The U.S. Department of Defense (DoD) has invested significantly in development and deployment of aerial high-capacity backbone (HCB) networks. Understanding the performance of these Mobile Ad-Hoc Networks (MANETs) is challenging, and requires insight into multiple layers of the protocol stack. Tons of data get generated at each layer of the stack. MIT LL has developed a data collection and visualization framework to parse through important data and help monitor and analyze the performance of these networks. The HCB data collection and analysis framework is comprised of pluggable data collection and reporting daemons, a persistent storage component based on a time-series database, and a visualization dashboard capable of displaying network performance metrics in real-time and playback modes. In this paper, we showcase the capabilities of this framework, identifying how it has helped further our research, and how it can be adapted to support other similar research programs that generate tons of data.

In highly dynamic wireless environments, link metrics such as link quality, availability, and others have become increasingly important to enable smart multi-hop routing decisions. In recent years, a number of radio-to-router interface (R2RI) protocols such as Point-to-Point over Ethernet RFC5578, Dynamic Link Exchange Protocol (DLEP), and Radio-Router Control Protocol (R2CP) have emerged to address the need to have a common set of link metrics exposed from the radio to the router to enhance multi-hop routing decisions. To fully evaluate R2RI functionality and specifications, differing implementations of both radio/client and router/server-side R2RI protocols must be prototyped and tested. In this paper, we present comparison tests of each of the three (3) radio-to-router interfaces with two (2) router/server-side R2RI experimental/beta R2RI implementations: one on a commercial Cisco router, and one on an open source Quagga platform. The goal of the comparison is not necessarily to provide a holistic performance comparison (as much of the code is experimental), but to highlight implementation differences and potential issues. In many cases, issues are already resolved in future releases.1

Timely and accurate situational awareness is one of the key elements of any effective military campaign. As such, significant effort has been invested in the DoD to share location, blue force tracking, and other track data to enable commanders to make effective decisions. What has been less available are network and link visualization tools for network administrators at the tactical edge to quickly diagnose problems and assess the performance of highly intermittent networks. Although many radio systems have built-in proprietary tools that show connectivity diagrams, these tools are extremely difficult to adapt to other systems due to their closed nature. In this paper, we present a real-time open source network visualization tool for rapid network and link situational awareness on heterogeneous networks. Using widely available software such as Google Earth and various distribution systems, we show that network connectivity over several heterogeneous radio systems can be easily visualized to aid in decision making1.

The current generation of long-range, high capacity, military radios are stove-piped systems that work well in a homogeneous environment, but require significant setup and configuration to interoperate with other radio systems. Each radio provides a subset of disparate link information in nonstandard interfaces and has built-in home-grown or industry-based routers running potentially different routing protocols. In a heterogeneous radio system airborne environment, wireless link characteristics change rapidly, often requiring direct link feedback from the radio to make routing decisions. In recent years, there has been a number of work in developing a common radio-to-router interface that standardizes a subset of per-link information to pass to the network layer for use in dynamic MANET routing. While simulations and emulation tests can provide a baseline for how systems will perform, field-tests are crucial to demonstrate capabilities in real-world operating environments. In this paper, we present measurement results from a field test involving three airborne and two ground assets with various radio systems that test an implementation of RFC4938, a radio-to-router interface protocol, and its interaction with a modified OSPFv3 routing protocol to support dynamic link metrics and OSPF cost generation. The assets participated in the exercise formed a high capacity, dynamically routed aerial IP backbone made of heterogeneous radio technologies over 250 nautical miles (Nm), allowing the passing of military operational traffic.1

The Airborne Network Definition (AND) project had the goal of creating and testing a robust, efficient network architecture networking all elements of the battlefield. This effort has since been generalized to fit the goals of the Mobile Edge Network System Architecture (MENSA) effort. The network is designed to be self-contained, attaching to fixed backbone infrastructure whenever possible. The fundamental building block of the architecture is the Small Combat Network (SCN), which integrates heterogeneous ground and air platforms, facilitating collaborative applications such as Blue Force Tracking, Cooperative Sensing and Targeting. The architecture uses the concept of an IP core to network unlike domains (radio types) and to connect the SCN to the backbone and the Global Information Grid (GIG). This paper describes the requirements of the network and outlines the technical design of the SCN architecture. We present step-by-step descriptions of a communication on the SCN which highlights some of the key features of the architecture. We present results of a simulation that applies our proposed architecture to realistic warfighting scenarios. Results show that the architecture enables cooperative applications and point to future work that will design and evaluate a deployable AN network architecture.