caching html pages - which is better: blobs or files?

I've written a simple package for caching pages I've fetched from the
web. It works like a champ, and makes it so I don't have to hit a
server multiple times.

I'm wondering whether it's better to store the cached pages in my db
(I'm using mysql) or in the file system. Either approach is
straightforward, but do people have experiences or even strongly held
beliefs as to which is better?

The floor is open...

- ff

I've written a simple package for caching pages I've fetched from the
web. It works like a champ, and makes it so I don't have to hit a
server multiple times.

I'm wondering whether it's better to store the cached pages in my db
(I'm using mysql) or in the file system. Either approach is
straightforward, but do people have experiences or even strongly held
beliefs as to which is better?

The moment you want to access those cached pages from another server, you'll want them in the database. Similarly, if you want to add any meta data.

But until then... I'd be inclined to just dump them to disk.

Just write your library so that the code doing the processing doesn't care where it is... then it will be easy to switch later.