We present empirical measurements of the upstream throughput of a DOCSIS 1 1.1 link. In contrast to all previous simulation-based studies, our measurements have been obtained from actual cable-modems (CMs) and head-ends, both from two different vendors each. We have constructed an exhaustive database of measurements of a large subset of the space of parameters affecting upstream throughput. Using a wellknown non-parametric hypothesis test, we query this database for obtaining statistically robust answers to key questions about the effect of parameter changes on the throughput. Our results indicate that for a single CM scenario, packet concatenation is most effective whereas piggybacking is effective and better than concatenation only in some cases. Using both enhancers decreases throughput for a single CM scenario. Our results are robust across head-end implementations and are of immediate interest to network and protocol architects as well as device developers.
I. MOTIVATIONThe DOCSIS 1.0 and 1.1 standard [1] was a result of research and development into the performance implications of various QoS mechanisms [2], [3], [4], [5], [6], [7]. Since these were preliminary, prototyping studies, they were conducted using analytical and/or simulation models. There has been no extensive study of the DOCSIS protocol using actual implementations, to our knowledge. However, studies based on actual implementations provide a wealth of complementary information of interest to designers and operators alike. Measurements from an implementation offer a tractable way of capturing and evaluating a system in its entirety. For network service providers and operators, empirical data obtained from actual implementations is an indispensable input for the design and upgrade process. Providers and operators require some form of field performance report from device vendors, before a device can be deployed in a live traffic environment. Needless to say, an empirical model must be tested for robustness using standard statistical techniques. Without such testing for confidence, no useful and reliable conclusions can be drawn from the data.In this paper, we report results from our extensive performance study of the upstream portion of a DOCSIS 1.1 link based on real devices. Using our measurements as a database, we can answer key performance questions that are of immense value to the network designer. We also hope that the results obtained provide some insight into the strengths and 1 DOCSIS is a registered trademark of CableLabs.shortcomings of the various performance enhancing features provided by the protocol.
II. AN OVERVIEW OF THE DOCSIS 1.1 MAC LAYERA DOCSIS network uses the existing cable television infrastructure to deliver data services to subscribers. The network, owing to the structure of the pre-existing cable television plant, forms a tree with the root connected to a Cable Modem Termination System (CMTS). Subscribers connect to the network through a cable modem (CM) connected as a leaf of the tree. The link from the CMTS to CM, te...