This page was last updated in May 2018.
Are there any other ways to mitigate network brute-forcing via significant
Is there any way to detect, flag and potentially remove 'bad floodfills'
similar to the Tor relay authorities, without actually needing to rely on a
form of central authority?
How could packet retransmission strategies and timeouts be improved?
Is there a way for I2P to obfuscate packets and reduce traffic analysis
more efficiently than other proposed ideas? (Padding transport layer,
developing mimic protocols, etc.)
Tunnels and Destinations
Is there a way that I2P could perform peer selection more efficiently or
Would it negatively impact anonymity to use geoip in order to prioritise
physically nearby peers for tunnel building? The primary goal would be to
increase tunnel success and reduce breakage.
- What are the benefits of unidirectional tunnels over bidirectional tunnels? What are the tradeoffs?
- More details are available here.
- How effective is multihoming at load-balancing?
- How does it scale? What happens as more routers host the same Destination?
- The benefit for anonymity is less correlation of router uptime to Destination uptime. Are there tradeoffs?
- How much is the effectiveness of timing attacks reduced by fragmentation and mixing of messages?
- What mixing strategies could I2P benefit from?
- How can high-latency techniques (e.g. message-dependent routing delays) be effectively employed within or alongside our low-latency network?
How significantly does browser fingerprinting impact the anonymity of I2P
users? Would developing a browser package be beneficial for the average
What is the overall impact on the network created by 'greedy users' (users
who take bandwidth from the network without contributing any back
themselves)? Would additional steps for encouraging bandwidth
participation be valuable?