13. Conclusions

13. Conclusions

Given the current efforts to implement AQM, we believe that now is an appropriate time for the deployment of congestion avoidance mechanisms that do not rely solely on packet dropping. As the deployment increases of applications and transports that are sensitive to delay and to individual packet losses (e.g., real-time traffic, short web transfers), relying on packet drops as the normal congestion notification mechanism appears insufficient (or at least non-optimal).

We have examined the consequences of modification of the ECN field within the network, analyzing all opportunities for an adversary to change the ECN field. In many cases, changes to the ECN field are no worse than dropping the packet. However, we note that some changes have the more serious consequence of subverting end-to-end congestion control. However, we note that even then, the potential damage is limited and is similar to the threat posed by end-systems deliberately not cooperating with end-to-end congestion control.