I am reading this thread trying to decide if the ~11Gbps DDOS attack is still underway. Just yesterday Slush said he wasn't coming back without implementing a distributed mining pool (has since changed his mind). A better place to get info may be IRC (I haved been using Bitcoin Watch, but that only covers the largest pools).

Vague details can be found here: Mass DDOS part 2.

MagicalTux's post disclosing the magnitude of the attack.

Slush's Giving up post.

Speculation: I suspect the mainframe mining pool was also hit by this DDOS attack. Most webhosts provide 100Mbps connectivity. A 11Gbps DDOS would saturate such a pipe 110 times over. If the datacenter is relatively small, a 11Gbps DDOS may saturate at least one of the pipes going into the datacenter (assuming 10Gbps or smaller). My own webhost has a "Triple OC-3 Backbone". According to Wikipedia, that works out to 148.608 Mbit/s of payload (x3=445.824Mbps). A 11Gbps DDOS would shutdown my webserver's datacenter 25 times over.

When faced with this situation, the hosting provider probably said "find a new provider." If the hosting provider actually has the capacity, the request was probably to upgrade to a more expensive plan.

What can be done? One possibility is to avoid pools: they are easy targets. Of course, so is every individual bitcoin node (but collectively a hard target). If you like the lower variance of a pool, there is at least one hybrid pool that combines pooling and solo mining,