As increasing numbers of major astronomical observatories are being established in very remote sites, in search of dark, clear, and stable skies, the time and money requirements for astronomical observing are rising. With the increasing availability of wide-area networks, it has become clear that this problem can be at least partially alleviated through the use of remote observing techniques. The ability to locally access a distant telescope provides a number of benefits, including reduced health risks associated with high altitudes, an increased degree of collaboration between researchers and students, and easy access for remote trouble-shooting and software development.
In order to demonstrate the viability of such an approach in the context of a major observatory (i.e., Keck), we have developed a high-speed network with sufficient bandwidth for transferring large astronomical images. A software environment identical to that used at the telescope has also been installed, providing the user with a familiar interface to the telescope and its instruments. Our experiments indicate that the remote observing paradigm is now sufficiently understood that it should be seriously considered by any observatory that is situated far from its user base.
As a technical demonstration project for the NASA Advanced
Communications Technology Satellite (ACTS), we have implemented remote
observing on the 10-meter Keck II telescope on Mauna Kea in Hawaii
from the California Institute of Technology campus in Pasadena. The
data connection consists of ATM networks in Hawaii and California,
running at OC-1 speeds (45 Mbit/sec) through optical fiber, and high
data rate (HDR) satellite antennae at the Jet Propulsion Laboratory
(JPL) in Pasadena and Tripler Army Medical Center (TAMC) in Honolulu.
The ACTS satellite, built by NASA and launched into orbit in 1993 to
explore new modes of high speed transmission for digital data, was
used to bridge the Pacific Ocean. Although ACTS represents a
prototype system and is highly susceptible to rain fade, the available
bandwidth of OC-3 (155 Mbit/sec) easily met our requirements for
remote observing. The ground networks in California and Hawaii were
installed and operated under contracts with Pacific Bell and GTE
Hawaiian Telephone, respectively.
On top of this ATM network, we have run Classical IP and the popular TCP/IP and UDP/IP protocol suites. This has enabled the use of the common network tools, such as ftp and telnet, as well as the X11-based observing software for the telescope. With the exception of the TCP-LFN Unix software extensions for high-bandwidth satellite networks, which were purchased from Sun Consulting, the implementation of the software environment was accomplished almost entirely with existing tools and software available via the Internet. The primary applications - the telescope and instrument control software - were used "as-is". Since our intention was to demonstrate remote operation of the Keck telescopes with a software environment identical to that actually used in Hawaii, alterations to the software were discouraged, in fact. A number of small operational modifications were made to the software to increase throughput and usability over the satellite network, as well as to minimize side-effects that might impact actual on-site observers at the telescope. A small number of applications were written and/or downloaded from the Internet to facilitate audio communication over the network, simplify data transfer, and test network bandwidth.
The implementation and operation of the Keck/ACTS remote observing system has led to a number of conclusions, many of which are applicable to ground-based remote observing efforts and non-astronomical communications satellite applications in general:
Remote astronomical observing techniques have the potential to save appreciable expenditures in terms of money and time, while simultaneously enabling increased levels of collaboration in the observing process. In the case of an observatory with large numbers of observers, short observing runs, and/or a very remote site, these savings may very well outweigh initial network costs to enable remote observing. This conclusion may be pertinent to any field in which travel for access to facilities is common, depending on the complexity and portability of the application.
Remote operation of a complex software/hardware system depends highly on the portability and modularity of the system. In the case of the Keck telescopes, the telescope and instrument control systems were originally designed to isolate the telescope hardware from the observing control computers. This feature has allowed us to integrate the hardware aspects of remote observing in a straightforward fashion, by simply extending a single Ethernet component of the network.
While it is beneficial (and easier!) to maintain an identical software interface for both local and remote observers, additional tools are needed to create a collaborative environment among the remotely-observing astronomers and the on-site telescope staff. Fortunately, such tools are becoming widely available with the expansion and increasing popularity of the Internet.
At the current time, high-speed terrestrial networks are the most viable source for adequate bandwidth to enable true remote observing. While the ACTS system is not sufficiently robust to enable remote observing, this testbed project suggests that future commercial-grade communications satellites may provide the reliability and affordability necessary for high-bandwidth remote software applications.
The most outstanding technical problem regarding the viability of geosynchronous communications satellites for Internet-based software applications concerns the performance of the standard TCP/IP protocol over high-bandwidth, long-delay time networks. Although the initial set of TCP extensions provides some relief, and several groups (e.g., Mitre Corporation) are working on this problem, its solution may ultimately determine the role for satellite communications in the WAN market.
This project represents a significant accomplishment in the fields of both astronomy and high-speed networking:
Although remote use of astronomical telescopes has been a topic of interest for many years, particularly among radio and space-based observatories, only very recently are ground-based optical observatories beginning to realize the benefits. This project represents one of the first attempts to develop a truly remote observing program for a major observatory. The Keck Observatory consists of two 10-meter telescopes, the largest optical telescopes in the world, which are managed by a staff of over 100 people for research by several hundred astronomers in California and across the country. Although temporary in nature, this project has introduced the concept of remote observing to the Keck community, thereby stimulating what will surely be an increasing role in future observatory operations.
The field of high-speed satellite data communications is still in its infancy, particularly with respect to extending existing terrestrial networks. Remote observing with the Keck telescopes involves a high-data-rate application which is only now becoming possible with terrestrial networks. This project has therefore been an ideal real-world application with which to experiment with high-speed satellite networking. We have discovered that while satellite communications involves a number of advantages over standard ground-based networking solutions, such as rapid deployment and potentially very high bandwidth, a number of problems remain to be solved, such as the operation of TCP-based applications over long-delay time connections.
Due to the temporary nature of the ground network and satellite arrangements, this project will not exist in its current state beyond September 1997. However, work has already begun to establish a dedicated ground-based network over which remote observing with the Keck telescopes can take place. Software applications are currently being developed for use over this network. The lessons which have been learned from this project have been, and will continue to be very important as we establish the new network and remote observing environment.
Larry Bergman
NASA/Jet Propulsion Laboratory
larry@jplopto.jpl.nasa.gov
(818) 306-6173
For more information, see the paper entitled,
Remote Observing
with the Keck Telescope from California Using NASA's ACTS Satellite