As a term of art, it's a way to distinguish the types of index. What it means, though, is that it's clustering the actual data (as opposed to just the index keys). So, when you have two sets of cluster keys that are "near" each other, their data rows are located "near" each other on the disk. The same cannot be said of non-clustered indexes.
Let's use a simple example:
Let's say I have a table of fictional characters with two columns: CharacterID
and Name
. If there are, say, 1 million rows in the table and the first two are (1, 'Antman')
and (2, 'Zorro')
and the cluster key is CharacterId
, then those two rows will be close to each other.
However, if I also have a row (1000000, 'Algernon')
and I have a non-clustered index on Name
, then despite Antman
and Algernon
being close to each other in the non-clustered index, the physical rows are likely located far apart because 1 and 1000000 are far apart.