Large experiments and high-performance computer models generate many petabytes of data. While Cloud Computing systems may meet the needs for analyzing these petabytes by harnessing the computing power of many distributed computers, the key challenge in effectively utilizing such a distributed system is the data management process, including storage, indexing, searching, accessing, and transferring data. Most analysis tasks perform computations on a subset of a large data records satisfying some user specified constraints on attribute (variable) values. This subsetting procedure is extremely important in that it reduces the network traffic to and from the cloud facilities. However, selected data records often span many different data files, and extracting the values out these files can be time-consuming especially if the number of files is large. This work addresses this challenge of working with a large number of files. We use a set of astronomical data set as an example and use an efficient database indexing technique, called FastBit, to significantly speed up the subsetting and thus optimize network usage. Overall, we aim to provide transparent and highly efficient attribute-based data access to scientists through a web-based Astronomy Data Analysis Portal. We will discuss the system design, and options for managing an extremely large number of files while minimizing network usage and latency.