Networked Embedded Systems: PHY and MAC

Transcription

1 Networked Embedded Systems: PHY and MAC Prof. António Grilo Instituto Superior Técnico (IST), Lisboa, Portugal

2 Overview Principal options and difficulties Contention-based protocols Schedule-based protocols IEEE Other protocols 2

3 Wireless Sensor and Actuator Networks Node Requirements: Wireless communications Autonomous = Rely on batteries and/or energy harvesting Low cost = Small batteries, slow proc., low mem., cheap oscillator, etc. Comms Requirements: Unlicensed bands As usual: high throughput, low delay, low error rates, Additionally: energy-efficient! = Multihop communications 3

4 ISM Bands Some typical ISM bands Frequency Comment 13,553-13,567 MHz 26,957 27,283 MHz Particularly interesting: ISM bands ( Industrial, scientific, medicine ) license-free operation 40,66 40,70 MHz MHz Europe 868 MHz Europe 2,4 2,5 GHz WLAN/WPAN 5,725 5,875 GHz WLAN 24 24,25 GHz 4

5 Free Space Radio Propagation: Friis Model Effect of attenuation: received signal strength is a function of the distance d between sender and transmitter Captured by Friis free-space equation Describes signal strength at distance d relative to some reference distance d 0 < d for which strength is known d 0 is far-field distance, depends on antenna technology 5

6 Radio Propagation: Log Distance Model Reference Distance: e.g., d 0 = 1 m PL 0 = Friis (d 0 ) Lognormal shadowing: X g = N( ) Pathloss Exponent: In the Friis model: = 2 and X g =0 Indoor propagation: typically > 2 6

7 Signal Strength and Bit Errors Extracting symbols out of a distorted/corrupted wave form is fraught with errors: Depends essentially on strength of the received signal compared to the corruption Captured by signal to noise and interference ratio (SINR) SINR = 10 log P I N + SINR allows to compute bit error rate (BER) for a given modulation: Also depends on data rate (nº bits/symbol) of modulation e.g., for simple DPSK, data rate corresponding to bandwidth: BER E /N = 0.5 e SINR = E R N B 7

8 Difficulties with WSN Communications Medium access in wireless networks is difficult mainly because of: Impossible (or very difficult) to send and receive at the same time No CSMA/CD Interference situation at receiver is what counts for transmission success, but can be very different from what sender can observe High error rates compound the issues Low Ptx and sporadic transmission means that the relevance of Prx and Pidle increases in overall consumption 8

9 Requirements for energy-efficient MAC Protocols Minimize energy wasting factors: Collisions wasted effort when two packets collide Overhearing waste effort in receiving a packet destined for another node Idle listening sitting idly and trying to receive when nobody is sending Protocol overhead Always nice: Low complexity solution 9

10 Centralized vs Distributed MAC Idea: Have a central station control when a node may access the medium Example: Polling, centralized computation of TDMA schedules Advantage: Simple, quite efficient (e.g., no collisions), burdens the central station Not directly feasible for non-trivial wireless network sizes But: Can be quite useful when network is somehow divided into smaller groups Clusters, in each cluster medium access can be controlled centrally compare Bluetooth piconets, for example Usually, distributed medium access is considered 10

11 Schedule- vs. contention-based MACs Schedule-based MAC A schedule exists, regulating which participant may use which resource at which time (TDMA component) Typical resource: frequency band in a given physical space (with a given code, CDMA) Schedule can be fixed or computed on demand Usually: mixed difference fixed/on demand is one of time scales Contention-based protocols Risk of colliding packets is deliberately taken Hope: coordination overhead can be saved, resulting in overall improved efficiency Mechanisms to handle/reduce probability/impact of collisions required Usually, randomization used somehow E.g. Aloha, CSMA, etc. Hybrid protocols 11

12 Schedule- vs. contention-based MACs II Scheduled Contention-based No collisions No overhearing Low idle listening Well-defined sleep times Simple Scalable with network size Adapts well to topology changes Need synchronization Need signalling to establish schedule Low signalling overhead No synchronization Automatically adapts to traffic load Cheap oscillators clock drift clock skew Scalability problems Not adaptive to topology changes (but ) Prone to collisions Idle listening Overhearing Do not easily adapt to traffic load 12

13 Overview Principal options and difficulties Contention-based protocols MACA PAMAS S-MAC, T-MAC Preamble sampling, B-MAC Schedule-based protocols IEEE Other protocols 13

14 Contention-based MAC Basic ideas for a distributed MAC ALOHA no good in most cases Listen before talk (Carrier Sense Multiple Access, CSMA) better, but suffers from sender not knowing what is going on at receiver, might destroy packets despite first listening Receiver additionally needs some possibility to inform possible senders in its vicinity about impending transmission (to shut them up for this duration) Hidden terminal scenario: A B C D Also: recall exposed terminal scenario 14

15 Main options to shut up senders Receiver informs potential interferers while a reception is on-going By sending out a signal indicating just that Problem: Cannot use same channel on which actual reception takes place Use separate channel for signaling Busy tone protocol Receiver informs potential interferers before a reception is on-going Can use same channel Receiver itself needs to be informed, by sender, about impending transmission Potential interferers need to be aware of such information, need to store it 15

16 Receiver informs interferers before Tx MACA Sender B asks receiver C whether C is able to receive a transmission Request to Send (RTS) Receiver C agrees, sends out a Clear to Send (CTS) Potential interferers overhear either RTS or CTS and know about impending transmission and for how long it will last Store this information in a Network Allocation Vector B sends, C acks MACA protocol (used e.g. in IEEE ) NAV indicates busy medium A B C D RTS Data CTS Ack NAV indicates busy medium 16

17 RTS/CTS RTS/CTS ameliorate, but do not solve hidden/exposed terminal problems Example problem cases (but there are more):

18 MACA Problem: Idle Listening Need to continuously sense carrier for RTS, CTS or DATA packets If traffic is exporadic, idle listening may spend more energy than Tx/Rx Simple sleeping will break the protocol Some form of scheduling is usually necessary! IEEE solution: ATIM windows & sleeping Basic idea: Nodes that have data buffered for receivers send traffic indicators at pre-arranged points in time Receivers need to wake up at these points, and can only sleep when sure that there is no traffic for them Requires tight synchronization (the standard does not specify how this is to be achieved) In a multihop network, unsynchronized ATIM windows may lead to network partition. 18

19 Power Aware Multiaccess with Signaling PAMAS Idea: combine busy tone with RTS/CTS Results in detailed overhearing avoidance, does not address idle listening Uses separate data and control channels Procedure Node A transmits RTS on control channel, does not sense channel Node B receives RTS, sends CTS on control channel if it can receive and does not know about ongoing transmissions B sends busy tone as it starts to receive data Control channel Data channel RTS A B CTS B A Busy tone sent by B Data A B Time 19

20 PAMAS Already ongoing transmission Suppose a node C in vicinity of A is already receiving a packet when A initiates RTS Procedure B C? A sends RTS to B A C sending busy tone (as it is receiving data) CTS and busy tone collide, A receives no CTS, does not send data Control channel Data channel RTS A B Busy tone by C CTS B A No data! Time Similarly: Ongoing transmission near B destroys RTS by busy tone 20

21 Sensor MAC S-MAC MACA s idle listening is particularly unsuitable if average data rate is low Most of the time, nothing happens Idea: Switch nodes off, ensure that neighboring nodes turn on simultaneously to allow packet exchange (rendez-vous) Only in these active periods, packet exchanges happen Need to also exchange wakeup schedule between neighbors When awake, essentially perform RTS/CTS Use SYNCH, RTS, CTS phases 21

22 S-MAC Synchronized Islands Nodes try to pick up schedule synchronization from neighboring nodes If no neighbor found, nodes pick some schedule to start with If additional nodes join, some node might learn about two different schedules from different nodes Synchronized islands To bridge this gap, it has to follow both schemes A A A A A B E B B B B E E E E E E A C C C C C D D D D Time 22

23 Timeout-MAC (T-MAC) In S-MAC, active period is of constant length What if no traffic actually happens? Nodes stay awake needlessly long Idea: Prematurely go back to sleep mode when no traffic has happened for a certain time (=timeout) T-MAC Adaptive duty cycle! One ensuing problem: Early sleeping C wants to send to D, but is hindered by transmission A B Two solutions exist A B C D CTS May not send Timeout, go back to sleep as nothing happened 23

24 D-MAC Hybrid TDMA/CSMA: tighter synchronization Staggered Wakeup Schedule (< latency) CSMA/CA within slot (but < contention due to separation) Extra slot request: more data flag Early Sleeping problem solved with data prediction Focused on for n 1 comms 24

25 Preamble Sampling So far: Periodic sleeping supported by some means to synchronize wake up of nodes to ensure rendez-vous between sender and receiver Alternative option: Don t try to explicitly synchronize nodes Have receiver sleep and only periodically sample the channel Use long preambles to ensure that receiver stays awake to catch actual packet Example: B-MAC, WiseMAC Start transmission: Long preamble Actual packet Check channel Check channel Check channel Check channel Stay awake! 25

26 B-MAC Clear Channel Assessment Adapts to noise floor by sampling channel when it is assumed to be free Samples are exponentially averaged, result used in gain control For actual assessment when sending a packet, look at five channel samples channel is free if even a single one of them is significantly below noise Random backoff if channel is found busy Optional: Immediate link layer acknowledgements for received packets 26

27 B-MAC II Low Power Listening (= preamble sampling) Uses the clear channel assessment techniques to decide whether there is a packet arriving when node wakes up Timeout puts node back to sleep if no packet arrived B-MAC does not have Synchronization RTS/CTS (may be implemented on top) Results in simpler, leaner implementation Clean and simple interface Currently: Often considered as the default WSN MAC protocol (e.g., supported in TinyOS and Nano-RK) 27

28 SCP-MAC Based on S-MAC and LPL In fact LPL with synchronous channel polling Requires tighter synchronization Short wake-up tone intead of long preamble Two contention windows (< collisions) TX tone Node 1 Node 2 wakeup wakeup sleep sleep wakeup TX RX sleep sleep wakeup wakeup sleep wakeup wakeup time time 28

29 Overview Principal options and difficulties Contention-based protocols Schedule-based protocols WiDom TRAMA IEEE Other protocols 29

30 WiDom Prioritized Collision-Free MAC Protocol for Wireless Sensor Networks Objective: to provide upper bounds on communication delays; also useful to perform efficient collaborative distributed computing (wireless/wired sensor networks and cyber-physical systems). Outline: Prioritized Medium Access Control (PrioMAC) Grants medium access to the computer node with the highest priority Originally created for wired networks, e.g Controller Area Networks (CAN) Applied to the wireless domain WiDOM Bit Dominant Recessive Medium Status Node 1 priority Node 2 priority Node 3 priority Node 3 is the only node that finishes the arbitration without losing 30

31 TRAMA Nodes are synchronized Time divided into cycles, divided into Random access periods Scheduled access periods Nodes exchange neighborhood information Learning about their two-hop neighborhood Using neighborhood exchange protocol: In random access period, send small, incremental neighborhood update information in randomly selected time slots Nodes exchange schedules Using schedule exchange protocol Similar to neighborhood exchange Schedules are bitmaps: for each slot, num bits = #1-hop neighbors 31

32 TRAMA Adaptive Election Given: Each node knows its two-hop neighborhood and their current schedules How to decide which slot (in scheduled access period) a node can use? Use node identifier x and globally known hash function h For time slot t, compute priority p = h (x t) Compute this priority for next k time slots for node itself and all two-hop neighbors Node uses those time slots for which it has the highest priority Priorities of node A and its two neighbors B & C t = 0 t = 1 t = 2 t=3 t = 4 t = 5 A B C

33 TRAMA Possible Conflicts When does a node have to receive? Easy case: one-hop neighbor has won a time slot and announced a packet for it But complications exist compare example What does B believe? A thinks it can send B knows that D has higher priority in its 2-hop neighborhood! Rules for resolving such conflicts are part of TRAMA 33

34 Overview Principal options and difficulties Contention-based protocols Schedule-based protocols IEEE Other protocols 34

35 IEEE IEEE standard for low-rate WPAN applications Goals: low-to-medium bit rates, moderate delays without too stringent guarantee requirements, low energy consumption Physical layer 20 kbps over MHz 40 kbps over MHz 250 kbps over GHz MAC protocol Single channel at any one time Combines contention-based and schedule-based schemes Asymmetric: nodes can assume different roles 35

36 IEEE PHY 2.4 GHz PHY 250 kb/s (4 bits/symbol, 62.5 kbaud) Data modulation is 16-ary orthogonal modulation 16 symbols are orthogonal set of 32-chip PN codes Chip modulation is O-QPSK at 2.0 Mchips/s 868MHz/915MHz PHY Symbol Rate 868 MHz Band: 20 kb/s (1 bit/symbol, 20 kbaud) 915 MHz Band: 40 kb/s (1 bit/symbol, 40 kbaud) Data modulation is BPSK with differential encoding Spreading code is a 15-chip m-sequence Chip modulation is BPSK at 868 MHz Band: 300 kchips/s 915 MHz Band: 600 kchips/s 36

37 IEEE PHY Frame PHY Packet Fields Preamble (32 bits) synchronization Start of Packet Delimiter (8 bits) PHY Header (8 bits) PSDU length PSDU (0 to 1016 bits) Data field Preamble Start of Packet Delimiter PHY Header PHY Service Data Unit (PSDU) 6 Octets Octets 37

38 IEEE MAC Functions Employs 64-bit IEEE & 16-bit short addresses Ultimate network size can reach 2 64 nodes (more than we ll probably need ) Using local addressing, simple networks of more than 65,000 (2^16) nodes can be configured, with reduced address overhead Three devices specified Network Coordinator Full Function Device (FFD) Reduced Function Device (RFD) Simple frame structure Reliable delivery of data Association/disassociation AES-128 security CSMA-CA channel access Optional superframe structure with beacons Optional GTS mechanism 38

39 IEEE MAC Topologies Star Peer-to-peer (Mesh) Cluster Tree PAN coordinator Full Function Device Reduced Function Device 39

40 IEEE General Frame Structure Payload MAC Layer MAC Header (MHR) MAC Service Data Unit (MSDU) MAC Footer (MFR) PHY Layer Synch. Header (SHR) PHY Header (PHR) MAC Protocol Data Unit (MPDU) PHY Service Data Unit (PSDU) 4 Types of MAC Frames: Data Frame Acknowledgment Frame MAC Command Frame Beacon Frame 40

41 IEEE Data Frame One of two most basic and important structures in Provides up to 104 byte data payload capacity Data sequence numbering to ensure that packets are tracked Robust structure improves reception in difficult conditions Frame Check Sequence (FCS) validates error-free data 41

42 IEEE Acknowledgement Frame The other most important structure for 15.4 Provides active feedback from receiver to sender that packet was received without error Short packet that takes advantage of standards-specified quiet time immediately after data packet transmission 42

43 IEEE Command Frame Mechanism for remote control/configuration of client nodes Allows a centralized network manager to configure individual clients no matter how large the network 43

44 IEEE Beacon Frame Beacons add a new level of functionality to a network Client devices can wake up only when a beacon is to be broadcast, listen for their address, and if not heard, return to sleep Beacons are important for mesh and cluster tree networks to keep all of the nodes synchronized without requiring nodes to consume precious battery energy listening for long periods of time 44

45 Frame Control field 45

46 IEEE MAC overview Star networks: devices are associated with coordinators Forming a PAN, identified by a PAN identifier Coordinator Bookkeeping of devices, address assignment, generate beacons Talks to devices and peer coordinators Beacon-mode superframe structure GTS assigned to devices upon request 46

47 IEEE Beacon Mode ( slotted CSMA/CA) Nodes synchronized with Coordinator Coordinator Node Coordinator Node Beacon Beacon Data frame Acknowledgement (opcional) Data Request Acknowledgement Data frame Acknowledgement Data to Coordinator Data from Coordinator 47

48 IEEE Non-Beacon Mode (CSMA/CA) Coordinator always active, Node with low duty cycle Coordinator Node Coordinator Node Data frame Data Request Acknowledgement Acknowledgement (opcional) Data frame Acknowledgement Data to Coordinator Data from Coordinator 48

49 IEEE Peer-to-Peer Nodes synchronized with each other Node 1 Node 2 Data frame Acknowledgement 49

50 Overview Principal options and difficulties Contention-based protocols Schedule-based protocols MAC protocols supported by Nano-RK IEEE Other protocols 50

51 Low-Energy Adaptive Clustering Hierarchy (LEACH) Given: dense network of nodes, reporting to a central sink, each node can reach sink directly Idea: Group nodes into clusters, controlled by clusterhead Setup phase: probabilistic clusterhead election About 5% of nodes become clusterhead (depends on scenario) Role of clusterhead is rotated to share the burden Clusterheads advertise themselves, ordinary nodes join CH with strongest signal Clusterheads organize CDMA code for all member transmissions TDMA schedule to be used within a cluster In steady state operation CHs collect & aggregate data from all cluster members Report aggregated data to sink using CDMA (long-range Tx) 51

52 LEACH rounds 52

53 SMACS Given: many radio channels, superframes of known length (not necessarily in phase, but still time synchronization required!) Goal: set up directional links between neighboring nodes Link: radio channel + time slot at both sender and receiver Free of collisions at receiver Channel picked randomly, slot is searched greedily until a collision-free slot is found Receivers sleep and only wake up in their assigned time slots, once per superframe In effect: a local construction of a schedule 53

54 SMACS Links Setup Case 1: Node X, Y both so far unconnected Node X sends invitation message Node Y answers, telling X that is unconnected to any other node Node X tells Y to pick slot/frequency for the link Node Y sends back the link specification Case 2: X has some neighbors, Y not Node X will construct link specification and instruct Y to use it (since Y is unattached) Case 3: X no neighbors, Y has some Y picks link specification Case 4: both nodes already have links Nodes exchange their schedules and pick free slots/frequencies in mutual agreement Message exchanges protected by randomized backoff 54

55 LEMMA Interference range is usually greater than transmission range 2-hop neighborhood assumption is nor accurate Propagation effects can be complex and hard to predict Slot allocation should be based on trial and error rather than a fixed heuristic. TDMA frame period Slot 0 TS i Parent node Node i (child) Node j (child) RTA 0 (broadcast, [i, TS i], [j, TS j]) CTA(parent, [i, TS i ]) CTA(parent, [j, TS j])... RTA (i, [i, TS i]) CTA(parent, [i, TS i ]) ACW1 RTA (i, [i, TS i]) CTA(parent, [i, TS i ]) Sink at node 0 Sink at node 7... ACW2 3 0 t TS j RTA (j, [j, TS j ]) Collision Detected 1 2 ACW1 4 5 Timeout: A new attempt shall be made in slot TS j ACW2 55

56 Wakeup radio MAC protocols Micro-power wake-up radio + normal radio Simplest scheme: Send a wakeup burst, waking up all neighbors Significant overhearing Possible option: First send a short filter packet that includes the actual destination address to allow nodes to power off quickly Not quite so simple scheme: Send a wakeup burst including the receiver address Wakeup radio needs to support this option Additionally: Send information about a (randomly chosen) data channel, CDMA code, in the wakeup burst Various variations on these schemes in the literature, various further problems One problem: 2-hop neighborhood on wakeup channel might be different from 2-hop neighborhood on data channel Not trivial to guarantee unique addresses on both channels 56

57 RT-Link TDMA slots + slotted Aloha slots (new nodes + mobile nodes) External out-of-band synchronization with special HW (!!!) AM transmitter uses building power grid as the tx antenna. Firefly mote includes a separate AM antenna. Centralized slot assignment at gateway node (does it scale?). Neighborhood map spanning tree (Dijkstra) coloring (thr vs delay)

58 Summary Many different ideas exist for medium access control in MANET/WSN Comparing their performance and suitability is difficult Especially: clearly identifying interdependencies between MAC protocol and other layers/applications is difficult Which is the best MAC for which application? 58