MultiChannel Wireless Networks Theory to Practice Nitin Vaidya

Multi-Channel Wireless Networks: Theory to Practice Nitin Vaidya Electrical and Computer Engineering University of Illinois at Urbana-Champaign Sept. 8. 2008 1

Multi-Channel Wireless Networks Acknowledgements Ph. D M. S. g Jungmin So (2006) g Priya Ravichandran (2003) g Pradeep Kyasanur (2006) g Chandrakanth Chereddi (2006) g Vartika Bhandari (2008) g Rishi Bhardwaj (2007) g Thomas Shen (2008) g Vijay Raman Post-docs g Wonyong Yoon g Cheolgi Kim Funded in part by: NSF, ARO, Motorola, Boeing 2

Preliminaries … 3

Wireless Networks g Wireless paradigms: Single hop versus Multi-hop g Multi-hop networks: Mesh networks, ad hoc networks, sensor networks 4

What Makes Wireless Networks Interesting? Significant path loss - Signal deteriorates over space + Spatial re-use feasible B A power g S distance 5 5

What Makes Wireless Networks Interesting? Interference management non-trivial B A power g I C D S distance 6

What Makes Wireless Networks Interesting? Many forms of diversity • Time • Route • Antenna • Path • Channel 7

What Makes Wireless Networks Interesting? Time diversity C D gain Time 8

What Makes Wireless Networks Interesting? Route diversity infrastructure AP 1 Access point B A F X C AP 2 D E Z 9

What Makes Wireless Networks Interesting? Antenna diversity D C A B 10 Sidelobes not shown

What Makes Wireless Networks Interesting? Path diversity 11

What Makes Wireless Networks Interesting? Channel diversity High interference Low gain B A High gain B A D C Low interference 12

Wireless Capacity g Wireless capacity limited g In dense environments, performance suffers g How to improve performance ? 13

Improving Wireless Performance g Exploit physical resources, diversity g Exploiting diversity requires appropriate protocols 14

This Talk Utilizing multiple channels in multi-hop wireless 15

Multi-Channel Environments Available spectrum Spectrum divided into channels 1 2 3 4 … c 16

Multiple Channels 3 channels 8 channels 4 channels 26 MHz 100 MHz 200 MHz 150 MHz 915 MHz 2. 45 GHz 5. 25 GHz 5. 8 GHz IEEE 802. 11 in ISM Band 17

Shared Access : Time & Spectrum D B C A One Channel Spectrum A B Time A Two Channels C A B C Time C 18

Why Divide Spectrum into Channels ? g Manageability: • Different networks on different channels avoids interference between networks g Contention mitigation: • Fewer nodes on a channel reduces channel contention 19

Why Divide Spectrum into Channels ? g g Lower transmission rate per channel • Slower hardware (simpler, cheaper) Reducing impact of bandwidth-independent overhead fixed time data size/rate 20

Outline capacity D Theory to Practice Net-X testbed E Fixed F B A Switchable C Capacity bounds channels Insights on protocol design OS improvements Software architecture User Applications Multi-channel protocol IP Stack ARP Channel Abstraction Module Linux box CSL Interface Device Driver 21

Interfaces & Channels g An interface can only use one channel at a time Channel 1 W c. W Channel c g Switching between channels may incur delay 22

Channel Switching g Unconstrained : An interface can tune to any available channel g Constrained : Restricted channel switching 23

This Talk g Assume unconstrained switching g Constrained switching results elsewhere 24

Multiple Interfaces g g Reducing hardware cost allows for multiple interfaces m interfaces per node 1 m 25

Practical Scenario g m<c A host can only be on subset of channels 1 1 m m m+1 c c–m unused channels at each node 26

Multi-Channel Mesh g How to best utilize multiple channels in a mesh network with limited hardware ? ? 27

Need for New Protocols m<c c = 4 channels m = 2 interfaces 1, 2 A B D 1, 2 C 1, 2 Some channels not used 1, 3 1, 2 3, 4 A B C D 2, 4 Network poorly connected 28

Multi-Channel Networks Many Inter-Dependent Issues g How to choose a channel for a transmission? B A g g How to schedule transmissions? How to measure “channel quality” C - gain, load g How to select routes ? 29

Outline capacity D Theory to Practice Net-X testbed E Fixed F B A Switchable C Capacity bounds channels Insights on protocol design OS improvements Software architecture User Applications Multi-channel protocol IP Stack ARP Channel Abstraction Module Linux box CSL Interface Device Driver 30

Capacity Analysis g g How does capacity improve with more channels ? How many interfaces necessary to efficiently utilize c channels ? 31

Network Model 32

Network Model [Gupta-Kumar] g Random source-destination pairs among randomly positioned n node in unit area, with n ∞ 33

Capacity = ? g g l = minimum flow throughput Capacity = n l 34

Capacity Constraints g g Capacity constrained by available spectrum bandwidth Other factors further constrain wireless network capacity … 35

Connectivity Constraint [Gupta-Kumar] g Need routes between source-destination pairs Places a lower bound on transmit power A D Not connected A D Connected 36

Interference Constraint [Gupta-Kumar] g g Interference among simultaneous transmissions Limits spatial reuse >r A r C D B 37

Capacity [Gupta-Kumar] g c=m capacity a 1 1 m=c c=m Capacity scales linearly with channels 38

Capacity g What if fewer interfaces ? 1 1 m m m+1 c 39

Interface Constraint g Throughput is limited by number of interfaces in a neighborhood N nodes in the “neighborhood” total throughput ≤ N *m * W Interfaces as a resource in addition to spectrum, time and space 40

Mutlti-Channel Capacity Order of Channels (c/m) 41

Capacity with n ∞ Are these results relevant ? g Yield insights on design of good routing and scheduling protocols g Insights relevant in smaller networks too 42

Outline capacity D Theory to Practice Net-X testbed E Fixed F B A Switchable C Capacity bounds channels Insights on protocol design OS improvements Software architecture User Applications Multi-channel protocol IP Stack ARP Channel Abstraction Module Linux box CSL Interface Device Driver 43

Insights from Analysis (1) Channel Assignment g Need to balance load on channels g Local coordination in channel assignment helpful 44

Insights from Analysis (2) g g Static channel allocation not optimal performance in general Must dynamically switch channels Channel 1 B A 2 C D 45

Insights from Analysis (3) g Optimal transmission range function of number of channels Intuition: # of interfering nodes ≈ # of channels 46

Insights from Analysis (4) g Routes must be distributed within a neighborhood D D F A B E F B A E C C m=1 interface c=1, 2 channels 47

Insights from Analysis (5) g Channel switching delay potentially detrimental, but may be hidden with icareful scheduling – create idle time on interfaces between channel switches iadditional interfaces 48

Protocol Design: Timescale Separation Upper layers g Routing: Longer timescales i(Optional) Multi-channel aware route selection g Interface management: Shorter timescales i. Dynamic channel assignment i. Interface switching Transport Network 802. 11 Link Physical Layer 49

Channel Management g g Two interfaces much better than one Hybrid channel assignment: Static + Dynamic Fixed (ch 1) Fixed (ch 2) Fixed (ch 3) A B C Switchable 2 1 Switchable 3 2 Channel assignment locally balanced Switchable 50

Selecting Channel Diverse Routes 1 A D 3 4 2 3 B 4 4 C 4 E F 4 2 A needs route to C Route A-B-C better More channel diverse 51

Impact of Switching Cost on Route Selection 1 3 4 A B C 2 3 D needs route to F D 2 Route A-B-C in use E 4 4 F 2 Route D-E-F better 2 Prefer routes that do not require frequent switching 52

CBR – Random topology (50 nodes, 50 flows, 500 m x 500 m area) (m, c) 53

Outline capacity D Theory to Practice Net-X testbed E Fixed F B A Switchable C Capacity bounds channels Insights on protocol design OS improvements Software architecture User Applications Multi-channel protocol IP Stack ARP Channel Abstraction Module Linux box CSL Interface Device Driver 54

Net-X Testbed g g Soekris 4521 g g Linux 2. 4 Two 802. 11 a radios per mesh node (m = 2) Legacy clients with 1 radio c = 5 channels Net-X source available 55

Phy-Aware Support g Additional mechanisms needed to choose channels based on destination Ch. 2 B A Ch. 1 C Next hop not equivalent to a wireless interface id g g Phy-aware forwarding not supported traditionally In general, need a “constraint” specification for desired channel(s), antenna beamform, power/rate, … to be used for the next hop 56

Phy-Aware Support g Multi-channel (phy-aware) broadcast Ch. 2 D g Channel switching from user space has high latency: frequent switching from user space undesirable Ch. 3 A B Ch. 1 C 57

New Kernel Support g Interface management needs to be hidden from “data path” – Buffering packets for different channels – Scheduling interface switching Ch. 2 Ch. 1 Packet to B buffer packet Interface switches to channel 1 Packet to C arrives 58

Net-X Architecture User Applications Multi-Channel Routing, Channel Assignment g IP Stack ARP Interface and Channel Abstraction Layer Interface Device Driver Abstraction layer simplifies use of multiple interfaces Implemented by extending Linux “bonding driver” 59

capacity D E Fixed F B A Switchable C Capacity bounds channels Insights on protocol design Wrap-up Net-X testbed OS improvements Software architecture User Applications Multi-channel protocol IP Stack ARP Channel Abstraction Module Linux box CSL Interface Device Driver 60

Current Status g ~ 25 node network operational g Protocol improvements … ongoing process g Further results for • Scheduling in multi-channel networks • Constrained channel assignment • Cross-channel interference 61

Summary g g g Significant performance benefits using many channels despite limited hardware Insights from analysis useful in protocol design Conversely, implementation experience helps formulate new theoretical problems Important to complete the loop from theory to practice 62

Research Opportunities g g Significant effort in protocol design needed to exploit available physical resources Other opportunities: • MIMO (multi-antenna) • Cooperative relaying • Dense wireless infrastructure 63

Thanks! www. crhc. uiuc. edu/wireless 64

Thanks! www. crhc. uiuc. edu/wireless 65

Thanks! www. crhc. uiuc. edu/wireless 66

Thanks! www. crhc. uiuc. edu/wireless 67

Scenario 1 g 1 m=c One interface per channel 1 1 m=c c=m 1 Common case With sufficient hardware 68

Constrained Switchability g g An interface may be constrained to use only a subset of channels Motivation: i. Hardware limitations (“untuned radio” [petrovic] ) i. Hardware heterogeneity (802. 11 b/g versus 802. 11 a/b/g) i. Policy issues (cognitive radios) 69

Impact of Constrained Switching (4, 5) (4, 6) (5, 6) (1, 2) (1, 7) (4, 5) A (6, 7) E B D (3, 4) C (2, 4) (7, 8) (1, 3) Reduced Connectivity Detour Routing (6, 7) (2, 5) 70

Impact of Constrained Switching 1 relay on channel b Z 3 relays on channel a X, Y, Z a, b S a X a Y Z a, b D Coupling between channel selection & relay choice 71

Impact of Constrained Switching a, c G a c c, f P c a, b X c, d Y b b, d H d d d, f Q 6 channels: a, b, c, d, e, f Bottleneck formed at Y 72

Destination Bottleneck Constraint g A node may be destination of multiple flows g Node throughput shared by all the incident flows f incoming flows D Node throughput T ≤ m. W Per-flow throughput T / f 73

Multi-Channel Network Capacity Interference and interface bottleneck Interface and destination bottlenecks Connectivity and interference Ratio c/m 74

Routing Approach g Legacy routing protocols can be operated over our interface management layer i. Does yield significant benefits from multiple channel i. Does not consider cost of channel switching g An alternative i. Develop a channel-aware metric (aware of channel diversity and switching costs) 75