Wesley M. Eddy

A bundle of problems

Delay-Tolerant Networking” (DTN) is a neologism used for a new store-and-forward architecture and protocol suite intended for disrupted networks where there is intermittent or ad-hoc connectivity. This has been proposed as one approach to supporting delay-tolerant networks. Work in the late 1990s on the “Interplanetary Internet” forms the basis for current DTN protocols and architecture. That early work considered transport protocols robust to the hours-long propagation delays of deep-space communications. DTN is also known, primarily in military circles, as Disruption-Tolerant Networking, due to the dynamic links and outages in the military tactical environment, rather than long-delay links. In both cases, DTN technologies are well-suited to applications that are mostly asynchronous and insensitive to large variations in delivery conditions. DTN networks differ sufficiently from traditional terrestrial networks in their characteristics and connectivity that link, network and transport protocols must be carefully considered and chosen to cope with these different characteristics, or new protocols can be designed that are suited for the problems that these DTN network conditions impose. The “Bundle Protocol” exists within the DTN architecture, which sends bundles over subnet-specific transport protocols, called “convergence layers.” “Bundling” has undergone a large amount of shared development and design over a period of years as a research effort. We examine the Bundle Protocol and its related architecture closely, and discuss areas where we have found that the current Bundle approach is not well-suited to many of the operational concepts that it was intended to support. Problems with the Bundle Protocol and its convergence layers exist in mechanisms for error detection and overall reliability. This weakens the Bundle Protocols suitability to disrupted and error-prone networks. We show that these reliability issues can lead to performance problems in DTN networks, requiring mitigation. Open research and development areas also exist with design choices in handling timing information, in determining necessary and sufficient security mechanisms, in its Quality of Service capabilities, and in other aspects of application or content identification. We show that the existing DTN bundling architecture has a number of open real-world deployment issues that can be addressed. We suggest possible remediation strategies for these weak areas of the bundle protocol that we have been working on. We also look at alternate approaches to DTN networking. Rather than only providing criticism, this paper identifies open issues, where work on modifying the Bundle Protocol is encouraged and approaches to address its various problems are suggested.

Experience with Delay-Tolerant Networking from orbit†

The disaster monitoring constellation (DMC), constructed by Surrey Satellite Technology Ltd (SSTL), is a multi-satellite Earth-imaging low-Earth-orbit sensor network where captured image swaths are stored onboard each satellite and later downloaded from each satellite payload to a ground station. The DMC is currently unique in its adoption of the Internet Protocol (IP) for its imaging payloads and for satellite command and control, based around reuse of commercial networking and link protocols. Earth images are downloaded from the satellites using a custom IP-based high-speed transfer protocol developed by SSTL, Saratoga, which works well in unusual link environments. Store-and-forward of images with capture and later download during passes over ground stations gives each satellite the characteristics of a node in a Delay/Disruption Tolerant Network (DTN). DTNs are under investigation in an Internet Research Task Force (IRTF) DTN research group (RG), which has developed a bundle architecture and protocol. We experiment with use of this DTNRG bundle concepts onboard the UK-DMC satellite, by examining how Saratoga can be used as a convergence layer to carry the DTN Bundle Protocol, so that sensor images can be delivered to ground stations and beyond as bundles. This is the first use of the Bundle Protocol from orbit. We use our practical experience to examine the strengths and weaknesses of the Bundle architecture for DTN use, paying attention to fragmentation, custody transfer, and reliability issues. We similarly examine and discuss an alternative network stack, based around proposed use of the Hypertext Transfer Protocol (HTTP) that we have been architecting, which we believe has potential applications across a range of DTN networks. We use our practical experience to make suggestions about DTN use and adoption of IP in sensor networks.

Experience with Delay-Tolerant Networking from Orbit

The disaster monitoring constellation (DMC), constructed by Surrey Satellite Technology Ltd (SSTL), is a multi-satellite Earth-imaging low-Earth-orbit sensor network where captured image swaths are stored onboard each satellite and later downloaded from each satellite payload to a ground station. The DMC is currently unique in its adoption of the Internet Protocol (IP) for its imaging payloads and for satellite command and control, based around reuse of commercial networking and link protocols. Earth images are downloaded from the satellites using a custom IP-based high-speed transfer protocol developed by SSTL, Saratoga, which works well in unusual link environments. Store-and-forward of images with capture and later download during passes over ground stations gives each satellite the characteristics of a node in a Delay/Disruption Tolerant Network (DTN). DTNs are under investigation in an Internet Research Task Force (IRTF) DTN research group (RG), which has developed a bundle architecture and protocol. We experiment with use of this DTNRG bundle concepts onboard the UK-DMC satellite, by examining how Saratoga can be used as a convergence layer to carry the DTN Bundle Protocol, so that sensor images can be delivered to ground stations and beyond as bundles. This is the first use of the Bundle Protocol from orbit. We use our practical experience to examine the strengths and weaknesses of the Bundle architecture for DTN use, paying attention to fragmentation, custody transfer, and reliability issues. We similarly examine and discuss an alternative network stack, based around proposed use of the Hypertext Transfer Protocol (HTTP) that we have been architecting, which we believe has potential applications across a range of DTN networks. We use our practical experience to make suggestions about DTN use and adoption of IP in sensor networks.

New techniques for making transport protocols robust to corruption-based loss

Current congestion control algorithms treat packet loss as an indication of network congestion, under the assumption that most losses are caused by router queues overflowing. In response to losses (congestion), a sender reduces its sending rate in an effort to reduce contention for shared network resources. In network paths where a non-negligible portion of loss is caused by packet corruption, performance can suffer due to needless reductions of the sending rate (in response to perceived congestion that is not really happening). This paper explores a technique, called Cumulative Explicit Transport Error Notification (CETEN), that uses information provided by the network to bring the transports long-term average sending rate closer to that dictated by only congestion-based losses. We discuss several ways that information about the cumulative rates of packet loss due to congestion and corruption might be obtained from the network or through fairly generic transport layer instrumentation. We then explore two ways to use this information to develop a more appropriate congestion control response (CETEN). The work in this paper is done in terms of TCP. Since numerous transport protocols use TCP-like congestion control schemes, the CETEN techniques we present are applicable to other transports as well. In this paper, we present early simulation results that show CETEN to be a promising technique. In addition, this paper discusses a number of practical and thorny implementation issues associated with CETEN.

Large File Transfers from Space Using Multiple Ground Terminals and Delay-Tolerant Networking

Abstract-We use Delay-Tolerant Networking (DTN) to break control loops between space-ground communication links and ground-ground communication links to increase overall file delivery efficiency, as well as to enable large files to be proactively fragmented and received across multiple ground stations. DTN proactive fragmentation and reactive fragmentation were demonstrated from the UK-DMC satellite using two independent ground stations. The files were reassembled at a bundle agent, located at Glenn Research Center in Cleveland Ohio. The first space-based demonstration of this occurred on September 30 and October 1, 2009. This paper details those experiments.

Sharing the dream

We reconsider desirability of the Bundle Protocol (BP) as a universal solution for Delay- and Disruption-Tolerant Networking (DTN). The BP is intended to provide a single solution that is applicable to a wide variety of differently-challenged DTN networks, even though those networks are unlikely to interact with one another. This paper asks whether such a single protocol can encompass all varied DTN networking needs. It asks whether attempting to repeat the previous success of the homogeneous Internet by layering over all networks is suitable for the heterogeneous DTN world, where diverse application needs and operational requirements lead to diverse, scenario-specific, applications. This position paper is intended to encourage discussion of the role, scope, and adoption of the BP.