Hacker News Reader

This should be voted higher up.

According to https://twitter.com/bgp4_table, we have just exceeded 768k Border Gateway Protocol routing entries, which may be causing some routers to malfunction.

Isn't it weird that it's happening now even though that number was surpassed nearly a month and half ago?

Different locations see different counts because of aggregation/de-aggregation.

Will this affect more than just Google? I haven't seen any outages from other cloud providers.

packet.net was hit. Specifically, also their San Jose DC. Internet only. It took less than an hour to recover. More than 20 minutes. I didn't ping it continuously, and I can say that the traceroute got stuck in Frankfuhrt (Where my ISP and their ISP first (as seen from me) meet).

I was actually surprised, as they tend to have excellent networking. Now I'm not nearly as distrusting as I was initially, knowing it was likely their ISP getting screwed by routing table overflow.

Reply on news.ycombinator.com