I have one site (UK) that has half a dozen ordinary pages (including home page) which, when presenting products for sale, splits into four frames plus the master page that holds the frames. It's been like this for about ten years and works well.
I have recently seen hits from (apparently) android that hammer each frame repeatedly, several times a second. I would normally put this down to robot activity but indications are it's really a human with a phone - although a probably really stupid human who can't tell when his phone is in overdrive.
The IPs are UK dynamic, from different ISPs, so probably really human, although today I had one that began with a single IP and seemed to split into two, from different ISPs (may have been two coincident visitors). The UA seems valid (although who knows with android's forever-incrementing version number!) and the headers are reasonable.
One of the oddities is: for many of the hits, especially on one of the seldom-used frames, the referer is http[://]www[.]google[.]co[.]uk[/] (my brackets).
I still allow that non-SSL, non-querystring referer because I see a fair number of hits from it - I think (still checking on that)!
It seems it's either a dumb frames-stupid device or a bot. If it's a scraper I need to find some very clever way of taking it out. If it's not a scraper, I need to discover why it's behaving like one.
Has anyone seen android activity like this? Is it even possible to store hundreds of pages on an android? Or could it be sending it direct to cloud?
