A secondary index structure in distributed data table

We described a secondary index structure which can be used in distributed data table in a system that horizontally range-partitions data over a large cluster of shared-nothing machines. In distributed data systems like Google Bigtable and Yahoo! PNUTS, data can be indexed only by the key which dominates the physical store order, but the secondary index is necessary in such systems. We implemented a system which uses the ordered table to store data and contains the secondary index structure. A distributed B+ tree is established among several index servers so that clients can fetch the data from any index server using the attribute on which the secondary index has been created. 1553-9105/ Copyright © 2009 Binary Information Press.