Combating the effects of Hidden Terminals in Multi Channel MAC Protocols Mthulisi Velempini

BACKGROUND Multichannel techniques reduce co-channel interference and collisions They do encourage concurrent transmissions They increase network capacity, flexibility and network reliability The use, access and coordination of multiple channels is a challenge Channel Switching gives rise to Missing Terminals Multichannel approach does not solve HTP

Motivation IEEE MAC protocol is not scalable Anticipated greater demand for QoS in B3G/4G More throughput will be required to improve the efficiency of access networks Effects of HTP can further be reduced or eliminated A move towards total connectivity and interoperability of networks

RELATED WORK Some schemes implement a control and one or more data channels, but do not solve HTP Some employ switchable and fixed transceivers which partitions a network Use of broadcast messages degrade network performance Time based schemes require global synchronization Randomly selecting a home channel increases instances of MRP

Continuation Receiver based approaches fail to calm transmitters in the sender neighbourhood A quiescent channel scheme has also been considered which partitions a network Energy based schemes reduce spatial reuse opportunities

T HE M ODEL In our model, HTP is treated differently from MRP MRP has been misunderstood as Multi- Channel HTP In this paper we investigate a HTP, a MRP will be explored in future In one common channel set up, we investigate whether the effects of HTP are limited to the control channel

Channel Selection Three channels are considered One control and two data channels The control channel is the signalling channel A sender selects a preferred idle channel to be confirmed by the receiver Other idle channels are included in RTS If the channels are busy at the receiver end, a CTS should contain NAV values of the channels Otherwise a RX sends its preferred channel

Dealing with Control Channel HTP Nodes in communication range of TX and RX to contend for channel access after CTS Nodes outside this range will receive erroneous packets Nodes in the RX neighbourhood will not cause interference after hearing an erroneous CTS TX nodes may interfere with CTS reception This is different from Single channel setup where DATA is affected

Packet Collision Figure 1Analysis ofSignaling in Multichannel Packet Channel Switching A B D B A C D F G G H H F E E J I I J K K L L E F E F A B B A D C C Channel 1 Channel 2 Channel3 C Analytical Diagram

ANALYTICAL RESULTS CTS packets are affected in the control channel DATA and ACK packets are also vulnerable in data channels The success of control channel handshake does not protect DATA and ACK packets These can be protected by a hold off scheme in the control channel We call the scheme, the Control IFS (CIFS) Switching delay to be factored into the scheme

Continuation A returning pair can successfully contend for the last used data channel immediately This is true only if the control channel is idle and there are no other nodes contending for the same data channel

Future Work A control and data channels scheme is to be implemented in our future work The CIFS scheme will be implemented in the control channel The model will be validated in a simulated environment Actual durations will be implemented MRP will be tackled in our future work

C ONCLUSION Multi-channel schemes do outperform single channel schemes This does not mean that the designs are good Poorly designed Multi channel schemes can outperform single channel schemes There is need for standardized schemes and performance benchmarks This will facilitate fair comparison of multi channel systems