Search RPD Archives
[rpd] A typical conversation with a service provider on v6
owen at delong.com
Mon Jun 16 22:41:56 UTC 2014
On Jun 16, 2014, at 04:37 , Seun Ojedeji <seun.ojedeji at gmail.com> wrote:
> Hello Mark,
> On Mon, Jun 16, 2014 at 9:19 AM, Mark Tinka <mark.tinka at seacom.mu> wrote:
> On Monday, June 16, 2014 09:14:09 AM Seun Ojedeji wrote:
> > Yeah "to some extent" it is; because i am just that rare
> > customer that also wants a 128bits even though
> > everything works fine on 32. I doubt end customers will
> > move ISPs as much especially if everything works fine on
> > v4. I think policy and its implementation at regulator's
> > level could expedite some of v6 visibility.
> Customers will go to where they can get service.
> If a customer is on IPv4 today, and it works, short of any other issues, they won't be looking to move.
> If an Internet resource is only on IPv6, and the customer's existing ISP only supports IPv4, the customer will experience connectivity issues and will, invariably, start shopping around unless their existing provider turns up IPv6.
> +1 but realistically speaking is there anything like v6 only resource at the moment?
I really need to put some more meaningful content there, but it's a start.
> If new Internet users are signing up to an ISP that is operating in an era where there is no longer any IPv4, the customer still doesn't care what protocol his services are running over provided there is end-to-end connectivity. Again, the customer's ISP (or their competition), will need to provide a solution to the customer that satisfies their connectivity needs (be they native to IPv6 resources, or translated to IPv4-only resources).
> Ultimately, the customer will go to where they can get
> service. That's the bottom line.
> Sure...this is a final future scenario. Just that at the moment we have few v6 champions. Instead folks are aiming for v4 and since they are aiming for v4 it also mean that v4 will stay longer that we may have presumed.
I don't think this is too much of a concern for the following reasons:
1. Continued growth in the IPv4 customer space is going to drive ISP costs in the residential markets in most of the developed world to a
point where those providers can no longer support IPv4 without increasing rates. These rate payers tend to be very price sensitive, so
there is strong disincentive here.
2. Continued fragmentation of the IPv4 address space as the address transfer market gains momentum in ARIN/RIPE/APNIC regions
will drive routing table growth which is already on the brink of breaking several core routers in active use today. This cost of sustaining
IPv4 will force changes which will lead to fragmentation and further degradation of the service that can be offered over IPv4.
I realize a lot of network engineers and network managers don't get it. That's really unfortunate. However, the simple facts on the ground are that IPv4 has been in critical condition on life support with multiple organ failure for more than 20 years. It's on the verge of the networking equivalent of a condition known as DIC (Diverse Intravascular Coagulopathy) at which point, it will be virtually impossible to keep it meaningfully running.
Like it or not, we are now in a race to get IPv6 deployed before IPv4 collapses under any one of several scaling limits all of which exacerbate each other. The IPv4 patient was GORK'd years ago. Now it's an organ donor. The sooner we face that reality, the better off we will all be.
GORK -- God Only Really Knows, a crass, but common term in the medical profession for a comatose patient on persistent life support with a low probability of any meaningful recovery.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the RPD