Slony-1 "classic" is the traditional method, and uses all the features that Slony-1 is known for. "Classic" is the original intended use of the product. In a "classic" environment, each Postgres instance to be replicated has Slony-1 fully installed and is designated with a node number. Each node communicates with the other nodes as an equal, via a software process called "slon". Origins and subscribers are defined. Slony-1 can be used in a master-slave configuration, but that is not the model up on which it was designed. The Slony-1 philosophy is that any node can be an origin for some data and a subscriber to other data. Likewise, any node can be all origin or all subscriber. Each slon process must be run as a database superuser, or a user with effectively superuser privileges. Otherwise, the slon process cannot execute the updates, inserts and deletes to keep the subscribers in sync. If all nodes are in-house, powerful access rights across Postgres instances is less of a concern. For the JSOC, with nodes distributed worldwide, "classic" represents a security risk for all distributed nodes. Running in "classic" mode is more complex and requires a tight connection between nodes.
Two nodes and a defined cluster are the minimum software items required to run Slony-1. It will not run without a second node, or without a defined cluster.