status-im
/
nim-eth
mirror of https://github.com/status-im/nim-eth.git
2
0
Fork 0
Code Issues Projects Releases Wiki Activity

Update discv5.md

This commit is contained in:
Kim De Mey 2021-01-14 09:26:19 +01:00 committed by GitHub
parent e25500c9f0
commit ab969f72c1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 17 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
doc/discv5.md
View File

@ -40,26 +40,32 @@ d = newProtocol(privKey, ddb, ip, tcpPort, udpPort,
d.open() # Start listening and add bootstrap nodes to the routing table. d.open() # Start listening and add bootstrap nodes to the routing table.
``` ```
Now there are two ways to run the protocol. Next there are two ways to run the protocol.
One can call `d.start()` and two loops will be started: One can call `d.start()` and two loops will be started:
1. Random lookup loop 1. Refresh loop
2. Revalidation loop 2. Revalidation loop
The first loop will at specific interval do a lookup with a random `NodeId`. The first loop will at specific interval do a query with a random `NodeId` if no
This lookup will add discovered nodes to the routing table. manual queries were done for more than that interval period.
This query will add discovered nodes to the routing table.
The second loop will at random ranged interval send a ping to the least recently The second loop will at random ranged interval send a ping to the least recently
seen node in a random bucket. This is to keep the routing table cleared of seen node in a random bucket. This is to keep the routing table cleared of
unreachable/dead nodes. unreachable/dead nodes.
Or, one can decide to do this manually within its application by using the Now within the application, manual queries or lookups can be done, for which
available calls: the discovered nodes can be used. Nodes discovered during this process will be
- `lookupRandom` and `lookup` for discovering more peers. attempted to be added to the routing table. One can use the `query`, `queryRandom`
- `revalidateNode` or directly `ping` for revalidating nodes. or `lookup` calls for this. `randomNodes` can also be used to find nodes,
but this will only look into the current routing table and not actively
search for nodes on the network.
In the future, the random lookup loop might be altered to only run in case no
lookup was done in the last x minutes. This way `d.start()` could still be run
while maintaining control over the lookups. Or, one can decide not to run `d.start()` and do this manually within its
application by using the available calls:
- `query`, `queryRandom` or `lookup` for discovering more peers.
- `revalidateNode` or directly `ping` for revalidating nodes.
Of course, in either scenario, lookups can still be done for actually finding a Of course, in either scenario, lookups can still be done for actually finding a
specific node. There is a `resolve` call that can help with this, it will first specific node. There is a `resolve` call that can help with this, it will first