These are all issues that the newly formed infrastructure team needs to solve (kernels, wpj, etc...). For now, we just need to stop delivering packets to the wrong people, dropping connections and be able to scale > 4k conns. pgm.