I am afraid that there is some kind of misunderstanding.
When we talk about hierarchy in OSPF, we talk about the split of the overall network topology into the backbone area (area 0) and other areas (areas whose numbers are non-zero) that are attached to the backbone area. By its design, OSPF has a two level hierarchy - backbone and other areas. There are no more levels of hierarchy in OSPF.
IP addressing and the length of an IP address (32 bits) have nothing to do with the hierarchy depth in OSPF. If you are perhaps referring to the fact that area numbers (area IDs) have 32-bit numbers that look like IP addresses, that is only a matter of convenience since 32-bit numbers are a natural quantity to work with to computers. However, area IDs are just that - 32-bit numbers - and they do not have any further structure, as opposed to IP addresses that have the network prefix part, and the host part.
Please feel welcome to ask further!
Hello Georg, Aneek,
Indeed. I agree with Georg. That paragraph does not talk about OSPF hierarchy, it talks about Internet routing hierarchy where obviously OSPF would not be even running (since OSPF is an IGP protocol). The hierarchy they talk about is just an attempt to "scope" an IP address into parts that would be hierarchically related to each other, for example, something like continent->country->owner->host (just an example). With 4B IPv4 addressing, a natural approach would be to take each of the bytes for one of those hierarchical units, and that's what they talk about. We never adopted this in real life, though - the best compromise between an efficient use of IP addresses and the numbers of hierarchy is achieved by a simple approach of splitting an IP address into a network prefix and a host ID part, and allowing the network prefix to be further subnetted later on.
Attempts to make IP addressing somehow related to geographical or geopolitical reality have come in and out several times; if you check the old and already superseded IPv6 addressing documents, in particular RFCs 2373 and 2374, they talk about structuring the IPv6 address into a whole set of parts (Top-Level Aggregation, Next-Level Aggregation, Site-Level Aggregation). It never gained much popularity. Imposing strict hierarchy into IP addressing does not go well with the paradigm of getting a prefix of an arbitrary length and subnetting it further as we need.
The bottom line is - what the book says in the paragraph you quoted is about IP addressing hierarchy, not about OSPF hierarchy, and it is an old take on how we could impose certain levels of hierarchy onto IP addressing. It's not exactly what we do today, and it is not related to OSPF.