As more and more information is being made available in the form of Linked Data and businesses begin to see the advantages of decentralizing their own information storage, federated queries across many different sources at once will become ever more common. While some queried sources may be deemed trustworthy, others might warrant further validation of the results. Also, when decentralizing user data privacy regulations come into play restricting what information can be queried by whom and mandating data minimization. It is paramount that these types of large-scale queries are executed efficiently. For instance, if a company uses the Solid Pods of their customers to store client details they will want to enable a customer service representative to quickly retrieve the information on a specific client using their e-mail or phone number. We want to investigate if for example knowledge on the data shape of information in a queried document, prior query results or indexing could be used to make future queries more efficient in terms of bandwidth usage, number of requests and time-to-first-response. But also what impact these optimizations have in terms of storage and computational cost at the client side and how they affect the overall decentralization of the data.
The end goal is to reduce the performance and security gap between centralized and decentralized information storage, so that decentralization can become a more viable option for businesses and governments to store their mission-critical information.