To add I think the reason why ntl.com and this site remained working is they are either cached on all the proxies and no lookup is occuring or they have a dns hard value set to avoid lookups meaning they not affected by dns outages.
Common sense would also dictate that ntl added new dns servers for redundancy and had them on different servers and networks so a ddos attack on the main 2 servers wouldnt affect the slave dns servers as such allowing them to keep their service live. They may also want to invest in some Juniper hardware
