Computer World –
Earlier this year, a NASA satellite used the Internet to phone home. Hardly an earth-shattering event -- or was it?
The call was made possible by OMNI (Operating Missions as Nodes on the Internet) engineers at NASA's Goddard Space Flight Center, who uploaded standard Internet software to an orbiting UoSAT-12 satellite and then received data via the Web. It worked, which didn't surprise the engineers. What it means, however, is that NASA satellites can have their own IP addresses and send and receive Internet messages and data. That's brand-new for space jockeys.
With NASA's current system, the Consultative Committee for Space Data Systems (CCSDS), transmissions are routed to special boxes and then to a protocol for Level 0 processing. After that, the data goes out.
"It's not that IP is better or worse [as a networking tool]; it's that CCSDS is just not compatible with the rest of the planet," says Ron Parise, a senior scientist at Computer Sciences Corp. (CSC) in El Segundo, Calif.
The glitch, Parise says, is that the CCSDS protocol doesn't have a network layer. Hurdles such as intermittent connectivity and noisy links seemed insurmountable, particularly when every project had customized protocols.
While the scientific world was stymied, private companies, which faced the same problems, forged ahead, solving forward-error correction problems and developing ricochet modems.
The OMNI team says that it's confident that commercial security products will also provide sufficient protection and privacy for NASA projects.
"The bottom line is that adequate measures are available now," says James Rash, OMNI project manager. "International stock markets and financial institutions use the open Internet for trillions of transactions every day. Those security measures are available to us."
As an added precaution, using closed communications channels -- a traditional NASA operation -- remains an option, even with standard IP.
The UoSAT-12, sent up last year by U.K.-based Surrey Satellite Technology Ltd., was chosen because it already uses high-level data-link control framing. Porting an IP stack to one of the spacecraft's onboard processors was a simple task.
The network approach is producing winning scenarios. Scientists are already familiar with the Internet's capabilities; collaborative science missions are possible because IP provides a common link. For example, earth-science mission data that's retrieved from sensors in ocean buoys or balloons can be relayed from the sensors to satellites, which can then be easily accessed by scientists.
Another mission advantage is commercial IP's ability to significantly reduce integration costs. Currently, instrumentation developed in one lab travels to where a spacecraft is being built. Teams of engineers must spend weeks on interphase documents to communicate among sites.
"With both on an Internet protocol, people can stay where they are, hook up to the Internet from a workstation and begin talking," says Rash.
An upcoming project at the University of California, Berkeley, is using this approach.
"We don't fly satellites to build communication infrastructure; we fly satellites to do science," explains Keith Hogie, a CSC senior consulting engineer. "If we spend less on infrastructure, we can do more science."
Because ground systems or end users don't need special communication hardware, a wide range of off-the-shelf hardwaare and software is available. Better yet, somebody else picks up the costs for development, debugging and ongoing maintenance.
"Thirty years ago, [space] communication was special," says Hogie. "Today, NASA doesn't need to invent new things. They can reap the benefit of the money they put into Internet connectivity."
Web Call to Mars
Although the OMNI project was initially geared toward establishing an IP network among LAN-based satellites and balloons, the technique is ideal for more distant missions, including those planned to Mars.
"Future NASA missions will require more networked assets, and the cheapest and quickest way to do this is with standard protocols," explains Rash.
The original test satellite was a minivan filled with the same equipment now on the UoSAT-12 satellite: computers, a power supply, a transmitter, an antenna, movable cameras and a standard Internet router. To gather data, the OMNI team drove the "spacecraft" around Goddard Space Flight Center in Greenbelt, Md., while people at NASA's Glenn Research Center in Cleveland were at a Web site controlling the video camera onboard the van. NASA's Tracking and Data Relay Satellite System sent data to a ground station at White Sands, Ariz.
For a real mission test, the OMNI team went to the Black Sea for the last solar eclipse of the century. In August of last year, the prototype satellite equipment sent live weather data and images via the Web.
Eclipse data went to Goddard and was distributed to mirror systems at centers in California and Florida. The sites also distributed Java applets that connected to the mirrored systems to receive real-time data streams. The applets did the final processing, reformatted the data and displayed it to users.
Standard data-delivery protocols are only part of the package. The team has already demonstrated spacecraft clock synchronization, and in June, it successfully tested standard File Transfer Protocol. Simple Mail Transfer Protocol e-mail tests are next.
"After all," says Hogie, "the spacecraft may want to send a message."