Last modified: 2014-09-02 11:53:30 UTC
All traffic currently goes through a single network node. This is a single point of failure in Labs and also a bottleneck.