I was wondering: are there attacks that target a load balancer's choice algorithm?
For example, if an attacker takes control of a load balancer, the traffic could be redirected to a single server and provoke a targeted denial of service.
Do such attacks exist? Are they common? The only thing I found so far talking about something related is this article where a security flaw in BIG-IP load balancers could allow an attacker to alter balancing rules.
There exist "Yo-Yo" attacks where traffic is sent in large quantities for just long enough that the load balancer's autoscaling increases the resource demand and therefore hardware use, then the attacker cuts off that traffic flow, wasting resources until the load balancer ascertains that there is no longer an excessive amount of traffic and scales down, at which point another wave of traffic is sent. This is relatively easy to mitigate, as the pattern should be easy to detect, so I doubt it's used often or for long, but it would be quite effective initially.
