an and d Im Impl plementa ementation tion (01 0120 20442 - PowerPoint PPT Presentation

Net etwork work Ke Kerne nel l Ar Archi chitect tectures ures an and d Im Impl plementa ementation tion (01 0120 20442 4423) ) Naming aming an and Ad Addressing ressing Chaiporn Jaikaeo Chaiporn.j@ku.ac.th Department of Computer Engineering Kasetsart University Materials taken from lecture slides by Karl and Willig

Name Na mes vs. . Add Addre resses Names: s: Refer to “things”  Nodes, networks, data, transactions, …  May or may not be globally unique  Ad Addres esse ses: : Information needed to fi find these  things Street address, IP address, MAC address  May or may not be globally unique  Services to map between names and addresses  E.g., DNS  Some names are also addresses  2

Na Nami ming ng in in WS WSN Nodes are not independent   But collaborate to solve a given task Better to shift view from naming nodes to  na naming ing data 3

Addre Add ress s Man Manage ageme ment nt Is Issue ues Address dress al allo loca catio ion: n: Assign an entity an  address from a given pool of possible addresses  Distributed address assignment (centralized like DHCP does not scale) Address dress deal allo locat cation: ion: Once address no  longer used, put it back into the address pool  Because of limited pool size  Graceful or abrupt, depending on node actions 4

Add Addre ress s Man Manage ageme ment nt Is Issue ues Address representation  Conflict detection & resolution (Duplicate  Address Detection - DAD)  What to do when the same address is assigned multiple times?  Can happen e.g. when two networks merge Binding   Map between addresses used by different protocol layers  E.g., IP addresses are bound to MAC address by ARP 5

Un Unique iqueness ness of f Add Addre resses Globally unique   Appears at most once all over the wo worl rld Network-wide unique   Appears at most once in a given netwo work rk Locally unique   Appears at most once in a defined neighbo borho rhood 6

Add Addre ressing sing Ov Overh rhead ead The fewer bits per address, the better  Global > Network-wide > Local  Tradeoffs   Address length  management overhead Typically, address negotiation runs only at  the beginning  Except when there is mobility 7

Di Distributed tributed Ad Addr dress ess As Assignment ignment Option ion 1: Random assignment   Unacceptable high risk of duplicate addresses  No-conflict probability for n addresses and k nodes is  By Stirlings approximation  Similar to the birthday paradox 8

Di Distributed tributed Ad Addr dress ess As Assignment ignment Option ion 2: Still random, but avoid addresses  used in local neighborhood  By overhearing exchanged packets  Good enough in many WSN apps where data sent to a certain sink 9

Di Distributed tributed Ad Addr dress ess As Assignment ignment Option ion 3: Repair any observed conflicts   Randomly pick a temporary address and a proposed fixed address  Send an addres ess requ quest t to the proposed address, using temporary address  If addres ess s reply y arrives, address already exists  Collisions in temporary address unlikely, as only used briefly Option ion 4: Similar to 3, but use a neighbor  that already has a fixed address to perform requests 10

Locally Lo cally Un Unique ique Add Addre resses Fewer bits are needed, due to   Each address can be reused several times across the same network  Lower-number addresses tend to be used more frequently Addresses can be compressed  E.g., using Huffman coding  11

Is Issues ues wi with th As Asym ymme metr tric ic Li Link nks Assume nodes communicate with bidirectional  neighbors only All bidirectional neighbors of each node must have  distinct addresses The address of any inbound neighbor must be different  from all bidirectional neighbors 12

Cont ntent ent-Base Based d Add Addre ressing ing Recall: Paradigm change from id-centric to  data-centric networking in WSN Supported by content-based  names/addresses  Do not described involved nodes (not known anyway), but the co content nt itself the interaction is about Classical option: Put a naming scheme on  top of IP addresses  Done by some middleware systems 13

De Descr cribing ibing In Inte terests rests Inter eres ests ts describe relevant data/event  Nodes match these interests with their locally  observed data Format: Attribute-Value-Operation (AVO)  E.g.: <TEMP, 20 ° C, GE>  Operations:  14

Des Describing cribing In Inte tere rest/Sensor/Data st/Sensor/Data List of AVOs  E.g.,  <type,temperature,IS> <x-coordinate,10,IS> <type,temperature,EQ> <y-coordinate,10,IS> <threshold-from-below,20,IS> <x-coordinate,20,LE> Sensor <x-coordinate,0,GE> <y-coordinate,20,LE> <y-coordinate,0,GE> <type,temperature,IS> <interval,0.05,IS> <x-coordinate,10,IS> <duration,10,IS> <y-coordinate,10,IS> <class,interest,IS> <temperature,20.01,IS> <class,data,IS> Interest Data 15

Mat Match ching ing Al Algo gorithm rithm Check whether an interest matches the  locally available data 16

Di Dire rected cted Di Diffu ffusion ion An example of data-centric networking  17

Geographic Ge graphic addr addressing ing Express addresses by denoting physical  position of nodes  Considered a special case of content-based addresses  Attributes for x and y (and z) coordinates Options   Single point  Circle or sphere centered around given point  Rectangle by two corner points  Polygon 18

Conclusion nclusion Addresses can be assigned distributedly  Non-id-centric addresses give additional  expressiveness, enables new interaction patterns than only using standard addresses These addresses have to be supported by  specific protocols, in particular, routing protocols 19