Let's ChronoSync: Decentralized dataset state synchronization in Named Data Networking

In supporting many distributed applications, such as group text messaging, file sharing, and joint editing, a basic requirement is the efficient and robust synchronization of knowledge about the dataset such as text messages, changes to the shared folder, or document edits. We propose ChronoSync protocol, which exploits the features of the Named Data Networking architecture to efficiently synchronize the state of a dataset among a distributed group of users. Using appropriate naming rules, ChronoSync summarizes the state of a dataset in a condensed cryptographic digest form and exchange it among the distributed parties. Differences of the dataset can be inferred from the digests and disseminated efficiently to all parties. With the complete and up-to-date knowledge of the dataset changes, applications can decide whether or when to fetch which pieces of the data. We implemented ChronoSync as a C++ library and developed two distributed application prototypes based on it. We show through simulations that ChronoSync is effective and efficient in synchronization dataset state, and is robust against packet losses and network partitions.