Traversal on server in-memory while querying

hi
i want to understand the procedure on Neo4j server in-memory traversal while querying the data.
i found below article

in the Database Management SystemsHigh-level Architecture section it says

  1. The lookup request moves the pointer to the first record in the global index
  2. It then computes in bytes the offset by multiplying the starting node (or other object) byte size or relationship ID byte size by the node store record size (also in bytes) to find the starting node address.
  1. here what is the global index and how it is selected/elected & defined?
  2. the nodes and relationships are stored in form of linked-list not an array then how can it calculate the offset and jump to the required starting node?
    please clearify
    thanks