Bash: How to extract portion from the end of a line that's in a variable?

In the context of a bash script, how do I extract the uniqueID from the uniqueID_db.dat part of URLs such as these, where the uniqueID can be anything, for example:

https://cdn.somedomain.com/fetch/uniqueID/uniqueID_db.dat 
https://server123456.eu.somedomain.com/789/storage/uniqueID/uniqueID_db.dat 
https://cdn.somedomain.com/fetch/6234449e1539130b/6234449e1539130b_db.dat 
https://server654321.eu.somedomain.com/0123/storage/afd85b3f9ae5bc9/afd85b3f9ae5bc9_db.dat 

Lines always end in _db.dat and it's the uniqueID before it that I'd like to extract. Any line is in the variable $link. Is this extraction possible with sed or some other tool? If so, how? And can you please explain the workings so I can learn? I imagine something like:

echo "${link}" | sed '...'

Thanks.