Authors:
            
                    Manuel Koschuch
                    
                        
                    
                    ; 
                
                    Matthias Hudler
                    
                        
                    
                     and
                
                    Zsolt Saffer
                    
                        
                    
                    
                
        
        
            Affiliation:
            
                    
                        
                    
                    FH Campus Wien - University of Applied Science, Austria
                
        
        
        
        
        
             Keyword(s):
            Iris, MicaZ, Sensor Motes, Hash Functions, Performance, ATmega128.
        
        
            
                Related
                    Ontology
                    Subjects/Areas/Topics:
                
                        Data Communication Networking
                    ; 
                        Implementation and Experimental Test-Beds
                    ; 
                        Performance Evaluation
                    ; 
                        Sensor Nets and Embedded Systems
                    ; 
                        Software Engineering
                    ; 
                        Software Project Management
                    ; 
                        Telecommunications
                    
            
        
        
            
                Abstract: 
                Cryptographic hash functions are an important building block used in many cryptosystems. The flexibility and ability of a system designer to choose the most fitting function for a given system enables fast, efficient and secure designs. In this position paper we give preliminary results of porting three selected hash algorithms to Iris and MicaZ Sensor nodes in terms of achieved performance, memory requirements and the influence of different compiler optimizations on these measurements. Our main goal is to provide a sort of baseline approximation of how much effort is needed to port reference code of these algorithms to a new platform without trying to optimize it, leaving all this work to the compiler; enabling designers to not having to stick to already ported algorithms, when they might be suboptimally suited for a given environment.