This work considers the problem of finding analytical expressions for the expected values of dis- tributed computing performance metrics when the underlying communication network has a complex structure. Through active probing tests a real distributed computing environment is analysed. From the resulting network, ensembles of synthetic graphs with additional structure are used in Monte Carlo simulations to both validate analytical expressions and explore the performance metrics under different conditions. Computing paradigms with different hierarchical structures in computing ser- vices are gauged, fully decentralised (i.e., peer-to-peer) environments providing the best performance. Moreover, it is found that by implementing more intelligent computing services configurations (e.g., betweenness centrality based mappings) and task allocations strategies, significant improvements in the parallel efficiency can be achieved. We qualitatively reproduce results from previous works and provide closed-form solutions for the expected performance metrics linking topological, application structure and allocation parameters when job dependencies and a complex network structure are considered.