|Data Freshness in Leader-Based Replicated Storage
|Amir Behrouzi-Far, Emina Soljanin, Roy D. Yates, Rutgers University, United States
|N.8: Timely Updates
|Networking and Network Coding
|Click here to download the manuscript
|Click here to watch in the Virtual Symposium
|Leader-based data replication improves consistency in highly available distributed storage systems via sequential writes to the “leader” nodes. After a write has been committed by the leaders, “mirror” nodes are written by a multicast mechanism and are only guaranteed to be eventually consistent. With Age of Information (AoI) as the freshness metric, we characterize the number of leaders influences the freshness of the data retrieved by an instantaneous read query. In particular, with a deterministic model for the leader writing time and a probabilistic model for the mirror writing time, we derive the average age of a read query. With exponential distribution for the mirror writing time, we derived a closed-form for the average age. Our numerical results show that, depending on the relative speed of the write operation to the two groups of nodes, there exists an optimal number of leaders which minimizes the average age of the retrieved data. Furthermore, the optimum number of leaders increases as the relative speed of writing on leaders increases.