[menog] Asia-Pacific IPv4 address depletion, Middle East next
owend at he.net
Sun Apr 17 16:55:04 GMT 2011
On Apr 16, 2011, at 11:40 AM, Ahmed Abu-Abed wrote:
> A few observations,
> a. RIPE is next, but the date is currently unpredictable as we are in a unique situation where both IANA and APNIC pools are depleted, an unprecedented event, and the effect of this is unknown. It could be as early as June 2011 or as late as Q1 2012.
I would argue that if you follow previous allocation rates, it comes pretty close to July 2011. If it accelerates (as it did in APNIC
region), that could be significantly earlier. I'm not sure what you think about the current unique situation would possibly cause
the consumption rate to slow so much as to move the date out to next year.
Finally, remember, Geoff's rather rosy predictions make no allowance whatsoever for the portability of demand. (The ability
of organizations with presence in multiple regions to obtain addresses from any or all of the applicable registries).
> b. RIPE's pool of available IPv4 addresses is shown below, but note that practical depletion happens once RIPE reaches 16.7 million addresses as this means they hit the last /8 block which has extremely restricted allocation policies (like APNIC):
Right... I'll also note that APNIC's current depletion state is exactly that. This merely reduces the amount of space
that needs to be consumed before referring to RIPE as exhausted from 3.93 to 2.93 /8s worth of space.
> c. IPv6 deployment at the ISP needs to be done a phased approach. First the core & connectivity, then provide access over tunneling to end users, then work on the rest of the network to get it dual-stack end-to-end. In the meantime if IPv4s are in short supply (depends on the ISP) then a parallel project is needed to phase out dependency on public IPv4s consumption by end users.
All of this is certainly a logical approach. However, I think that if ISPs are not already moving aggressively on such an approach, then, it is
likely to get overtaken by events which will, in turn, cause significant disruption.
> d. Migrating from IPv4 to IPv6 and running both IP stacks in parallel, while having an upcoming shortage of IPv4 and a lot of IPv4 content and websites, is not the same as implementing pure IPv6. It is a more complex job which needs careful planning by the ISP.
I don't agree that it is more complex than running pure IPv6 in the current environment. I do agree that it is more complex that running
IPv6 alone in an idealized environment where there are not any important IPv4 only resources remaining, but, unfortunately, the large
number of multilateral dependencies involved in changing from the current situation to that situation makes me think that will take
a substantial amount of time.
Failing to provide IPv4 access services of some form will likely be considered an unacceptable business decision by most ISPs.
However, failing to at least add IPv6 capabilities will also become an untenable business decision fairly soon. The difference is
that IPv4 is a well understood and highly visible requirement now. IPv6 is writing on the walls, but, not everyone reads it the
same way as yet. I postulate that if you wait until you are losing customers for failing to deploy IPv6, you are likely to have suffered
a pretty serious impact to your business by the time you can deploy it.
> Depletion is real and is happening all around us. ISPs need to start v6 migration now, dedicate staff and funding, and implement a phased approach.
We agree on virtually every point here. I think the only place where we have some difference is in the level of urgency and the perspective
on how quickly some of these "phases" may e overtaken by external events.
> From: Lu Heng
> Sent: Saturday, April 16, 2011 6:40 PM
> To: Owen DeLong
> Cc: Ahmed Abu-Abed ; menog at menog. net
> Subject: Re: [menog] Asia-Pacific IPv4 address depletion, Middle East next
> Hi...."we" means RIPE community.http://inetcore.com/project/ipv4ec/index_en.html. It shows RIPE still have 4 /8 at this moment. While it still has second most amount RIRs.
> IPv6 was not that easy, pure IPv6 environment is still not practical at this time, as small ISP, we tested just few days ago with pure IPv6 with few customers, but the thing is, it just not practical in the real world now. Many discussion around how and when IPv4 will become history, most of my college would agree later 2015 as earliest.
> On Sat, Apr 16, 2011 at 4:18 PM, Owen DeLong <owend at he.net> wrote:
> I don't think that is accurate at all.
> 1. As soon as anyone is out and forced to start deploying customers without
> native IPv4 (if said customers have IPv4 at all, it's through LSN/DS-Lite/NAT64)
> you have a situation where the user experience for your customers trying to
> reach those customers is degraded if you aren't providing IPv6.
> 2. It will take most network operators at least 6 months and probably more like
> 18 months to get from starting to deploy IPv6 on their backbones to being
> able to roll it out to the majority of their customers. Arguably if there's 6
> months until you have to have IPv6 to your customers, you needed to start
> 12 months ago just to be on schedule.
> 3. I expect RIPE will be the next RIR to run out. I expect they will run out
> probably around June or July. That's not 6 months and that's where
> most of the middle east gets their addresses.
> 4. I'm not sure what you mean by "we are the last". I'm not familiar enough
> with your network to apply the proper context, so, perhaps in some way
> you may have 6 months before you face it in your own environment, but,
> what about your user's ability to reach other environments and/or the
> ability for users in other environments to have a good experience
> reaching yours?
> 5. The organization who gets the last allocation in each RIR has a slight advantage
> over all the organizations who were in line behind them because they have
> enough IPv4 addresses to meet their needs for some (limited) amount
> of time whereas the others have no supply of addresses available to them.
> Using that advantage as an excuse to delay your IPv6 deployment is,
> IMHO, both short-sighted and self-destructive.
> On Apr 16, 2011, at 6:53 AM, Lu Heng wrote:
>> well, we are the last, we still have another 6 month to go before face it, correct me if I was wrong.
>> On Sat, Apr 16, 2011 at 2:56 PM, Owen DeLong <owend at he.net> wrote:
>> Well said.
>> On Apr 16, 2011, at 2:06 AM, Ahmed Abu-Abed wrote:
>>> Dear colleagues,
>>> APNIC, the IP address registry for the region stretching from Pakistan to Japan and down to Australia/NZ, announced yesterday it has reached the final /8 IPv4 address block, which is practical IPv4 depletion for most ISPs.
>>> From now on APNIC will highly restrict the amounts of IPv4 addresses it issues, with ONLY one block 1024 addresses per ISP, and this will be the last IPv4 address block given to each ISP.
>>> Expect IPv6 only services to start come up, so even if you have enough IPv4 addresses or thinking of implementing an IPv4 Carrier Grade NAT in your network, you will still need to start IPv6 migration all the way to the subscribers.
>>> The Middle East's address registry, RIPE NCC, is expected to reach a similar situation soon when RIPE reaches its final /8.
>>> Best Regards,
>>> Ahmed Abu-Abed, P.Eng.
>>> VP, IPv6 Forum Jordan
>>> GSM +962 777 669 100
>>> Menog mailing list
>>> Menog at menog.net
>> Menog mailing list
>> Menog at menog.net
>> Kind regards.
>> This message is a private communication. It may contain information that is privileged or confidential. If you are not the intended recipient, please do not read, copy, or use it, and do not disclose it to others. Please notify the sender of the delivery error by replying to this message, and then delete it and any attachments from your system. Use of this e-mail is prohibited when received in error. Thank you.
> Kind regards.
> This message is a private communication. It may contain information that is privileged or confidential. If you are not the intended recipient, please do not read, copy, or use it, and do not disclose it to others. Please notify the sender of the delivery error by replying to this message, and then delete it and any attachments from your system. Use of this e-mail is prohibited when received in error. Thank you.
> Menog mailing list
> Menog at menog.net
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Menog