Adopting the grpc mem package

[jzelinskie]

gRPC now has an package for managing memory more efficiently with the lifetimes of messages, I wonder if it makes sense to leverage this library instead of implementing internal pools: https://pkg.go.dev/google.golang.org/grpc/mem

[mostynb]

This looks interesting, but it's relatively new and marked as experimental. I think it might be worth trying out, but we should wait to see if it gets picked up and used more widely before considering using it.