Having played around for a bit, it's a nice toy for the propeller heads but there's no way I'd trust AWS with my business online computing requirements.

And to be honest the server I was 'playing' with was quietly racking up $60 dollars a month not doing very much. I think I'd rather spend a bit more and host my site with a more conventional provider.

Whilst the work done in this area by Amazon is interesting and good on them if it helps their own business processes tick over, but lets face it they're a retailer not a service company.

Paris...because she enjoys a good 'botting now and again.

Amazon could also be blocking bots that harvest content. If your service has hit its performance limits, the first thing you do is shut off everything that's not making money. That would mean blocking online stores that serve up your images, blocking deal-finder web sites that harvest your content and price lists, blocking review-finder web sites that harvest your user comments, and blocking sleazy search engines that never stop crawling. It's a pretty standard practice to buy time during an emergency.

Time will tell what's going on. Amazon can't keep having outages without informing investors.

...using Layer 7 for intercepting our web browsing?

"a single attacker perpetrated a "layer 7" attack on IMDB"

> "a single attacker perpetrated a "layer 7" attack on IMDB"

Thousands of gopher sessions? ^^

Declined for comment?

Sounds like they frakked up....big time.

Yeah a just a few Gopher sessions needed , a few thousand, but if you look at the Narus post you'll see they were flooding port 80, gopher works of port 70..

As to the amount of packets / sessions ...

Take packet size roughly 1,400 bytes MTU size - so that works out at rough 2,143 packets a second.

I'm old enough to remember the "Gopher" protocol.

Paris - because she wishes she was old enough to get herself indexed by one.

@Kevin McMurtrie,

Amazon have special pools of robot servers for search engines and the like. They don't block them as such - they just get a potentially lousy interactive experience which really doesn't matter too much if it is an automated scrape. The IPs get automatically added to a robots redirect list when they get in the top hit list. They may have got more clever since when I learnt this of course...