AT TOPPaper 9048 AT BOTTOM
Deriving Traffic Models from Netflow Tactical Data Collection
Schmidt,TatyanaThe MITRE Corporation
Dimarogonas,JimThe MITRE Corporation
Current tactical units rely heavily on communications to exchange information within the emerging “network centric” paradigm of modern warfare. Network resources are starting to be viewed as critical resource to a units force effectiveness as ammo and armor has in the past. Surprisingly though, little is known or documented about network utilization by “digitized” units and the effects on network utilization of a unit’s operations. Questions such as how much capacity do tactical units use or how does the use of network resources change in the context of an operational mission remain largely unanswered. Consequently, the current approach to network design and capacity planning is based largely on educated guesses and a “more is better” philosophy, rather than a rigorous approach of estimation and prediction of warfighters needs. To create a rigorous framework for predicting future traffic loads we need information about the distribution of usage of the different network services, parameters such as message sizes and frequencies, as well as their statistical distribution in time to develop a “profile” of the traffic within an operational context. These traffic profiles can be used by the M&S community for refining and configuring current networks and applications and to provide the basis of future C4ISR system designs. This paper describes the methodology of analyzing data collected from tactical unit in training, identifies the information necessary to build a traffic profile and finally presents the results as applied to NIPR traffic collected from the unit. From this we will extract what is a typical traffic profile by echelon and sender-receiver operational relationship and show how we can extrapolate this information onto other force structures and deployments creating a holistic view of the units network resource needs based upon a planned deployment and the best available information collected from past deployments.