A common reflex that some users have is to compare the bandwidth level expressed in Strings with the statistics coming through other tools. While we understand the quest to search for a "unique source of truth", we believe that any tool should use the method that best serves its other features.
That is to say that the way in which we measure bandwidth is aimed at getting the best out of the analytical effort of our users, which for us is based around Strings, the unit of analysis that brings together many of the features we provide.
A String is a container of other devices, such as Cablemodems, and it ties together information from the network into a cohesive, easy to grasp context that could be commercial (i.e: a String with the name of a customer), geographical (i.e: a String with the name of a part of a city) or technical (i.e: a String with the name of a part of the network).
In that sense, the CMTS bandwidth levels can be different to other tools, because we do not measure or detect the bandwidth at the CMTS level. Instead, we calculate the sum of every Cablemodem that is under the control of that CMTS, and display this value as the CMTS bandwidth for inbound/outbound use cases.
This is essential for us to make sure that each String can feature a percentage bandwidth in relation to the total network bandwidth, and that the CMTS detail screen can show an accurate display of the Cablemodems associated to it, and each bandwidth percentage in relation to the total bandwidth for that CMTS.
Since the purpose of the calculation is different, the method is also different. This does not mean that one method is correct, and the other incorrect, but rather that they are using the best way to help people make decisions given their feature set.