Been reading about how approximately half of internet traffic is bots (https://www.imperva.com/blog/bot-traffic-report-2016). Makes me wonder if cloud providers who charge based on usage (like AWS and GCP) charge for bot traffic.
0
votes
1 Answers
0
votes
Do cloud providers charge for bot traffic
The answer is "it depends" on your architecture.
If you are not doing any form of authentication/blocking at your service, then yes all traffic that reaches your service is chargeable.
If you are using an authentication service in-front of your service, for example, Cloud Identity Aware Proxy (IAP), then unauthenticated traffic will not reach your service, so you will not pay for the prohibited traffic.
If you are using a load balancer with Cloud Armor in front, you do not pay for traffic that Cloud Armor blocks.
If you are using Security Groups to only allow traffic from specified CIDR blocks, you do not pay for the blocked traffic.
The key point is if traffic reaches your service's network interface and you respond to that traffic you pay for the egress traffic. Implement options such as authentication, firewalls and security groups to block unwanted traffic. The key is to DROP (network term) unwanted traffic so that there is no response.
There exist both good and bad bots. Some bots want to be a nuisance and run up your bill. Implement policies on your services that do not respond to bots that are just generating traffic. For example, if a bot is hitting pages that return large amounts of data, block them after X requests. Block them if they hit more then X missing pages (404s). The list goes on and one of ideas to limit the problems that bots cause.
