[ALAC] 45 days remaining supply of IPV4?

Patrick Vande Walle patrick at vande-walle.eu
Tue Jan 4 07:13:18 UTC 2011

 Allow me to add some nuances and comments.

 On Tue, 4 Jan 2011 09:16:35 +0800, James Seng wrote:

> It isn't as bad as most thinks....IANA may be exhausted but the RIRs
> have reserves. And even RIRs runs out, the ISPs have their buffer.
> This will drag out for another 4-5 years at least before the Internet
> users is going to feel an effect of IANA exhaustion

 The common wisdom here is that RIRs will run out of IPv4 blocks 9 to 12 
 months after IANA.
 ISPs might revert to carrier grade NAT (CGN) to extend the life of IPv4 
 a bit, but that itself will create other problems. On the hosting side, 
 one can have hundreds of web sites behind one IPv4 address.

 From the ISPs' point of view, this not necessarily as bad as it sounds. 
 To the contrary, CGN will allow them to better control what their 
 customers get access to. Depending on the local context, it may be an 
 additional value for their business, or the governement's political 

> On Tue, Jan 4, 2011 at 7:04 AM, Olivier MJ Crepin-Leblond wrote:

>> Fingers *will* be pointed at the incompetence of CIOs

 CIOs are not necessarily "incompetent". Their company expects them to 
 deliver a trouble-free service.

 To summarize the result of an assessment made by my employer's IT dept:

 Major problem: HW equipment (load balancers, firewalls, proxy servers, 
 routers) not yet mature. Many local developments need to be updated. 
 Some legacy apps cannot.
 Minor problem: many off-the-shelf software products do not yet support 
 Service provisioning: Our ISPs are only offering "best effort" for IPv6 
 support. No SLAs, risk of lack of service quality

 In summary, CIOs cannot be expected to deliver a second rate Internet 
 access to the company.
 At least in the corporate environment, expect IPv4 to be there for 
 another 10 to 15 years, in parallel with IPv6.

>> We *are* going to hit the wall. It *is* going to hurt.

 Agree. But the reasons may not be those one might think of in first 

 The industry is being slow in delevering tools that allow the 
 operations of IPv6 just as easily as it is now for IPv4.
 I tend to agree with this post on the Gartner blog,as far as companies 
 are concerned:

 For individuals, home users, i.e. the major population served by the 
 At-Large, it might be slightly different. Still, content providers take 
 no risk. Their current offering in v6 is clearly separate from their 
 main v4 offering (eg ipv6.cnn.com), because there is no such thing yet 
 as a global IPv6 Internet. Many peering agreements are still missing, 
 misconfigured client computers are millions, etc. If IPv6 does not work 
 well on the client side, the average user will think that the "web site 
 is slow" and go elsewhere.

 It is not a surprise that Google, for example, only peers in v6 with 
 ISPs who demonstrate excellent connectivity to Google. They just do not 
 want their users to switch to Bing. See 

 On a related note, I am quite disturbed to see now a flow of new RFCs 
 coming out related to IPv6. It just gives the impression to it is now 
 yet ready for mass deployment.

 So, I agree we should continue our efforts to get a larger support for 
 IPv6, but we should at the same time avoid the simplistic view that 
 "everything is ready, we just need to turn it on". It is factually 

 Sorry to be even more dull ...


More information about the ALAC mailing list