Networking

Recent events at NASA have overtaken the internal debates described in your August article ["The Interplanetary Internet"]. It is important to note that there has been absolutely no disagreement between NASA Goddard Space Flight Center and NASA Jet Propulsion Laboratory that the Internet Protocol suite could be operational in space environments. Extended Internet protocols were developed and test-flown by JPL in the mid-1990s, and, as the article described, IP capabilities were demonstrated during the Goddard CANDOS experiment on the shuttle in 2003.

Disagreements between JPL and Goddard centered not on the viability of the IP suite but on the selection of the standard underlying space-link protocol over which they should run. Those disagreements have now been largely resolved.

We therefore all agree (and more important, we have never disagreed) that IP-based communications are appropriate for those space applications that share a connected, short-delay environment—that is, to and from Earth-orbiting spacecraft, on and around the surface of the moon or Mars, and probably between the Earth and the moon.

Delay Tolerant Networking (DTN) protocols were not conceived to compete with or to replace TCP/IP, but to complement it. With DTN, the characteristics of the underlying data path are irrelevant. DTN-based communications can thus reliably transport data even in the presence of extreme network disruptions. It is no accident that DARPA is now investing in DTN technology for extension of Internet services into highly stressed tactical military communications environments.

We therefore expect that while some future space mission communications will be IP-based, others will use a DTN overlay to augment the IP suite or will run DTN over non-IP underlying protocols. We all agree that the important thing is that NASA should move toward a networked model of space communications.

Vinton G. Cerf & David Israel

The writers are, respectively, with the NASA Jet Propulsion Laboratory, Pasadena, Calif., and NASA Goddard Space Flight Center, Greenbelt, Md.

Safe Hoist

"A Hoist to the Heavens" [August] discusses a space elevator and climbers (elevator cars) to take payload from the surface of Earth to geostationary orbit. There was nothing on the tangential motion of the climber relative to the center of Earth. A point on the surface at the equator moves at about 0.5 kilometer per second toward the east. An object directly above this point at geostationary altitude moves at about 3.1 km/s toward the east. A climber moving up the tether would require a force parallel to the earth's surface (perpendicular to the tether) to change its tangential velocity from 0.5 to 3.1 km/s. The tether is too long to provide this force without excessive distortion. An obvious (but impractical) way to provide the force would be with sideways rocket thrusters on the climber. This needs to be clarified.

William W. Shrader

IEEE Fellow

Stow, Mass.

Author Bradley Carl Edwards responds: Let's look at the numbers: the ribbon weighs about 800 tons, the counterweight 650 tons, and the climber 20 tons. The ribbon has a natural period of oscillation of 7.2 hours, and the climber ascends from Earth to the far end in about 14 days.

As the climber ascends, the ribbon is deflected backward in orbit a small fraction of a degree. This deflection then provides the force needed to accelerate the climber in the orbital direction. Because the mass of the ribbon and counterweight is 70 times that of the climber and the ascent rate is slow compared with the restoring period, the orbital velocity required is easily supplied by the ribbon and the rotating Earth. The ribbon never deflects more than a small part of a degree. With the climbers out of phase, the deflection does not build up and is manageable.

Dreaming

I appreciate Robert Lucky's "college dream" [Reflections, July] and the account of his tech-oriented dreams and nightmares. I had a professor in a college electronics class who never gave out notes or handouts. When we asked for references, he mentioned a dozen books, but never the book he had copied from.

We learned about the book, which was out of print, before the final exam. The four copies in our library had all been borrowed and were unavailable. Dr. Lucky, I still dream about that dreadful college examination.

M.K. Haldar

IEEE Senior Member

Singapore

Robert W. Lucky responds: I've gotten more e-mail about that column than any other I've ever written. Maybe a hundred e-mails from engineers with the identical dream—there's a final exam in some course that you never attended, or you can't find the place where the exam is being given. How strange!

Corrections

In the September issue, Sherry Higgins's name was misspelled on the Contents page. Also, in "Who Killed the Virtual Case File?" the release date of the final Inspector General's audit was 2005, as correctly stated on the timeline, not 2002, as stated on the second page.