Cardano SL Network Topology
This is an overview of Cardano SL network topology.
Though Cardano SL is designed and implemented as a distributed network, for purpose of DDoS protection additional facilities were added to Cardano SL implementation and topology. Currently all nodes are divided into 3 groups:
Let’s describe each of these groups.
Core nodes are the most important ones.
As described here, for Byron release we will have stake effectively locked on federation of core nodes. Only these few nodes can be slot leaders, so only these nodes will be able to create new blocks over this period. This is essential for network core nodes to operate well and continue maintaining the blockchain. We also bring additional security level around this set of core nodes: we put them inside a perimeter of relay nodes, so core nodes are isolated from the public internet and only relay nodes are allowed to communicate with core nodes. This is how we reduce probability for them being attacked.
Please note that core nodes never create currency transactions (only edge nodes can do it, see below).
You can think of relay nodes as of proxy between core nodes and public internet.
Since relay nodes are not isolated, they can be attacked, but they are in principle stateless and do not have any stake, so they can be moved, or their number can be increased. If relays are taken down, this does deny service, but the integrity of the core nodes (and the blockchain) should not be compromised.
Since relay nodes do not have any stake, they cannot be slot leaders. Moreover, they cannot create currency transactions as well.
Relay nodes are fully under the control of the federated committee of initial stakeholders.
Edge node is a simple node that anyone can run on their computer. Only these nodes can create currency transactions.
Since edge nodes do not have any stake, they cannot be slot leaders as well as relay nodes. Furthermore, edge nodes cannot directly communicate with core nodes, only with relay nodes and with other edge nodes.