Mathematical basis for the chaining of lossy interface adapters

Abstract

Despite providing similar functionality, multiple network services may require the use of different interfaces to access the functionality, and this problem will only become worse with the widespread deployment of ubiquitous computing environments. One way around this problem is to use interface adapters that adapt one interface into another. Chaining these adapters allows flexible interface adaptation with fewer adapters, but the loss incurred because of imperfect interface adaptation must be considered. This study outlines a matrix-based mathematical basis for analysing the chaining of lossy interface adapters. The authors also show that the problem of finding an optimal interface adapter chain is NP-complete with a reduction from 3SAT.