The ISP Column
A column on things Internet
January 2017
Geoff Huston
Addressing 2016
Time for another annual roundup from the world of IP addresses. Let's see
what has changed in the past 12 months in addressing the Internet, and
look at how IP address allocation information can inform us of the
changing nature of the network itself.
The process of exhausting the remaining pools of IPv4 addresses in each
of the Regional Internet Registries (RIRs) is continuing, yet, despite
this hiatus in the supply of IPv4 addresses, the Internet continues to
grow. The reasons for this growth are changing, however. Personal
computers have going the way of mainframes, and are no longer the stable
of the consumer IT market. It looks as if laptops are now also feeling
the pinch as the mainstream computer industry continues to focus on the
smart device market. But even in the personal device sector we see a
market that is no longer expanding at massive rates. In January 2017
Gartner released its estimate of the worldwide device sales for PCs for
2016. Shipments are estimated to total some 270 million units for 2016,
representing an 6% decline from 2015. "Peak" personal computers have now
come and gone, and these personal devices are following their mainframe
predecessors off to silicon heaven! Smartphone sales are more than four
times greater than PCs in terms of volumes, selling an estimated 1.8
billion units for the year, but there is evidence that this market is
saturating and projections for sales volumes in the coming years show no
signs of further growth. The total volume of deployed phones, tablets
and PCs sits at some 7 billion devices and further sales appear to be
dominated by replacement units rather than expansion into greenfield
markets. We appear to be getting very close to a current ceiling for
these human-driven devices.
However, the silicon industry is far too rapacious to be stalled by such
a mundane consideration as market saturation. Today we are seeing
computers disappear as recognisable computers and reappear cloaked in
some other functional wrapper. While the smart watch has not proved to
be a game changer, there are a myriad of other clever gadgets. We're
seeing the consumer offerings with internet-based home lighting systems
and other forms of household automation that involve sensors and
appliance management, such as energy management, irrigation management
and similar. Today's opportunity encompasses all these embedded devices
that collectively have been labelled "The Internet of Things".
However, the exuberant optimism of 12 months ago when predictions of
between 50 and 100 billion connected devices by 2020 has been tempered.
Gartner now project that this world of chattering silicon will get to 20
billion devices by 2020. But there is much uncertainty in these numbers.
Other indicators point to a highly capable silicon production line where
more than 20 billion microprocessors and 10 billion RFID tags were made
in 2016 alone. But it's challenging to place these numbers into a solid
analytical predictive framework, so there is a considerable uncertainty
with these numbers when we try to look forward over the next three to
five years.
What does all this mean for the Internet?
Obviously, the device population of the Internet continues to grow but
it appears that most of the growth of the network is occurring behind
various forms of IPv4 Network Address Translators (NATs). These devices
are then largely invisible to the public network, so efforts to track
their population are challenging. The deployment of these devices behind
NATS places very little in the way of pressures on address consumption.
While the Internet may have absorbed in 2016 a production quantity of
some 270 million personal computers, 1.8 billion smart phones and a
further 1.8 billion connected devices, that does not mean that there has
been a demand for some 4 billion additional IP addresses. Part of this
volume has replaced older equipment, and almost all these additional
devices find themselves positioned behind NATs, making only minor
demands on the overall address structure. The total drain on the
remaining unallocated IPv4 address pool was just 22 million addresses
for 2016.
This was the issue that IPv6 was primarily intended to solve. The
copious volumes of address space were intended to allow us to uniquely
assign a public IPv6 address to every such device, no matter how small,
or in what volume they might be deployed. Why this has not happened so
far, and why we are still concentrating a significant proportion of our
efforts on stretching IPv4 to encompass ever larger population of
attached devices is a critical question.
Answers to this question are not simple. There is the lack of backward
compatibility for IPv6, in that an IPv6 device can only communicate with
other IPv6 devices, which excludes the IPv4 world. There is the
networked effect where the maximal value for a connection is attained
when it can reach all connected services. There is the issue of sunk
cost for network operators, where the cost of the deployment of the new
protocol does not generate any economies in the costs of network
operation. There is also the observation that human use devices
represent higher value and therefore greater market power than low cost
and intrinsically low value devices even if their volumes are highly
disparate. The result is a protracted gradual transition in the
Internet, where network operators and service providers are adding IPv6
to their services on a piecemeal, and some more inpatient folk might say
even lackadaisical, basis. Today's Internet is close to 100% IPv4 and
around 10% IPv6. Or, in other words, around 10% of the Internet's end
user population use devices that have both protocol stacks provisioned
within the device, and have coupled that with a dual stack network
service. All the remainder is IPv4.
Given this level of relative deployment, it is not surprising that we
have still not seen any major service provider or service move one step
further on from a dual stack environment and discard IPv4 support
completely. It remains the case that the forays into the dual stack
world continue to enlarge the pool of IPv6 capable users and devices,
but at this stage this IPv6-capable user population still does not
represent a viable market in isolation from the IPv4 Internet. Services
providers across the Internet are still required to support IPv4 using
whatever approaches they have available, so it should be unsurprising to
observe the almost universal use of IPv4 NATs as a means of stretching
the limited pool of available public IPv4 addresses across ever larger
pools of connected devices.
But it's not that hopeless. We don't have to await near universal
adoption of dual stack services before we can reduce the pressure on
IPv4 and NATs. Most popular browsers, and many other dual-stack aware
applications, are willing to test if a network transaction is viable
using IPv6 before opening an IPv4 session, and as the extent of IPv6
availability increases we can expect to see increasing numbers of
transactions over IPv6, and this may relieve some of the pressures on
the NAT space.
IPv4 in 2016
The process of exhausting the remaining pools of unallocated IPv4
addresses is proving to be as protracted as the process of the
transition to IPv6.
The allocation of 22 million addresses in 2016 on top of a base of 3,558
million addresses that are already allocated at the start of the year
represents a growth rate of 1.1% for the year for the allocated public
address pool. This is less that one tenth of the growth in 2010 (the
last full year before the onset of IPv4 address exhaustion).
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Allocated
Addresses (Millions) 203.9 203.3 189.4 248.8 201.0 114.9 65.1 63.9 34.8 22.2
Total Volume
(Billions) 2.32 2.52 2.72 2.90 3.14 3.34 3.43 3.50 3.59 3.62
Relative Growth 8.4% 7.9% 6.6% 8.3% 6.4% 2.9% 1.9% 1.8% 1.0% 0.6%
Table 1 - IPv4 Allocated addresses by year
The record of address allocations per RIR over the past 10 years is
shown in Table 2.
RIR 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
APNIC 69.6 87.8 86.9 120.2 105.2 1.0 1.3 3.7 4.1 3.8
RIPE NCC 60.7 44.0 43.4 56.0 43.1 40.0 2.0 2.5 3.3 3.4
ARIN 53.0 57.1 41.1 45.2 23.5 45.0 26.5 26.0 8.6 1.6
LACNIC 14.2 12.0 10.5 13.0 24.4 21.0 28.5 19.1 1.8 1.6
AFRINIC 5.5 1.6 5.9 8.5 9.2 7.9 6.8 12.5 16.9 11.8
Table 2 - IPv4 Allocated addresses (millions) - Distribution by RIR
In terms of the IPv4 Internet there is a considerable diversity in the
situation in each region. As of the end of 2016, AFRINIC was the last
remaining Regional Internet Registry (RIR) with remaining IPv4 addresses
available for general allocation, with slightly over 20 million
addresses left in its address pool. APNIC and the RIPE NCC have both
adopted "Last /8" policies, where each applicant can receive just a
single allocation of up to 1,024 addresses from their respective last /8
address pools. ARIN and LACNIC both have much smaller residual IPv4
address pools, and to all intents and purposes they have both run out of
any addresses.
The current position of each RIR, and the projections of IPv4 address
consumption in the coming years is shown in Figure 1.
Figure 1 – IPv4 Exhaustion – from
In 2016 APNIC allocated 3.8 million IPv4 addresses. APNIC effectively
exhausted its general use pool of addresses in April 2011, and since
then the registry has been operating under the terms of a "Last /8"
policy that limits each entity to at most 1,024 addresses drawn from
this residual address pool.
APNIC recorded some 4,325 individual address allocations in 2013. Of
these, 3,019 entries refer to allocations from APNIC's last /8 address
block (103.0.0.0/8), and represent a total of 2.53 million allocated
addresses. The average size of allocations from this block is 840
addresses, or between a /23 and a /22. There were 7.1 million addresses
remaining in this particular address pool at the end of 2016, and
assuming an ongoing consumption rate of some 2.5 million addresses per
year, the pool will last until late 2019. The remaining 1,306
allocations were drawn from those addresses assigned to APNIC from the
recovered IPv4 address pool. The average allocation size from this pool
was 956 addresses. By the end of 2016 this second address pool was
effectively exhausted, so that all that remains for APNIC is the 7.1
million addresses in the last /8 pool.
However, a further 4.1 million addresses are marked as ‘reserved' by
APNIC. There are a variety of reasons for this marking, including
non-contactability of the original address holder, or addresses
undergoing a period of ‘quarantine' following a forced recovery.
Evidently, efforts are being made to reduce the size of this pool, as in
August 2015 the size of the reserved pool was 4.6 million addresses.
RIPE NCC exhausted its general use pool of addresses in mid-September
2012. This RIR allocated some 3.4 million addresses in 2016, and
recorded 3,306 allocations, using their "last /8" address allocation
policy, which is similar to that used by APNIC. The average allocation
size was 1,024 addresses per allocation, which is comparable to the 2015
figures for this registry. At this allocation rate, the remaining pool
of IP addresses, some 13.15 million addresses at the end of 2016, will
last for a further 4 years, or until late 2020. At the end of 2016 the
RIPE NCC has some 1.05 million addresses marked as ‘reserved, up from
850,000 addresses a year earlier.
LACNIC exhausted its general use pool of IPv4 addresses in June 2014,
leaving some 4 million addresses in its residual address pool. This
registry allocated some 1.6 million addresses in 2016, and there are
314,000 addresses left as "available". LACNIC are holding some 5.0
million as "reserved" at the end for 2016, up from 4.7 million at the
start of the year.
ARIN exhausted its general use IPv4 pool in September 2015. The registry
is holding some 6.3 million addresses as "reserved", which represents no
change to the number of reserved addresses held by ARIN 12 months ago.
AFRINIC address consumption rate increased in 2015 from its 2014 level,
and the 16.9 million addresses allocated in 2015 is the highest address
consumption level for AFRINIC to date. At this stage, it appears that
AFRINIC's remaining 32.7 million addresses as for the start of 2016 will
last a further two years, until the end of 2018. AFRINIC are holding
770,304 addresses in a "reserved" pool.
Finally, the IANA is holding 80,128 addresses in its recovered address
pool in 101 discrete address blocks. This pool is large enough to assign
each RIR a /19 in March 2017, but not quite large enough to support a
/18 round of allocations. This relatively small allocation to each RIR
will have little in the way of impact on the overall IPv4 picture.
The RIR IPv4 address allocation volumes by year are shown in Figure 2.
Figure 2 – IPv4 Allocations by RIR by year
Which countries received the largest pool of IPv4 address allocations in
2016?
Rank 2012 2013 2014 2015 2016
1 USA 28.2 USA 25.0 USA 24.5 USA 7.6 Morocco 3.1
2 Canada 16.7 Brazil 17.4 Brazil 10.9 Egypt 7.4 Seychelles 2.1
3 Brazil 8.4 Colombia 3.8 Morocco 2.6 Seychelles2.1 USA 1.7
4 Russia 5.3 Argentina 1.6 Colombia 2.1 S Africa 2.0 China 1.3
5 Iran 4.5 Egypt 1.6 S Africa 1.7 Tunisia 1.8 Brazil 1.3
6 Germany 3.4 Canada 1.4 Egypt 1.6 Brazil 1.4 S Africa 1.2
7 S Africa 3.4 Nigeria 1.2 China 1.5 China 1.3 India 1.1
8 Italy 3.3 Chile 1.1 Canada 1.5 India 1.3 Egypt 1.1
9 Colombia 2.6 Mexico 1.1 Kenya 1.4 Canada 1.1 Kenya 1.1
10 Romania 2.6 Seychelles1.0 Mexico 1.1 Ghana 0.6 Algeria 1.1
Table 3 - IPv4 Allocated addresses - Top 10 Economies (millions of
IPv4 addresses allocated in the year)
This distribution is not surprising, and the countries in this year's
top 10 list in Table 3 largely corresponds to the RIR with remaining
IPv4 addresses.
IPv4 Address Transfers
In recent years, several RIRs (RIPE NCC, ARIN and APNIC) have included
the registration of IPv4 transfers between address holders, as a means
of allowing secondary re-distribution of addresses as an alternative to
returning unused addresses to the registry. This has been in response to
the issues raised by IPv4 address exhaustion, where the underlying
motivation as to encourage the reuse of otherwise idle or inefficiently
used address blocks through the incentives provided by a market for
addresses, and to ensure that such address movement is publically
recorded in the registry system.
The numbers of registered transfers in the past four years is shown in
Table 4.
Receiving RIR 2012 2013 2014 2015 2016
ARIN 79 31 58 277 727
APNIC 255 206 437 514 581
RIPE NCC 10 171 1,050 2,852 2,411
Total 344 408 1,545 3,643 3,719
Table 4 - IPv4 Address Transfers per year
A slightly different view is that of the volume of addresses transferred
per year (Table 5).
Receiving RIR 2012 2013 2014 2015 2016
ARIN 6,728,448 5,136,640 4,737,280 37,637,888 15,613,952
APNIC 3,434,496 2,504,960 4,953,088 9,836,288 7,842,816
RIPE NCC 65,536 1,977,344 9,635,328 10,835,712 9,220,864
Total 10,228,480 9,618,944 19,325,696 58,309,888 32,677,632
Table 5 – Volume of Transferred IPv4 Addresses per year (millions of
addresses)
A plot of these numbers is shown in Figures 3 and 4.
Figure 2 – Number of Transfers: 2012 - 2016
Figure 4 – Volume of Transferred Addresses: 2012 - 2016
The total volume of addresses transferred in this way is approaching
double the volume of allocated addresses. The aggregate total of
addresses in the transfer logs is some 130 million addresses, or the
equivalent of 7.75 /8s.
This data raises some questions about the nature of transfers.
The first question is whether address transfers have managed to be
effective in dredging the pool of unadvertised public IPv4 addresses. It
was thought that by being able to monetize these addresses, holders of
such addresses may have been motivated to convert their networks to use
private addresses and resell their holding of public addresses. The
numbers appear to show that this has happened, although progress has
been slow. At the onset of address exhaustion in 2011 the unadvertised
pool was at the equivalent of 54 /8s and it was down to 48 /8s at the
end of 2016 (Figure 5). In relative terms the pool dropped from 27% of
the total allocated address pool to 22% in the same period (Figure 6).
Figure 5 – IPv4 Unadvertised Address Pool size
Figure 6 – Ratio of Unadvertised Pool to Total Pool
There is a slightly different aspect to this question, and this concerns
whether the transferred addresses are predominately recently allocated
addresses, where there may be the potential for arbitrage between the
costs of receiving an address allocation from an RIR and the potential
returns from selling these address holdings on the transfer market, and
longer held address addresses where the holder is wanting to realise
otherwise unused assets. The basic question concerns the "age"
distribution of transferred addresses where the "age" of an address
reflects the period since it was first allocated or assigned by the RIR
system.
The cumulative age distribution of transferred addresses is shown on a
year-by-year basis in Figure 7. In 2012 one half of the transferred
address blocks were originally assigned or allocated by an RIR within
the previous 7 years. In 2016 this has dropped to 10% of transferred
addresses. In 2012 30% of transferred addresses had original
registration dates older than 20 or more years in the past. In 2016 this
number has risen to 70%. It appears that these days most of the
transferred addresses are delving into the original legacy address
allocations and reusing them in new deployments.
Figure 7 – Age distribution of transferred addresses 2012 - 2016
The second question is whether the transfer process is deaggregating the
address space and splitting up larger address blocks into successively
smaller address blocks. There are 9,534 address blocks described in the
transfer registries, and of these 6,930 entries list transferred address
blocks that are smaller than the original allocated block. However, this
could be a misleading number, in that the registry entry lists blocks
that are not necessarily aligned to normal address prefix masks, while
the transfer logs list the transferred blocks in conventional prefix and
mask notation. For example, a single registry entry for a transferred
block of 768 addresses would be listed as two transfer transactions, one
for a /24 and the second for a /23 in the transfer log. When this is
taken into account, we have some 893 address blocks that have been split
up into 5,196 smaller address blocks that have diverse address holders.
In other words, some 54% of the listed transfers involve deaggregating
an address block and distributing the smaller blocks to several holders.
In summary, it appears that address transfers involve some level of
address deaggregation. On average, approximately one sixth of the
original address blocks that are transferred (17%) are split into
smaller pieces with multiple holders, and on average this results in
approximately six different holders of transferred address fragments.
The third question concerns the inter-country flow of transferred
addresses. Let's look at the ten countries that sourced the greatest
volume of transferred addresses (Table 6), the ten largest recipients of
transfers (Table 7), and the ten largest country-to-country address
transfers (Table 8).
The transfer logs contain 6,474 domestic address transfers, with a total
of 81,805,824 addresses, while 3,060 transfers appear to result in a
movement of addresses between countries, involving a total of 39,777,792
addresses.
Rank CC Addresses Country Name
1 US 67,903,232 USA
2 CA 13,445,888 Canada
3 RO 6,921,728 Romania
4 RU 4,234,240 Russia
5 GB 4,013,056 UK
6 JP 3,061,504 Japan
7 DE 3,023,616 Germany
8 HK 2,956,288 Hong Kong
9 CN 1,731,328 China
10 AU 1,429,504 Australia
Table 6 – Top 10 Countries Sourcing Transferred IPv4 addresses
Rank CC Addresses Country Name
1 US 67,330,816 USA
2 CN 7,230,976 China
3 IN 5,853,440 India
4 JP 4,676,096 Japan
5 IR 3,836,928 Iran
6 RU 3,119,616 Russia
7 GB 3,054,592 UK
8 SA 2,418,688 Saudi Arabia
9 DE 2,225,920 Germany
10 NL 1,862,912 Netherlands
Table 7 – Top 10 Countries Receiving Transferred IPv4 addresses
Rank From-CC To-CC Addresses From To
1 US US 58,322,432 USA USA
2 CA US 8,785,408 Canada USA
3 US IN 4,150,528 USA India
4 JP JP 2,989,568 Japan Japan
5 RU RU 2,940,416 Russia Russia
6 GB GB 2,513,920 UK UK
7 CA CN 2,359,296 Canada China
8 HK CN 1,934,848 Hong Kong China
9 RO SA 1,656,832 Romania Saudi Arabia
10 CN CN 1,655,552 China China
Table 8 – Top 10 Country-to-Country IPv4 address transfers
The total volume of addresses reassigned in this manner, some 121
million IPv4 addresses over four years, is far less than the underlying
pre-exhaustion address demand levels that peaked at some 250 million
addresses in a single year. It appears that the address supply hiatus
has motivated most providers to use address sharing technologies, and,
in particular, Carrier Grade NAT (CGN), on the access side and server
pooling on the content side as a means of increasing the level of
sharing of addresses.
The outstanding question about this transfer data is whether all address
transfers that have occurred have been duly recorded in the registry
system. This question is raised because registered transfers require
conformance to various registry policies, and it may be the case that
only a subset of transfers are being recorded in the registry as a
result. This can be somewhat challenging to detect, particularly if such
a transfer is expressed as a lease or other form of temporary
arrangement, and if the parties agree to keep the details of the
transfer confidential.
But it might be possible to place an upper bound on the volume of
address movements that have occurred in any period is to look at the
Internet's routing system. One way to shed some further light on what
this upper bound on transfers might be is through a simple examination
of the routing system, looking at addresses that were announced in 2016
by comparing the routing stable state at the start of the year with the
table state at the end of the year (Figure 8).
Figure 8 – Change in the size of the BGP routing table across 2016
While the routing table grew by 59.141 entries over the year, the nature
of the change is slightly more involved. Some 67,504 prefixes that were
announced at the start of the year were removed from the routing system
through the year, and 126,645 prefixes were announced by the end of the
year that were not announced at the start of the year. (I have not
tracked the progress of announcements through the year, and it is likely
that more prefixes were announced and removed on a transient basis
through the course of the year.) A further 16,928 prefixes had changed
their originating Autonomous System number, indicating some form of
change in the prefix's network location in some manner (Table 9).
Jan-16 Jan-17 Delta Unchanged Re-Homed Removed Added
Announcements 586,918 646,059 59,141 502,846 16,928 67,504 126,645
Root Prefixes 286,249 309,902 23,653 252,411 10,803 20,080 46,238
Address span 156.35 158.40 2.05 147.31 2.52 5.58 8.57
(/8s)
More Specifics 300,699 336,967 36,268 250,435 6,125 45,424 80,407
Address Count 51.86 56.04 4.18 47.06 0.81 4.94 8.17
(/8s)
Table 9 – Routing changes across 2016
We can compare these changed prefixes against the transfer logs for
2016. Table 10 shows the comparison of these routing numbers against the
set of transfers that were logged in 2016.
Type Listed as Transferred Unlisted Ratio
Re-Homed
All 1,539 15,389 9.1%
Root Prefixes 1,184 9,551 11.0%
Removed
All 3,287 64,287 4.9%
Root Prefixes 1,877 20,203 8.5%
Added
All 8,663 117,982 6.8%
Root Prefixes 4,617 41,621 10.0%
Table 10 – Routing changes across 2016 compared to the Transfer Logs
These figures show that 5-10% of changes in advertised addresses are
reflected as changes as recorded in the transfers. This should not imply
that the remaining 90-95% of changes in advertised prefixes reflect
unrecorded address transfers. There are many reasons for changes in the
advertisement of an address prefix and a change in the controller of the
address is only one potential cause. However, it does establish some
notional upper ceiling on the number of movements of addresses in 2016,
some of which relate to transfer of operational control of an address
block, that have not been captured in the transfer logs.
Finally, we can perform an age profile of the addresses that were Added,
Removed and Re-Homed during 2016, and compare it to the overall age
profile of IPv4 addresses. This is shown in Figure 9. In terms of
addresses that were added in 2016, they differ from the average profile
due to a skew in favour of "recent" addresses, and 20% of all announced
addresses were allocated or assigned in the past 18 months. In terms of
addresses that were removed from the routing system there is a
disproportionate volume of removed addresses that are between 2 and 10
years old. 20% of removed addresses are more than 20 years old, where
almost 40% of all registered addresses are more than 20 years old.
Addresses that Re-Home appear to be disproportionally represented in the
age bracket of between 7 to 15 years old.
Figure 9 – Change in the size of the BGP routing table across 2016
IPv6 in 2016
Obviously, the story of IPv4 address allocations is only half of the
story, and to complete the picture it's necessary to look at how IPv6
has fared over 2016.
IPv6 uses a somewhat different address allocation methodology than IPv4,
and it is a matter of choice for a service provider as to how large an
IPv6 address prefix is assigned to each customer. The original
recommendations published by the IAB and IESG in 2001, documented in
RFC3177, envisaged the general use of a /48 as an end site prefix.
Subsequent consideration of long term address conservation saw a more
flexible approach being taken with the choice of the end site prefix
size being left to the service provider. Today's IPv6 environment has
some providers using a /60 end site allocation unit, many use a /56, and
other providers use a /48. This variation makes a comparison of the
count of allocated IPv6 addresses somewhat misleading, as an ISP using
/48's for end sites will require 256 times more address space to
accommodate a similarly sized same customer base as a provider who uses
a /56 end site prefix, and 4,096 times more address space than an ISP
using a /60 end site allocation!
For IPv6 let's use both the number of discrete IPv6 allocations and the
total amount of space that was allocated to see how IPv6 fared in 2016.
Comparing 2015 to 2016 the number of individual allocations of IPv6
address space has risen by some 20%. By contrast, the number of IPv4
allocations has fallen by 16% in this same period (Table 11).
Allocations 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
IPv6 473 841 1,243 2,477 3,700 3,403 3,840 4,407 4,733 5,594
IPv4 6,312 6,969 6,701 7,758 10,061 8,619 7,110 10,853 11,732 9,787
Table 11 - Number of individual Address Allocations, 2007 - 2016
The amount of IPv6 address space distributed in 2013 had risen by some
40% over 2012 levels, but in 2014 the total volume of allocated
addresses fell by the same amount, back to the same total volume of
addresses as in 2012. The number of allocations increased, however,
indicating that 2014 there were no anomalous extremely large allocations
of IPv6 address space through this last year. 2015 showed a visible
level of growth over 2014 levels in IPv6 (Table 12).
Addresses 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
IPv6 (/32s) 6,916 15,634 1,555 4,754 20,009 18,136 23,935 17,513 20,225 25,301
IPv4 (/32s)(M) 203.9 203.3 189.4 248.8 201.0 114.9 65.1 63.9 34.8 22.2
Table 12 – Volume of Address Allocations, 2007 - 2016
Regionally, each of the RIRs saw IPv6 allocation activity in 2015 that
was on a par with those seen in the previous year (Table 13).
Allocations 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
RIPE NCC 150 413 595 1,012 1,565 1,661 2,057 2,143 2,206 2,141
ARIN 196 213 357 567 959 545 523 505 602 646
APNIC 61 158 185 637 610 561 505 503 778 1,681
LACNIC 38 43 93 212 447 560 683 1,196 1,061 1,010
AFRINIC 18 14 13 49 119 76 72 60 86 116
473 841 1,243 2,477 3,700 3,403 3,840 4,407 4,733 5,594
Table 13 - IPv6 allocations by RIR
The assignment data tells a slightly different story. Table 14 shows the
number of allocated IPv6 /32's per year. Interestingly the volume of
IPv6 addresses assigned by ARIN in 2015, the year that ARIN exhausted
its remaining pools of available IPv4 space, was approximately one tenth
of the address volume of the previous year. The opposite was seen in
AFRINIC, where the 2015 address volumes were ten times the 2014 volumes.
This is largely due to two large IPv6 allocations in 2015 to Telecom SA
(/20) and Vodacom SA (/24).
IPv6 Addresses 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
(/32s)
RIPE NCC 1,468 964 1,052 2,406 3,174 3,892 6,286 8,217 12,031 21,707
ARIN 148 14,486 236 780 6,344 1,660 12,558 5,241 641 1,088
APNIC 5,236 139 170 1,335 9,486 3,783 4,442 2,644 2,109 1,236
LACNIC 51 35 87 197 948 4,605 597 1,359 974 1,182
AFRINIC 13 10 9 36 147 4,196 51 51 4,471 78
6,916 15,634 1,555 4,754 20,099 18,136 23,935 17,513 20,225 25,301
Table 14 - IPv6 address allocation volumes by RIR
Dividing addresses by allocations gives the average IPv6 allocation size
in each region (Table 15). The average allocation size for RIPE
effectively doubled from a /30 to a /29 through 2016, while it halved in
APNIC from a /31 to a /32. There was a large IPv6 allocation by Afrinic
in 2015 which had no counterpart in 2016, so the allocation size is back
to a /33. Overall, the average IPv6 allocation size remains a /30.
Avg Alloc 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
RIPE NCC /28.7 /30.8 /31.2 /30.8 /31.0 /30.8 /30.4 /30.1 /29.6 /28.7
ARIN /32.4 /25.9 /32.6 /31.5 /29.3 /30.4 /27.4 /28.6 /31.9 /31.2
APNIC /25.6 /32.2 /32.1 /30.9 /28.0 /29.2 /28.9 /29.6 /30.6 /32.4
LACNIC /31.6 /32.3 /32.1 /32.1 /30.9 /29.0 /32.2 /31.8 /32.1 /31.8
AFRINIC /32.5 /32.5 /32.5 /32.4 /31.7 /26.2 /32.5 /32.2 /26.3 /32.6
All /28.1 /27.8 /31.7 /31.1 /29.6 /29.6 /29.4 /30.0 /29.9 /29.8
Table 15 – Average IPv6 address allocation size by RIR
The number and volume of IPv6 allocations per RIR per year is shown in
Figures 10 and 11.
Figure 10 – Number of IPv6 Allocations per year
Figure 11 – Volume of IPv6 Allocations per year
Table 16 shows the countries who received the largest number of IPv6
allocations, while Table 17 shows the amount of IPv6 address space
assigned on a per economy basis for the past 5 years (using units of
/32s).
Rank 2012 2013 2014 2015 2016
1 USA 549 USA 485 Brazil 946 Brazil 815 Brazil 774
2 UK 199 Brazil 473 USA 457 USA 540 USA 603
3 Germany 187 UK 248 UK 239 China 267 China 509
4 Russia 186 Russia 246 Germany 215 Germany 249 Germany 266
5 Netherlands 124 Germany 195 Russia 201 UK 216 Australia 219
6 Australia 113 Netherlands 134 Netherlands 181 Russia 183 UK 211
7 France 111 France 132 France 122 Netherlands 170 Netherlands 198
8 Sweden 90 Sweden 112 Switzerland 103 Australia 123 Russia 173
9 Argentina 78 Australia 102 Italy 103 Spain 119 India 161
10 Poland 77 Italy 98 Australia 101 France 116 Indonesia 159
Table 16 - IPv6 allocations by Economy
Rank 2012 2013 2014 2015 2016
1 Argentina 4,177 USA 12,537 USA 4,930 Sth Africa 4,440 UK 9,571
2 Egypt 4,098 China 4,135 China 2,127 China 1,797 Germany 1,525
3 China 3,136 UK 782 UK 1,090 UK 1,297 Netherlands 1,312
4 USA 1,337 Germany 651 Brazil 863 Germany 1,269 USA 1,137
5 Italy 635 Russia 523 Germany 749 Netherlands 1,010 Russia 1,005
6 Russia 403 Netherlands 463 Netherlands 719 Russia 864 France 926
7 Germany 399 Brazil 450 Russia 716 Brazil 755 Brazil 727
8 UK 356 France 435 France 436 Spain 708 Spain 702
9 Canada 323 Italy 339 Italy 410 Italy 707 Italy 679
10 Brazil 294 Switzerland 265 Switzerland 369 USA 662 China 596
Table 17 - IPv6 Address Allocation Volumes by Economy (/32s)
Four of the countries in Table 17 listed as having received the highest
volumes of allocated addresses in 2016, namely Russia, Spain, Italy and
China all have IPv6 deployments that are under 2% of their total user
population. To what extent are allocated IPv6 addresses visible as
advertised prefixes in the Internet's routing table?
Figure 12 shows the overall counts of advertised, unadvertised and total
allocated address volume for IPv6 since mid 2009. Aside from the obvious
discontinuity in early 2013, when a registration of a single /18
national address allocation of Brazil of a /18 was replaced by the
actual end user allocations, it's clear that the pool of allocated but
unadvertised IPv6 addresses far exceeds the total sum of allocated
addresses.
Figure 12 – Allocated, Unadvertised and Advertised IPv6 addresses
It is probably clearer to see the ratio of advertised to unadvertised
addresses expressed as a percent, as shown in Figure 13. By the end of
2016 less than 8% of the total pool of allocated IPv6 addresses was
advertised in BGP, while slightly more than 92% of these addresses were
unadvertised.
Figure 13 –Advertised IPv6 Addresses as a percentage of the Allocated
Address Pool
Every country is different in this respect, and Table 18 shows the same
ratio of advertised addresses to the total pool of addresses allocated
to entities in each of the 25 countries holding the largest allocated
IPv6 address pools. The table also shows the current estimated level of
usage of IPv6 in that country.
Rank CC Allocated Advertised Ratio IPv6 Use Country
1 US 43,030 138 0.3% 32.9% USA
2 CN 21,196 29 0.1% 0.5% China
3 GB 17,139 2,148 12.5% 25.5% UK
4 DE 16,107 226 1.4% 44.0% Germany
5 FR 11,432 38 0.3% 18.2% France
6 JP 9,415 93 1.0% 19.7% Japan
7 AU 8,864 4,109 46.4% 14.8% Australia
8 IT 7,143 50 0.7% 22.2% Italy
9 SE 5,736 4,148 72.3% 31.1% Sweden
10 KR 5,251 29 0.6% 8.8% Rep. Korea
11 NL 4,939 600 12.1% 76.6% Netherlands
12 AR 4,793 4 0.1% 10.5% Argentina
13 ZA 4,640 9 0.2% 8.0% Sth Africa
14 EG 4,105 4 0.1% 41.0% Egypt
15 RU 3,954 6 0.2% 23.2% Russia
16 PL 3,740 31 0.8% 35.5% Poland
17 BR 3,651 19 0.5% 13.2% Brazil
18 ES 2,800 9 0.3% 9.5% Spain
19 TW 2,359 2,159 91.5% 1.7% Taiwan
20 CH 2,090 111 5.3% 31.0% Switzerland
21 NO 1,618 286 17.7% 14.5% Norway
22 IR 1,491 3 0.2% 4.0% Iran
23 TR 1,326 1 0.1% 49.0% Turkey
24 CZ 1,319 41 3.1% 11.3% Czech Rep.
25 UA 1,082 1 0.1% 4.0% Ukraine
Table 18 - IPv6 Address Allocation Volumes and Advertised Address Count by Economy (/32s)
There is considerable variation in these numbers. Taiwan advertises the
majority of its allocated IPv6 addresses, yet the extent of IPv6
deployment is still quite low. On the other hand, the United States
advertises only 0.3% of its allocated address pool, yet has a IPv6
deployment level spanning one third of its user population. What this
figure points to is that registries, and service operators do not feel
as constrained with their address management practices with IPv6, and
are willing to operate with very low address utilization efficiencies in
IPv6. This is perhaps an instance where one of the original design
objectives of IPv6 is evident, in that IPv6 was intended to provide such
a large address space that it was no longer necessary to pay the price
of operating extremely conservative address management practices. In
that respect, IPv6 appears to have succeeded!
The Outlook for the Internet
Once more the set of uncertainties that surround the immediate future of
the Internet are considerably greater than the set of predictions that
we can be reasonably certain about.
There has been much in the way of progress in the transition to IPv6 in
2016, but that does not necessarily mean that other providers will
follow the lead. We do not appear to think that the collective costs of
NAT deployment are unsupportable, and there would appear to be a school
of thought that says that NATs could cost effectively absorb some
further years of Internet device population growth. At least that's the
only rationale I can ascribe to a very large number of service providers
who are making no visible moves to push out dual stack services. Given
that the objective of this transition is not to turn on Dual Stack
everywhere, but to turn off IPv4, there is still some time to go, and
the uncertainty lies in trying to quantify what that time might be.
The period of the past few years has been dominated by the mass
marketing of mobile internet services, and the growth rates for 2014
through to 2016 perhaps might have been the highest so far recorded were
it not for the exhaustion of the IPv4 address pool. In address terms
this growth in the IPv4 Internet is being almost completely masked by
the use of Carrier Grade NATs in the mobile service provider
environment, so that the resultant demands for public addresses in IPv4
are quite low and the real underlying growth rates in the network are
occluded by these NATs.
In theory, there is no such requirement for IPv6 to use NATS, and if the
mobile world were deploying dual stack ubiquitously then this would be
evident in the IPv6 address allocation data. Unfortunately, no such very
large scale broad scale of deployment of IPv6 was visible in the address
statistics for 2016. This points to a mobile Internet whose continued
growth in 2016 remains, for the most part, highly reliant on NATs, and
this, in turn, points to some longer term elements of concern for the
continued ability of the Internet to support further innovation and
diversification in its portfolio of applications and services.
We should also be seeing IPv6 address demands for deployments of large
scale sensor networks and other forms of deployments that are
encompassed under the broad umbrella of the Internet of Things. This
does not necessarily imply that the deployment is merely a product of an
over-hyped industry, although that is always a possibility. It is more
likely to assume that such deployments take place using private IPv4 (or
IPv6 ULA addresses) addresses, and once more rely on NATs or application
level gateways to interface to the public network. Time and time again
we are lectured that NATs are not a good security device, but in
practice NATs do offer a reasonable front line defence against network
scanning malware, so there may be a larger story behind the use of NATs
and device based networks than just a simple conservative preference to
continue to use an IPv4 protocol stack.
We are witnessing an industry that is no longer using technical
innovation, openness and diversification as its primary means of
propulsion. The widespread use of NATs in IPv4 limit the technical
substrate of the Internet to a very restricted model of simple
client/server interactions using TCP and UDP. The use of NATs force the
interactions into client-initiated transactions, and the model of an
open network with considerable flexibility in the way in which
communications take place is no longer being sustained in today's
network. Incumbents are entrenching their position and innovation and
entrepreneurialism are taking a back seat while we sit out this
protracted IPv4/IPv6 transition.
What is happening is that today's internet carriage service is provided
by a smaller number of very large players, each of whom appear to be
assuming a very strong position within their respective markets. The
drivers for such larger players tend towards risk aversion, conservatism
and increased levels of control across their scope of operation. The
same trends of market aggregation are now appearing in content
provision, where a small number of content providers are exerting a
dominant position across the entire Internet.
The evolving makeup of the Internet industry has quite profound
implications in terms of network neutrality, the separation of functions
of carriage and service provision, investment profiles and expectations
of risk and returns on infrastructure investments, and on the openness
of the Internet itself. The focus now is turning to the regulatory
agenda. Given the economies of volume in this industry, it was always
going to be challenging to sustain an efficient, fully open and
competitive industry, but the degree of challenge in this agenda is
multiplied many-fold when the underlying platform has run out of the
basic currency of IP addresses. The pressures on the larger players
within these markets to leverage their incumbency into overarching
control gains traction when the stream of new entrants with competitive
offerings dries up, and the solutions in such scenarios typically
involve some form of public sector intervention directed to restore
effective competition and revive the impetus for more efficient and
effective offerings in the market.
As the Internet continues to evolve, it is no longer the technically
innovative challenger pitted against venerable incumbents in the forms
of the traditional industries of telephony, print newspapers, television
entertainment and social interaction. The Internet is now the
established norm. The days when the Internet was touted as a poster
child of disruption in a deregulated space are long since over, and
these days we appear to be increasingly looking further afield for a
regulatory and governance framework that can continue to challenge the
increasing complacency of the newly-established incumbents.
It is unclear how successful we will be in this search. We can but wait
and see.
Disclaimer
The above views do not necessarily represent the views of the Asia
Pacific Network Information Centre.
About the Author
GEOFF HUSTON B.Sc., M.Sc., is the Chief Scientist at APNIC, the Regional
Internet Registry serving the Asia Pacific region.
www.potaroo.net