STATE OF THE ART IN OPPORTUNISTIC SPECTRUM ACCESS MEDIUM ACCESS CONTROL DESIGN Pawelczak, P.; Pollin, S.; So, H.-S.W.; Motamedi, A.; Bahai, A.; Prasad, R.V.; Hekmat, R.; Department of EEMCS, Delft University of Technology, The Netherlands Cognitive Radio Oriented Wireless Networks and Communications, Presenter: Han-Tien Chang Advisor: Professor Dr. Yeali S. Sun

Outline  Introduction  Survey of Opportunistic Spectrum Access MACs  The essential features (functionalities) of OSA MAC  Measurements of Licensee Channel Occupancy  Conclusion  Comments 2

Introduction 3  Opportunistic Spectrum Access (OSA)  A promising new spectrum management approach  Allow co-existence of both licensed and opportunistic users in each spectrum band  Potentially decreasing the spectrum licensing costs for both classes of users Boost the throughput of the networks

Introduction (cont’d) 4  Exclusive: exclusive channel use to each user  Spectrum commons: everybody competes to use the spectrum  Hierarchical: give PU more rights to use the spectrum  Overlay: only one user can use a frequency band in particular time and space  Underlay: SU can transmit in an occupied channel with PU, but it cannot increase the interference to PU over than a threshold Modern spectrum management: classification with the application examples Extensively discuss in this paper [Ref] Przemyslaw Pawelczak ; et. al, “Quality of Service Assessment of Opportunistic Spectrum Access: A Medium Access Control Approach”, IEEE Wireless Communications, Oct. 2008

Introduction (cont’d) 5  In this paper survey,  Catalogue the features about the deign of OSA MAC in the literature  Focus on the distributed channel access OSA MAC protocol The OSA nodes locally decide on, where and how to access the channel  Comparing different design choices

Introduction (cont’d) 6  Many different coexistent scenarios for OSA networks and licensed users  But it’s still unclear what are the most realistic or useful scenarios  Thus, in this paper  performing some measurements of licensed bands get an idea of the expected licensed channel use

Survey of OSA MACs 7 Catalogue the features of 20 OSA MACs  explain the features later

Survey of OSA MACs 8  Bootstrapping  A pre-process during which an OSA node decides which PU channels are allowed for opportunistic spectrum communication Consult such an external entity (spectrum regulator)  Radio Front Ends (RFEs)  The number of RFEs significantly affects the operation of an OSA node Increasing RFEs higher reliability, lower delay (channel switch), but increasing the total cost  Channel Joining (6/20)  Combine the together to increase the throughput  Variable bandwidth

Survey of OSA MACs (cont’d) 9  Multiple Channel Management  Dedicated (Common) Control Channel (DCC) One SU channel is dedicated solely to the transport of control messages Drawback When a PU is active on the control channel, all communication is obstructed. Need extra dedicated RFE  Hopping Control Channel (HCC) All nodes hop between all channels following a predefined hopping pattern HCC doesn’t require a particular single channel to be free from PU activity

Survey of OSA MACs (cont’d) 10  Split Phase Control Channel (SPCC) Time is divided into control and data phases During the control phase 1. All nodes switch their RFEs to the dedicated control channel 2. Decide on the channels to use for the upcoming data transfers Need no extra RFE, but need stronger synchronization  Multiple Rendezvous Control Channel (MRCC) Multiple nodes can exchange control information at the same time, using all available channels Also need a stringent synchronization

Survey of OSA MACs (cont’d) 11  Discuss the performance about these multichannel management mechanism  Offer different primary users’ load in all channel  Observe the average delay and throughput

Survey of OSA MACs (cont’d) 12 Dashed lines: the shorter PU packet size Solid lines: the longer PU packet sizes MRCC and HCC are better  MRCC doesn’t suffer a lot from PU packet size  MRCC randomizes the channel access maximally among all PU channels PU Load vs. average packet delay PU Load vs. average throughput

Survey of OSA MACs (cont’d)  Scanning Process  Since an SU cannot use the channel when a PU is present, it should obtain information about PU activities on each channel Via PU detector or broadcast by a central device  Scanning increases the overhead since nodes cannot transmit when they are scanning  It’s hard to distinguish SU and PU signals, the whole SU network has to be quiet during sensing. 13

Survey of OSA MACs (cont’d)  Quiet period management Let all SU network be quiet during sensing The approach can be done periodically or before each transmission attempt  The more tolerant the PU to interference, the less often sensing should be done.  Scanning performance is measured in terms of the probability of detecting a PU when present, and the probability of falsely detecting a PU.  The measurement of protecting PU’s QoS 14

Survey of OSA MACs (cont’d) 15  Policies  The operation of every OSA network has to be governed by the radio regulator policies. Time based Define time frame during which the OSA node is allowed to transmit on the occupied PU channel Power based Define certain power regimes that the OSA nodes needs to fulfill to be able to use certain PU channels Collision based Define the probability that OSA traffic will collide with the PU packets

Survey of OSA MACs (cont’d) 16  Multi-hop Communications  Depending on the targeted infrastructure  Interaction Between Licensed and Opportunistic Users  The SU may know a priori of the transmission feature from the other party  But the ultimate scenario OSA network shall be deployed into a completely unknown radio environment

Survey of OSA MACs (cont’d) 17  Negotiated Parameter  Between the transmitter and receiver in the secondary users  tune to the best channel Channel Number Queuing parameters: queue size and packet length Channel history: collision prob. and PU occupancy distribution Physical properties of the channel ARQ level, link adaptation strategy  Traffic Estimation  A good prediction of the PU traffic can result in significant performance

Measurement of licensee channel occupancy 18  Objective of this measurement [22]  Get some information on what type of channel occupancy distributions that can expect while observing the PU channels  Measurement setup  Using Rhode&Schwartz ESPI07 spectrum analyzer (SA)  Time: on March , between 11 AM and 8 PM.  Frequency band: between and MHz. Measure the On/Off period [22] P. Pawełczak, S. Pollin, H.-S. W. So, A. Bahai, R. V. Prasad, and R. Hekmat, “Performance analysis of multichannel medium access control algorithms for opportunistic spectrum access,” IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 58, NO. 6, JULY 2009

Measurement of licensee channel occupancy 19  Statistical analysis of the measured data  Using the Kolmogorov-Smirnov test (K-S test) To determine which continuous distribution fits the obtained data best  Result The majority of traces had no identified type of distribution The exponential distribution usually assumed in the analysis of OSA MACs has only a reflection in a handful of the identified activities.

Conclusion 20  Identified different MAC designs proposed so far in the literature  different designs have been classified and assesses in systematic manner  Performance analysis of the multichannel MAC design feature (channel access)  spreading the communication across channels enables to optimally avoid PU activity

Comments  Know the features of the OSA MAC Protocols  With the importance and the functionality  The channel access management mechanism is very important to both the PU and SU QoS  The realistic whitespace usage is not considered here 21