Hi Arunima.
I come with some questions regarding these scenarios:
- How do you imagine the initial deployment. What I mean by this is, for example taking into account “Rare Service Discovery in a Large Noisy Network“ scenario, were you describe a set of 500 nodes. Are these nodes connected to an initial bootstrap node? Or they connect between themselves?
Then, specifically for scenarios:
”Rare Service Discovery in a Large Noisy Network”
- You mention 500 nodes: 250 + 150 + 80 + 3 + 40 is a bit more than that. Just doublechecking that we are talking about 523 nodes, since I assume that every described set of nodes is independent from any other group.
- How we set the behavior of a node as “registrar”? Is this not the behavior by default?
- What is considered a popular service and a rare service. Is it correctly to assume that popular is “popular” because it is advertised by 80 nodes?
- What is “normal random peer discovery”? Normal Kad-DHT nodes?
“REGISTER Storm / Retry Explosion“
- Is client mode finished? Where can I see an example of it?
My last question is, are the required things to check exposed by the service-discovery protocol? Either by logs or metrics.
That would be it for now. I think scenarios are fairly doable so far.