     Infrequent handoff
         Wireless LAN Design
   Physical (link) layer design
   MAC design
   Network architecture
     centralized or ad-hoc

   Interaction with higher layer protocols
       TCP
       QOS mechanisms
              Physical Layer

   Want high speeds and good performance
   Must compensate for ISI and slow fading
   Must compensate for interference
   Most designs use DSSS or OFDM
     Wideband techniques for high speeds
     Both techniques compensate for ISI
     DSSS rejects interference while OFDM
      allocates subchannels to avoid interference
                MAC Design
   CSMA or centralized control (reservations)
     CSMA efficient for short data bursts
     Centralized control (with QOS) needed   for
      media streaming
     Centralized control does not work well for
      multihop networks

   What will be the “killer” application for
    wireless LANs???
        Network Architecture

   Centralized (Star)

   Ad-Hoc

   Implications
     Single hop versus multihop
     Centralized versus distributed        control
     Tradeoffs:
          Infrastructure, performance, flexibility, and range
       Higher Layer Protocols
   TCP
     Congestion   control protocol designed for the
      Internet (transport layer)
     Lost/delayed/corrupted packets trigger
      congestion control
     Doesn’t work well for wireless links

   QOS Mechanisms
     ATM guarantees QOS (delay, rate, BER, etc.)
     Needed for media streaming applications
     Hard to provide QOS on wireless channels
          Flavors: Guaranteed, predictive, best effort
               Performance Gap with
                   Wired Systems
                                                                         100 M               ATM

                      10,000                                      FDDI
                                                                                    wired- wireless
                                                       Ethernet                     bit-rate "gap"
                        1000         User
                                     (kbps)                                                       2nd gen
         DATA RATES

                         100                                                                      WLAN
                                                                           1st gen
                                            Polling                        WLAN



                                   1970                   1980                   1990                       2000

802.11a promises about 5 times higher data rates than 802.11b
 - Still 2-10 times slower than wired LANs, for a single user!
     Wireless LAN Predictions
   “802.11 is on fire” Benchmark Capital, 12/00.

   “Although its been the pot of gold just around the corner for the last
    five years, we still don’t have a hughly profitable business case for
    wireless data” Dr. David Roddy, Chief Telecom Economist, Deloitte & Touche, 4/98
   “Up to now, the wireless data party has been more flop than fete,
    with precious few suscribers showing even a hint of interest in using
    unwired services for anything more than voice calls. This year may
    finally see wireless data make its long-expected splash” Meg McGinity,
    Wireless Editor, 4/98

   “Wireless LANs finally appear poised for the growth and popularity
    long expected of them. The wireless LAN market in the U.S. will
    enjoy a sixfold expansion by the year 2000, reaching more than 1
    billion in revenue” Morris Edwards, Communications News, May, 1997.
                    Wireless LAN Reality

                                  Internet users
                                                   Cellular users
            80       USA market
            70                                      pager users

            40                                     laptop users
            20                             dedicated wireless
            10                                  data subs
             1995                                                   2000
            Closing Thoughts
   802.11a promises 50 Mbps speeds (shared
    rate) at relatively low cost
     QOS   support not really available
 Is this the breakthrough that will finally
  realize the wireless LAN promise?
 What killer applications will need to be
  supported on wireless LANs?
 What technical breakthroughs are still

To top