MNEMOSYNE

Progetto Distributed System @ Politecnico di Milano - Anno 2019/2020

Stefano Fedeli - Andrea Pozzoli - Lipei Liu

REPLICATED DATA STORE

Implement a replicated key-value store that offers causal consistency.

Requirements

• Implement causal consistency with limited (coordination) overhead.
• New replicas can be added or removed at runtime.
• The store will be implemented in Java using RMI for the server-side part together with some client code written in Python.

Assumptions

• Processes are reliable.
• Channels are point-to-point (no broadcast)
• The same fault model of the Internet (congestions or partition).
• Clients are "sticky": they always interact with the same replica.

USING

• Go to /deliveries and start the LoadBalancer

java -jar LoadBalancer.jar

• Start as many replicas as you want even on different machines on the same LAN

java -jar Replica.jar [LoadBalancer IP] [Local machine IP]

• Start the as many client as you want even on different machines on the same LAN

python clientGUI.py

TECNOLOGIES