Technically, a distributed database/cache at DNS scale, pure push-based invalidation is just not practical.
I wrote a post about DNS invalidation earlier https://blog.the-pans.com/dns/.
Technically, a distributed database/cache at DNS scale, pure push-based invalidation is just not practical.
I wrote a post about DNS invalidation earlier https://blog.the-pans.com/dns/.
I've always used the term "propagate" even though I understand it to be a pull system due to the existence of the TTL. Either way I hold no association between pull/push and "propagate" in my mind either.
Maybe the author and their friend have some bias towards this definition, perhaps due to technical contexts they have most used that word within.
I don't think so? Using the same example, I would say news propagates, and there is no specific topology it spreads through.
2. As with the OP, the content marketing piece referenced fails to discuss the DNS beyond resolver activity, and thereby presents the same painfully half-baked construction of the system.