Opportunistic computing forwireless sensor networks

small. Because ofall these constraints, hardware resources ofnodes areextremely limited. Wireless sensor networks aremoving fromacademia to Inthelast fewyears several hardware platforms, with real world scenarios. Thiswill involve, inthenearfuture,similar characteristics, havebeendesigned andputinto prothedesign andproduction ofhardwareplatforms character-duction toenable thedevelopment ofWSN applications. ized bylow-cost andsmallform factor Asaconsequence, Examples include MicaandMica2, produced byCrossbow, theamountofresources available onasingle node, i. e. andTelos, produced byMoteiv. Allthese systems havebeen computing power, storage, andenergy, will beevenmore usedasthehardware baseinalarge numIber ofresearch constrained thantoday. Thispaper faces theproblem of projects, especially inthearea ofsoftware systems. Generstoring andexecuting anapplication that exceeds themem- ally, theamount ofprogram memorythat isavailable ona oryresources available onasingle node. Theproposed so- single noderanges from48to128KBytes. lution isbased ontheideaofpartitioning theapplicationItshould benoted that, since these kindsofsystems codeinto anumberofopportunisticall y cooperating mod- havebeendesigned andmanufactured tobeusedinreules. Eachnodecontributes totheexecution oftheoriginalsearch projects, boththeformfactor andproduction costs application byrunning asubset oftheapplication tasks and arenotadequate fortheir adoption inreal ubiquitous comproviding service totheneighboring nodes.