Edward J. Birrane

Contact graph routing in DTN space networks: overview, enhancements and performance

Delay- and Disruption Tolerant Networks (DTNs) are based on an overlay protocol and on the store-carry-forward paradigm. In practice, each DTN node can store information for a long time before forwarding it. DTNs are particularly suited to cope with the challenges imposed by the space environment. This paper is focused on routing in space DTNs, and in particular on contact graph routing (CGR) and its most representative enhancements, available in the literature, which are briefly surveyed in this work. Moreover, the applicability and the obtained performance of the DTN protocol stack and of the CGR have been evaluated by presenting results from real experimental experiences such as the Deep Impact Network experiment (employing the EPOXI space cruise), the JAXA jointly performed space link demonstrations with NASA (where the JAXAs GEO relay satellite called Data Relay Test Satellite has been used), the Space Data Routers European Project, and the pilot operation of a DTN implementation on the International Space Station (ISS).

Congestion modeling in graph-routed Delay Tolerant Networks with Predictive Capacity Consumption

We present Predictive Capacity Consumption (PCC), a congestion modeling extension to graph-based routing protocols. This extension provides a solution to the problem of flow control in Delay-Tolerant Networks (DTNs) and other overlays that can neither synchronize physical link state across the network nor negotiate bandwidth consumption bridging heterogeneous link layers. PCC enables the construction of a distributed, predictive congestion model independent of the underlying link layer without requiring excessive broadcasts or other mechanisms unfeasible in DTNs. PCC examines information generated by routing protocols and adjusts local routing graphs to account for predicted message paths, correcting for downstream congestion and message retransmission. Unlike other mechanisms, the flow control provided by PCC can be implemented anywhere a graph-based routing methodology is used and the adoption of this method requires only minor modification to the in-situ routing framework. We describe the PCC algorithm, analyze its operation, and demonstrate its performance by simulating multiple data streams driving a set of constrained networks to saturation. The simulation results show that PCC improves the throughput of the network by 97% over table routing approaches and by 37% over graph routing approaches without congestion models.

Congestion modeling and management techniques for predictable disruption tolerant networks

Delay and disruption tolerant networks (DTNs) are becoming an appealing solution for extending Internet boundaries so as to embrace disruptive communications. In particular, if node trajectory can be predicted as in space networks, routing schemes can take advantage of the a-priori knowledge of a contact plan. Despite mechanisms such as Contact Graph Routing (CGR) exist, they might derive in harmful overbooking of forthcoming contacts, also known as congestion. In order to tackle congestion, we initially formulate the problem by means of a linear programming (LP) model so as to establish an upper theoretical bound of performance. Next, we survey existing congestion mitigation mechanisms for predictable DTNs to later contribute with a CGR extension named PA-CGR. Finally, in the pursuance of an optimal congestion avoidance approach, we also propose and evaluate in a realistic scenario a novel multi-graph technique (MG-CGR) that outperforms existing solutions by exploiting traffic predictability.

Defining Tolerance: Impacts of Delay and Disruption when Managing Challenged Networks

Challenged networks exhibit irregularities in their communication performance stemming from node mobility, power constraints, and impacts from the operating environment. These irregularities manifest as high signal propagation delay and frequent link disruption. Understanding those limits of link disruption and propagation delay beyond which core networking features fail is an ongoing area of research. Various wireless networking communities propose tools and techniques that address these phenomena. Emerging standardization activities within the Internet Research Task Force (IRTF) and the Consultative Committee for Space Data Systems (CCSDS) look to build upon both this experience and scalability analysis. Successful research in this area is predicated upon identifying enablers for common communication functions (notably node discovery, duplex communication, state caching, and link negotiation) and how increased disruptions and delays affect their feasibility within the network. Networks that make fewer assumptions relating to these enablers provide more universal service. Specifically, reliance on node discovery and link negotiation results in network-specific operational concepts rather than scalable technical solutions. Fundamental to this debate are the definitions, assumptions, operational concepts, and anticipated scaling of these networks. This paper presents the commonalities and differences between delay and disruption tolerance, including support protocols and critical enablers. We present where and how these tolerances differ. We propose a set of use cases that must be accommodated by any standardized delay-tolerant network and discuss the implication of these on existing tool development.

Improving graph-based overlay routing in delay tolerant networks

We present CGR-EB, a modification of, and extension to, the Contact Graph Routing (CGR) protocol — a forwarding mechanism for interplanetary communication. CGR-EB enables graph-based overlay routing for a variety of networks, including those using vehicular assets as data mules. It improves CGR by storing end-to-end message paths and encoding these paths, and the sub-graphs that spawned them, with the message. Simulation results demonstrate that CGR-EB reduces processing by up to two orders of magnitude, better tolerates errors in the network graph, and enables the use of cost functions that optimize system-level network state over individual message delivery. Graph-based overlay routing improves data exchange in mobility-enabled delay-tolerant networks by reducing or eliminating link negotiation and node discovery message overhead. CGR-EB provides the mechanism through which terrestrial networks may implement this routing approach.

Maximizing data return for the Europa lander: A trade study in the application of CCSDS protocols

In support of NASA, Caltechs Jet Propulsion Laboratory and The Johns Hopkins Applied Physics Laboratory are studying concepts for two missions to explore Europa: a multiple flyby spacecraft and a surface lander. This paper analyzes the use of packetized, multi-hop, multi-path communications protocols for the Europa lander concept and assesses their potential for reducing power requirements while increasing data return. Analysis includes three protocols standardized by the Consultative Committee for Space Data Systems (CCSDS): the CCSDS File Delivery Protocol (CFDP), the Bundle Protocol (BP), and the Licklider Transmission Protocol (LTP). A spacecraft may implement a networking stack of one or more of these protocols, with each of these stacks exhibiting different strengths and weaknesses. We present heuristic and analytical methods for evaluating protocol performance including a priori computations of protocol overheads, Monte-Carlo analysis across bit error rates and packet sizes, and high fidelity simulations. Quantitative metrics such as retransmission efficiency, packet overhead, and end-to-end transaction duration characterize individual protocol options. Qualitative metrics such as cost of ownership, mission operations complexity, and computational processing load characterize the mission impacts of various networking stacks. We generate results using anticipated mission link characteristics, data volumes, and network geometries and provide recommendations relating to the value of software protocols and multi-protocol networking stacks. Results demonstrate that each candidate protocol combination can be tuned to within 15% of optimal performance over links of up to 10−4 bit error rate, although achieving this efficiency with solely CFDP incurs up to 800% greater computational processing load versus the other stacks. We conclude multiprotocol stacks separate concerns when optimizing performance for multiple stakeholders. A CFDP/BP/LTP networking stack solves a joint optimization problem where CFDP can be tuned for onboard data operations, BP can be used to provide standardized priority and store-and-forward operations, and LTP can be tuned for retransmission and acknowledgement. This approach enables efficient end-to-end communications for the Europa lander concept that maximizes data return with minimal power requirements.

The path to space-terrestrial internetworking

A fundamental challenge for the information age is to extend networked communication and its benefits (increased quality of life, better understanding and utilization of our planet, more efficient industry) across the globe. One obstacle to that challenge is the fact that the majority of the Earth and its space environs cannot support the timely, reliable, low-cost telecommunications infrastructure that currently enables rapid growth. We address the question of whether adding spacecraft as nodes in a global network sustains the financial and technological progress necessary to overcome this obstacle. Space-based networked nodes, offer unique technical advantages for global coverage and are approaching, at scale, economic feasibility. Such approaches provide the critical communications infrastructure for a step increase in our use of space, and provide a complementary resource for global terrestrial coverage. However, the vision of space-terrestrial internetworking is threatened by the lack of a converging vision for how such a network would operate. We propose a logical architecture, informed by the evolution of terrestrial cellular networks, that provides a unifying vision of space networking that considers existing infrastructure, unique spacecraft constraints, and the capabilities that make such an internetwork useful. We further provide a roadmap for organizing the space networking community in an attempt to avoid the inefficiencies of proprietary, non-interoperable network constellations. We conclude that spacecraft can provide a compelling way to network those portions of our planet that lack infrastructure and efficiently disseminate common information to large populations.

Quality of Service and Message Aggregation in Delay-Tolerant Sensor Internetworks

We present traffic-shaping and message-aggregation algorithms that provide reservation-based quality-of-service mechanisms for delay-tolerant internetworks utilizing graph-based routing protocols. We define a Traffic Shaping with Contacts (TSC) method that alters the edge weights in a graph structure to represent service level specifications, rather than physical capacity. This adjustment allows existing routing mechanisms to implement bandwidth reservations without additional processing at the node. We define a Payload Aggregation and Fragmentation (PAF) algorithm that calculates preferred payload sizes over traffic-shaping contacts. PAF aggregates too-small payloads together and fragments too-large payloads to optimize contact capacities. Unlike other mechanisms, TSC/PAF are unaffected by heterogeneous physical, data-link, and transport layer protocols across an internetwork and require only minor modifications to internetwork-layer graph-routing frameworks. Simulation results show that together TSC/PAF reduce the number of messages in a sensor internetwork by 43 % while increasing the goodput of the network by 63 % over standard graph-routing techniques.