Database architecture

I was looking at how Elgg stored data and it made start wondering.

Usually when your site has posts, comments, etc you would have tables something like (posts, comments), but everything here is stored in two tables.

The two tables are: `entities` and `metadata` and they are structured as so:

Entities:
```
------------------------------------------------------------
| Name | Type |
------------------------------------------------------------
| guid | int(20) |
| type | enum('object', 'user', 'group', 'site') |
| subtype | varchar(252) |
| owner_guid | int(20) |
| container_guid | int(20) |
| access_id | int(11) |
| time_created | int(11) |
| time_updated | int(11) |
| last_action | int(11) |
| enabled | enum('yes', 'no') |
------------------------------------------------------------
```
and the metadata structure is
```
----------------------------------------------
| Name | Type |
----------------------------------------------
| id | int(11) |
| entity_guid | int(20) |
| name | text |
| value | longtext |
| value_type | enum('integer', 'text') |
| owner_guid | int(11) |
| access_id | int(11) |
| time_created | int(11) |
| enabled | enum('yes', 'no') |
----------------------------------------------
```

If properly indexed, is this better at scale too? What are the benefits and drawbacks to this method?