During the last years, two approaches have been developed to overcome these restrictions by placing satellites closer to the earth: Low earth-orbit (LEO) satellites reside typically between 500 and 1500 km above the earth and have very low round-trip delays (10 to 30 ms) while their middle-earth orbit (MEO) counterparts are 5000 to 12000 km high and have round-trip delays of about 100 ms. For global coverage, obviously, the number of required satellites varies with altitude. Examples of commercial LEO projects are Iridium (780 km, 66 satellites) and Globalstar (1400 km , 48 satellites). A typical example of a MEO global satellite system is ICO (10335 km, 10 satellites). Figure 1 gives an overview over characteristic GEO and LEO/MEO satellite configurations.
A property of LEO/MEO satellites grouped into constellations is that they move at a different pace than the earth; it is therefore not only necessary to provide a larger number of ground stations, but the number of satellites in their reach varies with time.
Currently, satellite ground segments are composed of mission control centres and satellite control centres linked to ground stations. Mission control centres aim to manage satellites in terms of payload. They use satellite control centres in order to configure satellites, manage the onboard resources and ensure the orbit control. Finally, to communicate with satellites, control centres use a network of ground stations. The geographical location of these ground stations allows a complete coverage of the satellites orbit. As described above, in the past, typical space systems used massive spacecraft, huge control centres and large ground stations. As depicted in figure 1(a), every station is equipped with its own management platform. In the future, however, constellations of LEO/MEO micro-satellites will be deployed. The control of these constellations will be based on many micro-ground-stations, inter-communicating through a network in order to keep expenses low and time-to-availability short.
This approach to satellite management follows a structure similar to that employed in other distributed systems. However, it poses an extreme example of such a structure, since it requires high system reliability. Thus, resolving the problems related to space systems would, as a by-product, solve the current management problems of other distributed systems with high availability requirements, such as power-line networks, telephony systems, or digital TV/data networks. Unfortunately, very limited experience with the distributed operations management of such ground station networks exists in the space industry. This constitutes a major obstacle for the deployment of micro-satellite constellations. Experience gained from other fields, such as Civil Aviation Air Traffic Control, shows that international co-operation and interoperable distributed management systems are essential to keep investment costs at a reasonable level, and to enable scalability of future systems. This experience also shows that new management paradigms should be carefully designed to address the specific requirements of the target systems.
With a constellation of satellites, there always will be several ground stations simultaneously in visibility as depicted in the lower part of figure 1. Although it is possible to concentrate enough equipment in a set of ground stations to allow each of them to handle its task correctly, a more efficient solution is to dynamically reallocate tracking slots among the ground stations. This way this otherwise redundant capacity will be better utilised. Another issue is an improved visibility pattern when tracking satellites. When a ground station is used to track one or several satellites, its work plan may be prepared by a management system in advance and it is possible to get the work done using scattered yet co-operating ground stations.
To meet these challenges, it is necessary to define up-to-date operation management software for ground stations. Existing systems management frameworks such as HP OpenView, Tivoli TME 10 and CA Unicenter TNG are relevant to such management contexts and may be used to reduce costs. The management requirements of LEO/MEO satellite networks and an approach for representing their topological relationships with available management platforms are described in [6]. However, many non-standard equipments remain to be managed at a lower level, and, at a higher level, space operations management not addressed by the standard management tools must still be performed. Therefore, an important research problem consists of creating a framework of inter-operable components. These components satisfy a market need which is to build rapidly and on-demand an operations management software to handle a network of ground stations. These stations are intended to co-operate via CORBA (see figure 1(b)) in order to manage a constellation of satellites.
To support this effort, application frameworks based on distributed processing standards such as CORBA and performant and easy-to-implement scripting languages like Java present several advantages to lower the costs and shorten the delay of building operations management systems:
The market of such operations management systems may be extended from the space stations operations management to other types of "networked facilities" management such as fixed services transmitting stations, power lines, digital-TV networks or digital data networks. Operations management is also emerging for terrestrial transportation, telecommunications and messaging, and operations scheduling: The control centres are currently equipped with software and tools on an as-is basis, due of the high cost of tailoring operations management centres. The availability of reusable application frameworks may open a new and rich market there.