Data management systems for "big science" often have tight memory and disk space constraints. In this paper, we introduce adaptive bitmap indexes, which conform to both space limits while dynamically adapting to the query load and offering excellent performance. So that adaptive bitmap indexes can use optimal bin boundaries, we show how to improve the scalability of optimal binning algorithms so that they can be used with real- world workloads. As the removal of false positives is the largest component of lookup time for a small-footprint bitmap index, we propose a novel way to materialize and drop auxiliary projection indexes, to eliminate the need to visit the data store to check for false positives. Our experiments with real-world data and queries show that adaptive bitmap indexes offer approximately 100- 300% performance improvement (compared to standard binned bitmap indexes) at a cost of 5 MB of dedicated memory, under disk storage constraints that would cripple other indexes.

The unique characteristics of scientific data and queries cause traditional indexing techniques to perform poorly on scientific workloads, occupy excessive space, or both. Refinements of bitmap indexes have been proposed previously as a solution to this problem. In this article, we describe the difficulties we encountered in deploying bitmap indexes with scientific data and queries from two real-world domains. In particular, previously proposed methods of binning, encoding, and compressing bitmap vectors either were quite slow for processing the large-range query conditions our scientists used, or required excessive storage space. Nor could the indexes easily be built or used on parallel platforms. In this article, we show how to solve these problems through the use of multi-resolution, parallelizable bitmap indexes, which support a fine-grained trade-off between storage requirements and query performance. Our experiments with large data sets from two scientific domains show that multi-resolution, parallelizable bitmap indexes occupy an acceptable amount of storage while improving range query performance by roughly a factor of 10, compared to a single-resolution bitmap index of reasonable size.

The data used by today's scientific applications are often very high in dimensionality and staggering in size. These characteristics necessitate the use of a good multidimensional indexing strategy to provide efficient access to the data. Researchers have previously proposed the use of bitmap indexes for high-dimension scientific data as a way of overcoming the drawbacks of traditional multidimensional indexes such as R-trees and KD-trees, which are bulky and whose performance does not scale well as the number of dimensions increases. However, the techniques proposed in previous work on bitmap indexes are not sufficient to address all problems that arise in practice. In experiments with real datasets, we experienced problems with index size and query performance. To overcome these shortcomings, we propose the use of adaptive, multilevel, multi-resolution bitmap indexes, and evaluate their performance in two scientific domains. Our preliminary experiments with a parallel query processor and index creator also show that it is very easy to parallelize a bitmap index

Scientific simulations are often very I/O intensive, requiring high I/O bandwidth to store the data generated by the simulation. Traditional supercomputers have specialized I/O systems with multiple I/O nodes and specialized interconnects to handle such high I/O loads. However, with the increased availability of inexpensive clusters of workstations, more and more simulations are now run on clusters. Unfortunately, cluster supercomputers are usually not very well equipped for I/O, making I/O a serious bottleneck for such applications. To address this problem, we propose log-based I/O (LBIO), an approach that can substantially increase the I/O performance of simulations on clusters by utilizing free space on the cluster's local disks to stage data on its way to remote storage. LBIO uses local disks to create a log of all I/O calls, and uses a background thread to replay the log at the rate that best utilizes the server and network resources. LBIO is implemented as an easy-to-use, non-intrusive library - a user can turn on LBIO by adding a single initialization call to the simulation code. LBIO also works with existing scientific I/O libraries like HDF, as well as collective libraries like ROMIO. Our performance studies on microbenchmarks and a real-world scientific simulation code show that LBIO can provide upto 35% improvement in I/O performance for raw I/O and over 50% for I/O through libraries like ROMIO or HDF