Gossip Observer: New project to monitor the Lightning P2P network

The discussion emphasizes the comparative inefficiency of using message flooding in Lightning Network (LN) implementations, as seen in Erlay's design. It highlights that a combination of limited flooding and set reconciliation could be beneficial for reducing the total propagation delay. However, the use of Minisketch, despite its potential to reduce communication overhead by filling in small omissions efficiently, is criticized due to its superlinear decoding costs which could significantly increase CPU usage when dealing with large sketches. This raises concerns about the scalability and efficiency of such an approach.

An alternative suggestion includes frequent reconciliations to mitigate the decoding cost issue associated with Minisketch. Yet, this method introduces its own challenges, notably increasing communication overheads due to the necessity of overshooting the unknown required amount for effective reconstruction. The inherent efficiency of Erlay’s flooding technique is acknowledged for its ability to alleviate the bulk of the load from the sketch, thereby optimizing the process by allowing the sketch technology to address only minor gaps effectively.

Furthermore, the communication presents the idea of potentially opting for IBLT (Invertible Bloom Lookup Tables) over Minisketch if relying solely on reconciliation. This consideration stems from IBLT’s different cost-benefit dynamics, where despite its higher overheads, it might become more viable if it were to handle all network traffic, especially when complemented with a minimal use of Minisketch to resolve any residual issues in IBLT decodings. This nuanced perspective underscores the need for careful selection of technologies based on the specific requirements and constraints of the LN's operational environment, balancing between overheads and effectiveness in data propagation tasks.