I am surprised and disappointed at the somewhat unhelpful discussion about 
BCP38.

Can BCP38 be an effective defence against the DDoS attacks? Yes, of course. 
Will it be? No, at least not  always. And almost certainly not in our lifetimes.

It is hard for any reasonable (ie multi-homed) network to implement BCP38 
unless they do so from the outset. It's also an expense/overhead which doesn't 
benefit the network operator but benefits everyone else. So the economic 
incentives simply are not there for ISPs to deploy this: why incur those costs 
so that your competitors benefit? This is why BCP38 has not been uniformly 
adopted since it was published ~15 years ago. That situation isn't going to 
change until the Internet Police come along and *make* everyone use it. In 
other words, it's never going to happen. Get over it.

Clearly BCP38 has a role in defeating these attacks. But it's not the only 
tool. IMO it would be better for the people here to discuss other defences: 
response rate limiting, traffic shaping, kernel filters, improvements to ntpd, 
share operational experiences, etc. Encouraging wider uptake of BCP38 is of 
course a very good thing. But please be realistic about what can be achieved 
and when. Assuming this is the magic bullet which will fix this problem forever 
is at best naive. It may even focus efforts away from other areas where 
achievable improvements could be made.

Please note too that critical DNS servers have been subjected to similar 
attacks for a couple of years now and this has had very little impact on the 
uptake of BCP38. If the operators of root and TLD name servers have been unable 
to ratchet up the use of BCP38 or put pressure on IXPs, what hope is there for 
contributors to the NTP pool (or other NTP operators)?

_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool

Reply via email to