Among the key advantages that led to the emergence of the TCP/IP protocol suite as the foundation of the Internet architecture was publication of the TCP/IP specifications as fully open standards, which could be implemented by anybody. Proprietary networking architectures such as IBM’s Systems Network Architecture (SNA), Digital Equipment’s DECnet, and the Xerox Network Systems (XNS) framework lent themselves less easily to widespread adoption.
In the Interplanetary Networking community we are trying hard to replicate that success by establishing universally available open Delay-Tolerant Networking standards. We hope to encourage a wide range of interoperable protocol implementations that can address all the use cases that anyone can think of.
So far, progress is encouraging. Among the implementations of Bundle Protocol that we know of are:
- DTN2, the original reference implementation, developed largely by Mike Demmer at UC Berkeley.
- DTN2’s lineal descendant DTNME, currently in use for International Space Station (ISS) operations and maintained by NASA’s Marshall Space Flight Center.
- ION, likewise in use for ISS operations, developed and maintained mainly at NASA’s Jet Propulsion Laboratory.
- cFS BPlib, soon to fly on the PACE mission, developed and maintained at NASA’s Goddard Space Flight Center.
- An implementation developed by the European Space Agency (ESA).
- IBR-DTN, developed at Technische Universität Braunschweig.
- uPCN, developed by D3TN GmbH, Dresden.
- HDTN, a high-speed implementation developed at NASA’s Glenn Research Center.
- PyDTN, written in Python X-Works.
- Experimental implementations written in Java, Go, and Rust.
More important than the number and variety of implementations, though, is the demonstrated interoperability of those implementations. Interoperation venues have ranged from the informal, as in the uPCN/PyDTN interoperability testing performed at the IETF 101 Hackathon, to the operational, as in the ION/DTNME-based architecture supporting ISS and the ION/BPlib framework supporting PACE.
In January an international team executed an especially gratifying testbed demonstration, in preparation for a planned interoperation experiment that will include Lunar Ice Cube mission communications. The testbed included:
- One DTN node running cFS BPlib, emulating the Lunar Ice Cube spacecraft.
- One DTN node running ION, emulating the Lunar Ice Cube mission operations center.
- One DTN node running ESA’s implementation of BP, emulating an ESA ground station, which forwarded bundles between the other two nodes.
The cFS BPlib code base does not include an implementation of Licklider Transmission Protocol, instead relying on Aggregate Custody Signaling (ACS) for reliability in bundle transmission. However, the ESA BP implementation does not include an implementation of ACS; instead, the ION node closes the custody transfer loop with the emulated spacecraft, with the ESA node forwarding bundles from the ION node including the aggregate custody signals.
This may be the first demonstration of sustainable Solar System Internet architecture, relying on the interoperability of different BP implementations developed by different national space agencies. We think it won’t be the last.
