|
|
@@ -26,25 +26,15 @@ The outbound session is responsible to ensure the hosts pool is populated, eithe
|
|
|
through currently connected nodes or using the seed session.
|
|
|
It performs this algorithm:
|
|
|
|
|
|
-1. Start $N$ slots, and set each slot with `status = ACTIVE`
|
|
|
+1. Start $N$ slots, and a sleeping peer discovery process
|
|
|
|
|
|
Then each slot performs this algorithm:
|
|
|
|
|
|
1. If no addresses matching our filters are in the hosts pool then:
|
|
|
- 1. If there is another slot where `status ≟ DISCOVERY` or `status ≟ SEED` then
|
|
|
- let `status = SLEEP` and wait for a wakeup signal.
|
|
|
- 2. If there are channels opened in `p2p` then let `status = DISCOVERY`
|
|
|
- else skip this step, and let `status = SEED`.
|
|
|
- 1. If `status ≟ DISCOVERY` and no hosts are found then let `status = SEED`.
|
|
|
- 3. In either case when `status ≟ DISCOVERY` or `status = SEED` and we manage to find
|
|
|
- new hosts, then wakeup the other sleeping slots.
|
|
|
- 4. If there are still no hosts found, then let `status = SLEEP`.
|
|
|
-
|
|
|
-The slots are able to communicate to each other through pipes to signal status changes
|
|
|
-such as wakeup requests.
|
|
|
-
|
|
|
-Sleeping slots are woken up periodically by the session. They can be forcefully woken up
|
|
|
-by calling `session.wakeup()`.
|
|
|
+ 1. Wakeup the peer discovery process. This does nothing if peer discovery is already active.
|
|
|
+ 2. Peer discovery tries first 2 times to poll the current network if there are connected nodes,
|
|
|
+ otherwise it will do a seed server sync.
|
|
|
+ 3. Peer discovery then wakes any sleeping slots and goes back to sleep.
|
|
|
|
|
|
## Security
|
|
|
|
