I2P Network Performance: Speed, Connections and Resource Management

The I2P network is fully dynamic. Each client is known to other nodes and tests locally known nodes for reachability and capacity. Only reachable and capable nodes are saved to a local NetDB. During the tunnel building process, the best resources are selected from this pool to build tunnels with. Because testing happens continuously, the pool of nodes changes. Each I2P node knows a different part of the NetDB, meaning that each router has a different set of I2P nodes to be used for tunnels. Even if two routers have the same subset of known nodes, the tests for reachability and capacity will likely show different results, as the other routers could be under load just as one router tests, but be free when the second router tests.

This describes why each I2P node has different nodes to build tunnels. Because every I2P node has a different latency and bandwith, tunnels (which are built via those nodes) have different latency and bandwidth values. And because every I2P node has different tunnels built, no two I2P nodes have the same tunnel sets.

En server/klient är känd som en "destination" och varje destination har åtminstone en ingående och en utgående tunnel. Förvalet är 3 hopp per tunnel. Det blir tillsammans 12 hopp (aka 12 olika I2P-noder) för en komplett tur och retur klient-server-klient.

Varje datapaket sänds genom 6 andra I2Pnoder innan det når servern.

client - hop1 - hop2 - hop3 - hopa1 - hopa2 - hopa3 - server

och på vägen tillbaka 6 andra I2Pnoder:

server - hopb1 - hopb2 - hopb3 - hopc1 - hopc2 - hopc3 - client

Traffic on the network needs an ACK before new data is sent, it needs to wait until an ACK returns from a server: send data, wait for ACK, send more data, wait for ACK. As the RTT (RoundTripTime) adds up from the latency of each individual I2P node and each connection on this roundtrip, it takes usually 1-3 seconds until an ACK comes back to the client. Because of TCP and I2P transport design, a data package has a limited size. Together these conditions set a limit max bandwidth per tunnel of 20-50 kbyte/sec. However, if ONLY ONE hop in the tunnel has only 5 kb/sec bandwidth to spend, the whole tunnel is limited to 5 kb/sec, independent of the latency and other limitations.

Encryption, latency, and how a tunnel is built makes it quite expensive in CPU time to build a tunnel. This is why a destination is only allowed to have a maximum of 6 IN and 6 OUT tunnels to transport data. With a max of 50 kb/sec per tunnel, a destination could use roughly 300 kb/sec traffic combined ( in reality it could be more if shorter tunnels are used with low or no anonymity available). Used tunnels are discarded every 10 minutes and new ones are built. This change of tunnels, and sometimes clients that shutdown or lose their connection to the network will sometimes break tunnels and connections. An example of this can be seen on the IRC2P Network in loss of connection (ping timeout) or on when using eepget.

Med ett begränsat set av destinationer och ett begränsat set av tunnlar per destination, använder en I2P-nod endast ett begränsat set tunnlar över andra I2P-noder. Om, t ex, en I2P-nod är "hopp1" i det lilla exemplet ovan, så ser vi endast 1 deltagande tunnel som utgår från klienten. Om vi summerar hela I2P-nätverket, så byggs bara ett ganska begränsat antal tunnlar med en begränsad bandbredd allt som allt. Om man distribuerar dessa begränsade mängder över antalet I2P-noder, så finns det bara en liten del av den befintliga bandbredden tillgänglig.

To remain anonymous one router should not be used by the whole network for building tunnels. If one router does act as a tunnel router for ALL I2P nodes, it becomes a very real central point of failure as well as a central point to gather IPs and data from clients. This is why the network distributes traffic across nodes in the tunnel building process.

Another consideration for performance is the way I2P handles mesh networking. Each connection hop-hop utilizes one TCP or UDP connection on I2P nodes. With 1000 connections, one sees 1000 TCP connections. That is quite a lot, and some home and small office routers only allow a small number of connections. I2P tries to limit these connections to under 1500 per UDP and per TCP type. This limits the amount of traffic routed across an I2P node as well.

If a node is reachable, and has a bandwidth setting of >128 kbyte/sec shared and is reachable 24/7, it should be used after some time for participating traffic. If it is down in between, the testing of an I2P node done by other nodes will tell them it not reachable. This blocks a node for at least 24 hours on other nodes. So, the other nodes which tested that node as down will not use that node for 24 hours for building tunnels. This is why your traffic is lower after a restart/shutdown of your I2P router for a minimum of 24 hours.

Additionally, other I2P nodes need to know an I2P router to test it for reachability and capacity. This process can be made faster when you interact with the network, for instance by using applications, or visiting I2P sites, which will result in more tunnel building and therefore more activity and reachability for testing by nodes on the network.

Prestandaförbättringar

För möjliga framtida prestandaförbättringar see Framtida Prestandaförbättringar

För tidigare prestandaförbättringar se Prestanda Historiken.