Mobile ad hoc networks. Sources and sinks revisited. Singleg versus multiple-hops. apps) & multimedia (voice, video); i.e., humans in the loop

Transcription

1 Mobile ad hoc networks 2 Wireless Sensor Architecture Nodes N d ttalking lki tto each h other th Nodes talking to some node in another network (Web server on the, ee.g.) g) GENERAL PRINCIPLES AND ARCHITECTURES FOR PUTTING SENSOR NODES TOGETHER TO FORM A MEANINGFUL NETWORK Tpicall requires some connection to the fixed network Applications: Traditional data (http, (http ftp ftp, or collaborative apps) & multimedia (voice, video); i.e., humans in the loop Access Point s and sinks revisited Singleg versus multiple-hops p p 3 4 s: An entit that provides data/measurements s: Nodes where information is required Belongs to the sensor network as such Ù Is an external entit, e.g., a PDA, but directl connected to the WSN Main difference: comes and goes, often moves around, Ù Is part of an external network (e.g., internet), somehow connected to the WSN Ù One common problem: limited range of wireless communication Essentiall due to limited transmission power, path loss, obstacles Option: multi-hop networks Send packets to an intermediate node Intermediate node forwards packet to its destination Store and forward multi-hop Store-and-forward multi hop network Basic technique applies to both WSN and MANET Note: Store&forward multi-hopping possible solution pp g NOT the onl p E.g., collaborative networking, network coding Do not operate on a per-packet basis Obstacle

2 Hopping pp g in WSNs s of mobilit 5 6 Node mobilit A node participating as source/sink (or destination) or a rela node might move around Deliberatel, Deliberatel self self-propelled propelled or b external force; targeted or at random Happens in both WSN and MANET mobilit In WSN, a sink that is not part of the WSN might move Mobile requester Event mobilit In WSN, event that is to be observed moves around (or extends, shrinks) Different WSN become responsible for surveillance of such an event mobilit Event mobilit 7 8 q Request Propagation p g of answers Movement direction

3 Challenging g goals: QoS In MANET: Usual QoS interpretation Throughput/dela/jitter High perceived QoS for multimedia applications In WSN, more complicated 9 Event detection and/or reporting probabilit Event classification error Detection dela Probabilit of missing a periodic report Approximation accurac (e.g., when WSN constructs a temperature map) Tracking accurac (e.g., difference between true and conjectured position of the pink elephant) Related goal: robustness Network should withstand failure of some Challenging g goals: Efficienc Energ per correctl received bit Counting all the overheads, in intermediate, etc. Energ per reported (unique) event After all, information is important, not paload bits! Dela/energ tradeoffs Network lifetime Time to first node failure Network half-life (how long until 50% of the died?) Time to partition Time to loss of coverage Time to failure of first event notification 0 Challenging g goals: Scalabilit Network should be operational regardless of number of At high efficienc Tpical node counts difficult to guess MANETs: 0s to 00s WSNs: 0s to 000s, mabe more (possible?) Requiring to scale to large node numbers has serious consequences for network architecture Might not result in the most efficient solutions for small networks! Carefull consider the actual application i needs before looking for man solutions! Distributed organization In-network processing Aggregation Design principles p Distributed source coding and distributed compression Distributed and collaborative signal processing Mobile code/agent networking Adaptive fidelit Data centricit Address data, not Implementation options Overla networks and distributed hash tables Publish/subscribe Databases 2

4 Exploit: Location information Activit patterns Heterogeneit Design principles 2p Component-based protocol stacks and cross-laer optimization Service interfaces Gatewa concepts 3 Distributed organization Participants i t should cooperate in organizing i the network E.g., with respect to medium access, routing, Centralistic approach as alternative usuall not feasible hinders scalabilit, robustness Potential shortcomings 4 Not clear whether distributed or centralistic organization achieves better energ efficienc (when all overheads considered) Option: limited it centralized solution Elect for local coordination/control Perhaps rotate this function over time In-network processing MANETs are supposed to deliver bits from one end to the other WSNs, however, are expected to provide information, not necessaril original bits Additional options E.g., manipulate or process the data in the network Main example: aggregation 5 Appl composable aggregation functions to a converge cast tree in a network Tpical functions: minimum, maximum, average, sum, Not amenable functions: median In-network processing: Aggregation g Reduce number of transmitted bits/packets t b appling an aggregation function in the network 6 3 6

5 In-network processing: Distributed 7 Depending on application, more sophisticated processing of data can take place within the network Example edge detection: locall exchange raw data with neighboring, compute edges, onl communicate edge description to far awa data sinks Example tracking/angle detection of signal source: Conceive of sensor as a distributed microphone arra, use it to compute the angle of a single source, onl communicate this angle, not all the raw data Exploit temporal and spatial correlation Observed signals might var onl slowl in time! no need to transmit all data at full rate all the time Signals of neighboring are often quite similar! onl tr to transmit differences (details a bit complicated, see later) Seek other programming paradigms: mobile code, swarm intelligence Adaptive fidelit Adapt the effort with which h data is exchanged to the currentl required accurac/fidelit Example: event detection When there is no event, onl ver rarel send short all is well messages When event occurs, increase rate of message exchanges Example: temperature When temperature is in acceptable range, onl send temperature values at low resolution When temperature becomes high, increase resolution and thus message length 8 Data-centric networking In tpical networks (including ad hoc networks), network transactions are addressed to the identities of specific A node-centric or address-centric networking paradigm In a redundantl deploed sensor networks, specific source of an event, alarm, etc. might not be important Redundanc: e.g., several can observe the same area Thus: focus networking transactions on the data directl instead of their senders and transmitters! data-centric networking Principal design change 9 Implementation options Overla networks & distributed ib t d hash h tables (DHT) Hash table: content-addressable memor Retrieve data from an unknown source, like in peer-to-peer networking with efficient implementation Some disparities remain 20 Static ke in DHT, dnamic changes in WSN DHTs tpicall ignore issues like hop count or distance between when performing a lookup operation

6 Publish/subscribe b ib Implementation options 2 Different interaction paradigm 2 Nodes can publish data, can subscribe to an particular kind of data Once data of a certain tpe has been published, it is delivered to all subscribes Subscription and publication are decoupled in time; subscriber and published are agnostic of each other (decoupled in identit) Databases Matches with the idea of using a data-centric organization of the networking protocols Interested in certain aspects of data == formulating queries for a database Other design principles p Exploit location information Required anwas for man applications; can considerabl increase performance Exploit activit patterns Watch for event shower effect Exploit heterogeneit 22 B construction: of different tpes in the network B evolution: some had to perform more tasks and have less energ left; some received more solar energ than others; Cross-laer optimization of protocol stacks for WSN Goes against grain of standard networking; but promises big performance gains Also applicable to other networks like ad hoc; usuall at least worthwhile to consider for most wireless networks Service interfaces The world s all-purpose network interface: sockets Good for transmitting data from one sender to one receiver Not well matched to WSN needs (ok for ad hoc networks) Expressibilit requirements Support for simple request/response interactions Support for asnchronous event notification Different was for identifing addressee of data B location, b observed values, implicitl b some other form of group membership, b some semanticall meaningful form room 23 Eas accessibilit of in-network processing functions Formulate complex events events defined onl b several Allow to specif accurac & timeliness requirements Access node/network status information i (e.g., batter level) l) Securit, management functionalit, 23 Gatewa concepts in WSNs Necessar to the t for remote access to WSNs 24 Same is true for ad hoc networks; additional complications due to mobilit ( (change route to the gatewa; use different gatewas) WSN: Additionall bridge the gap between different interaction semantics (data vs. address-centric networking) in the gatewa Gatewa needs support for different radios/protocols, Gatewa node

7 WSN to communication to WSN communication Example: Deliver an alarm message to an host Issues: I Issues Need to find a gatewa (integrates routing & service discover) Choose h b best gatewa if severall are available il bl How to find Alice or Alice s IP? Alert Alice How to find the right WSN to answer a need? How to translate from IP protocols to WSN protocols, semantics? Remote requester Ali desktop Alice s d k Gatewa Gatewa Gatewa Alice s PDA WSN tunneling g Summar Use U the th I t t to t tunnel t l WSN packets k t b between t ttwo remote WSNs Network N t k architectures hit t ffor ad dh hoc networks t k are in i principle relativel straightforward and similar to standard networks Mobilit is compensated for b appropriate protocols, but interaction paradigms don t change too much WSNs, on the other hand, look quite different on man levels Gatewa Gatewa Data-centric D i paradigm, di the h need d and d the h possibilit ibili to manipulate i l data as it travels through the network opens new possibilities for protocol design Next, we will look at how these ideas are realized b actual protocols