Google disclosed key details on Maglev, the software-based, network load balancer that is used to ensure consistent performance for the Google Compute Engine.
In a blog posting, Google said Maglev load balancing has been handling most of the traffic to Google services since 2008. It runs on the same, ordinary servers that the services themselves use. The system "uses Equal-Cost Multi-Path routing (ECMP) to spread incoming packets across all Maglevs, which then use consistent hashing techniques to forward packets to the correct service backend servers, no matter which Maglev receives a particular packet."
https://cloudplatform.googleblog.com/2016/03/Google-shares-software-network-load-balancer-design-powering-GCP-networking.html
In a blog posting, Google said Maglev load balancing has been handling most of the traffic to Google services since 2008. It runs on the same, ordinary servers that the services themselves use. The system "uses Equal-Cost Multi-Path routing (ECMP) to spread incoming packets across all Maglevs, which then use consistent hashing techniques to forward packets to the correct service backend servers, no matter which Maglev receives a particular packet."
https://cloudplatform.googleblog.com/2016/03/Google-shares-software-network-load-balancer-design-powering-GCP-networking.html