I am implementing a decentralised data storage system, and I am currently researching the scalability of Libp2p. I would like to know if there are any established performance benchmarks or studies on how well Libp2p scales. In particular:
-
What is the maximum (or practical upper bound) number of peers that can realistically participate in a Libp2p network?
-
Are there known differences in scalability between subsystems, such as PubSub and the DHT (e.g., performance limits, network overhead, message propagation, etc.), in, for instance, local machine setup, local networks and global network?
I’m asking this both from a research perspective and with an interest in potential production deployments, so any references to experiments, case studies, or best practices would be greatly appreciated.