Objective 2: Decrease latency for end users by deploying smarter load balancing mechanisms.
We can make significant improvements to user-facing performance metrics and to user-perceived performance delays by updating the way traffic is balanced across the Tor network.
Tor clients use Tor relays in proportion to their measured spare capacity, with the goal of giving high capacity relays more user traffic, since they can support it. This system works to correct imbalances in relay use for the public Tor network. However, it is currently not used to ensure that the bridges and pluggable transports given to our censored users are not overloaded, and it does not yet take into account the additional overhead being devoted to cover traffic for security defenses. It also currently does not respond to any overload conditions other than bandwidth (such as CPU, memory, or connection count limits). The goal of this objective is to fix these limitations.
This objective will have a significant impact on censored users in places like China, where the use of pluggable transports and bridges is required to access the Tor network. In Objective 2.1.2, we will improve the load balancing for selected pluggable transports and bridges so that all users who need to circumvent censorship against Tor have a fast connection. In particular this means that a surge of users (e.g., during a censorship event) can be easily handled through the addition of more bridges. Right now, such an influx of users may cause some bridges to become overloaded and too slow to use, for no reason other than the lack of load balancing while handing them out.
To complete this objective, we will: