Efficient management of cached data in the global alzheimer’s association interactive network

Abstract

A key tenet of the Global Alzheimer’s Association Interactive Network (GAAIN) is to protect the data ownership rights of its members. This prohibits data shared by its federated data repositories from being copied to any GAAIN disk drive and requires all GAAIN server caches to be managed in memory only. Further, the different data repositories collect different attributes for their subjects, and often there are different amounts of data collected for the subjects within the same data repository. We present a relational database design to manage this sparse cached data using elementary bit operators to perform queries and extract results from compact value representations.