Partial Answers for Unavailable Data Sources

Overview & Synopsis

Many heterogeneous database system products and prototypes exist today; they will soon be deployed in a wide variety of environments. All existing systems suffer from an {\em Achilles' heel}: if some sources are unavailable when accessed, these systems either silently ignore them or generate an error, i.e. they {\em ungraciously fail}. This behavior is improper in environments where there is a non-negligible probability that data sources cannot be accessed (e.g., Internet). In this paper, we propose a novel approach to this issue where, in presence of unavailable data sources, the answer to a query is a {\em partial answer}. A partial answer is itself a query that results from theo partial evaluation of the original query; it is composed of the data that have been obtained and processed during the evaluation and of a representation of the unfinished work to be done. Partial answers can be resubmitted to the system in order to obtain the final answer to the original query, or another partial answer. Additionally, the application program can extract information from a partial answer through the use of a secondary query. This secondary query is called a {\em parachute query}. In this paper we give a taxonomy of partial answers and parachute queries. We present algorithms for the evaluation of queries in presence of unavailable data sources, and we describe an implementation.