Quantitative evaluation Sample Clauses

Quantitative evaluation. The Matlab based tool uses log-files generated by Hermes for calculation of quantita- tive measures for specified key performance indicators (i.e. journey time, resilience, punctuality, energy consumption, resource usage ). The result can also be visualized in form of a number of diagrams. See examples below and in the document D4.3 from the ON-TIME project.
Quantitative evaluation. To quantitatively evaluate the performance of the system, we simulated contract gen- eration with varying (i) modes of operation (TCP/TLS/HTTP/HTTPS), (ii) number of facts per contract, and (iii) certificate chain lengths of both peers. We quantitatively evaluate the (a) total duration of the handshake, (b) the number of bytes transferred, as well as (c) the size of the generated contract in its default JSON representation. We employ a TCP proxy to investigate the impact of varying network latency on the proto- col’s performance and measure the amount of data transferred. Overhead for TLS and HTTP are included in the results. We split the evaluation into two orthogonal parameter combinations to facilitate visualization and discussion, which we show in Table 1. Contract Size & Communication Overhead. In Fig. 3, we show how the number of facts in one contract influences contract bytes and communication overhead per fact, split by handshake mode. A per-fact plot brings better comparability to other approaches than per-contract, as contracts are a concept that is specific to our approach. Thus, metrics are plotted on a per-fact (i.e., per-resource) basis. The total contract size and communication overhead per handshake increase linearly with the number of facts, as fact data is of constant size, consisting of a RID and a checksum. Thus, if m models one of the per-contract metrics, this gives m(n) = sn + c, where s is the slope of the curve, i.e., the bytes by which the metric increases if one fact is added per contract, n is the number of facts per contract, and c is the constant overhead which is not influenced by the number of facts. Then, we can model the per-fact metric Table 1. Overview of the evaluated parameter combinations. 1 TCP, TLS, HTTP, HTTPS 1, 5, 10, 20, 30, 40, 50, 0 1, 2, 3, 4, 5 20 2 TCP, TLS, HTTP, HTTPS 10 0, 10, 20, 30, 40, 50, 1 50 Contract Bytes HTTP TCP HTTPS TCP+TLS Theoretical limit: 127 Contract B/fact 10K Bytes per fact 1K 100 1 5 10 20 30 40 50 60 70 80 90 100 Facts per contract