[menog] IPv4 depletion update
Maher El Zein
mz-ccie at hotmail.com
Tue May 18 20:23:42 GMT 2010
The move to IPv6 should really very well tested before deploying in the Service Provider class networks, from few days I read an article from IPv6tf saying that many UK Service Providers failed on adding IPv6 in their existing network infrastructures because of unseen complications exactly on the security layer.
The best way to deploy IPv6 in Service Provider MPLS infrastructures is to add the IPv6 addresses only on the facing internet/WAN edges with encapsulating IPv4/IPv6 & Ipv6/IPv4 because currently there are really many complexities exactly on the distribution layer and security layer because most of the firewalls and intelligent security systems still not supporting advanced IPv6.
I found these articles interesting to read:
Also, the IPv4 Exhaustion Counter:
maher el zein
From: ahmed at tamkien.com
To: remco.vanmook at gmail.com
Subject: Re: [menog] IPv4 depletion update
Date: Tue, 18 May 2010 15:23:56 +0300
CC: menog at menog.net; IPv6Jordan at ipv6forum.com
Thanks for your feedback. No one is sure yet how long RIPE NCC will have v4 address blocks; it’s a matter of speculation and depends on factors that we have not yet gone thru before (last minute rush, increase in uptake rate due to LTE, etc.). But we can predict the IANA depletion date, hence my message.
For the 4 countries you mention the utilization rate varies from 75% to 92%, with the highest being for the Netherlands. So there will be demand for limited resources within RIPE. Unfortunately there is no way to predict the depletion on a country by country basis (no framework exists for collecting and sharing LIR data for address block utilization as far as I know ), but all indications/predictions I have seen say RIR depletion should be by 2012, and probably as early as the first half.
As for the IPv6 resources discussion at the RIPE meeting, having an address block and connectivity is a step in the right direction. But these are the easier parts of the transition, while enabling the core, edge and access, then providing the v6 service (without the need of a public v4 address), plus porting the applications, needs a lot more work and time.
If the ability to offer IPv6 native access (and coexist with v4 using private v4 addresses) can be measured by the RIR then this will be a more accurate prediction of readiness. A large US based ISP recently mentioned, publicly, that it took them over 4 years to work on v6 transition to get to a stage to offer native IPv6 access trials.
From: remco van mook
Sent: Tuesday, May 18, 2010 10:52 AM
To: Ahmed Abu-Abed
Cc: menog at menog. net ; IPv6jordan
Subject: Re: [menog] IPv4 depletion update
I think you're reading this the wrong way. The amount of allocated IPv4 space per country is not a fixed number, but rather an aggregation of the blocks of space that have so far been allocated to the different ISPs inside of a country. Since we've not run out of IPv4 space, and it's likely that the RIPE NCC will have space for at least 2 more years, that number doesn't really reflect anything else than a very commendable utilization rate of the resources allocated so far. On top of that, particularly the US and Japan numbers are skewed in a huge way because of the way IPv4 addresses were distributed until about 1995 (remember classfull routing?). Comparing numbers with for example France, Germany, Sweden and the Netherlands would give you a far more apples-to-apples comparison.
I fully agree more action is called for when it comes to deploying IPv6 in the Middle East, but the simple graph that describes this was used at the RIPE meeting two weeks ago - it shows the percentage of ISPs per country that have IPv6 resources, and what percentage actually uses those.
Remco van Mook
On Tue, May 18, 2010 at 8:24 AM, Ahmed Abu-Abed <ahmed at tamkien.com> wrote:
Looking at the IPv4 depletion statistics, it looks like some countries in the MENOG region are in a more critical situation.
The latest data show the IANA IPv4 depletion date varying between March 2011 (aggressive prediction, http://ipv4depletion.com/) and August 2011 (conservative, www.ipv6forum.org ) . In other words, we are only 10 to 15 months away from start of depletion.
Geoff Huston, APNIC Researcher and Scientist, publishes IPv4 stats on all countries. A snap shot is shown below, and I only included a few MENOG countries as well as 3 non-MENOG members for comparison. If there is demand for the full Excel spreadsheet I can send it.
Note that Egypt and Jordan are using more than 90% of their allocated IPv4 addresses (I.e. /32s) to date. Jordan, being a RIPE region member, will probably have more difficulty in getting new allocations as RIPE has one of the higher burn rates.
One can only say its time to act on offering IPv6 to subscribers. And this needs significant work on the core, OSS/BSS, international gateway as well as access systems.
# 0. cc code
# 1. allocated /32s
# 2. % V4 pool
# 3. % Allocated /32s
# 4. /32s per user
# 5. /32s per capita
# 6. advertised /32s
# 7. % cc allocated /32s
# 8. % Advertised /32s
# 9. Adv /32s per user
# 10. Adv /32 per capita
# 11. unadvertised /32s
# 16. users
# 17. population
# 20. cc num
# 21. region num
# 23. cc name
IPv4 Resource Allocations as of May 17th 2010
Source: Reformatted to .XLS from http://bgp.potaroo.net/iso3166/v4cc.html
Menog mailing list
Menog at menog.net
The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Menog