IPv4 trading remains active through Q3 2017, especially in the small block market.  Strong demand for large, legacy blocks continues but  supply still lags.  Prices are on the rise across all block sizes.  

Transfer Activity

This past quarter, the number of IPv4 transactions and volume of IP addresses flowing into and out of the ARIN region fell very slightly compared to number of transactions and volume in Q2 2017.  This small drop in number of transactions is the first we’ve seen in any quarter since the market started picking up steam in 2014.  Current levels, however, are still higher than any quarter in 2016, and reflect a continued healthy level of trading activity, particularly in the small block market (< 65,536 numbers) which comprises about 90% of traded blocks.

The 5.3 million IPv4 numbers traded in Q3, albeit a decline from Q2, are still 20% higher than numbers transferred in the first quarter of 2017, and represent an even larger increase over volumes in Q1, Q2 and Q4 of 2016, all of which were in the range of 3-3.5 million numbers transferred per quarter.   Compared to this time last year, however, Q3 2017 numbers pale.

In Q3 2016, the market experienced a 2015-like bubble with over 11 million numbers transferred, 80% of which were attributable to 3 large block transactions.  We have yet to see a similar burst of massive large block buying in a single quarter in 2017.  Although demand for large blocks remains high, a shortage of supply, which began in early 2016, continues.  As prices rise (see below), we anticipate that more supply will enter the market over the next 12-18 months to meet at least some of this demand.

The /24 block size continues to be the most prevalent transferred in Q3, and offers a reasonable proxy for measuring growth in the IPv4 small block market in general.  In 2015, 67 /24 blocks transferred within or from ARIN in total.  In Q3 2017 alone, there were 90 such transfers.   Among the larger block sizes, the /16 continues to dominate, appealing both to small block buyers who are willing to invest in excess capacity, and those looking for larger blocks but willing to buy in smaller increments.

Pricing

Block prices are on the rise.

At around this time last year, small block pricing (< 4,000 numbers) was in the range of $12- $13 whereas mid-size blocks (65K – 135K numbers) and large blocks (>1M) had converged at around $8 - $9.  As we head into the last quarter of 2017, small block pricing is closer to $14-$15 per number, and mid-blocks are in the $10-$12 range.  But the big news is in large block trading.  Due to the supply/demand imbalance, large block pricing expectations are closing in on $15 per number, which we anticipate will soon be the floor pricing, not the ceiling.   Whether large block pricing continues to escalate throughout 2018 or begins its decline will depend on whether there is an infusion of large block supply to keep the market adequately liquid, and whether there is a change in the rate of IPv6 deployment that affects long term deployment strategy.

IPv6 Deployment

Worldwide global native IPv6 hit 20% in July according to Google user stats.  By the end of the quarter, user connectivity was ranging from around 16.5% to 17.5% during weekdays to closing in on 21% during weekends.   In our 2016 State of the IPv4 Market Report, we had predicted an acceleration in IPv6 deployment and estimated v6 deployment would hit 24% by the end of 2017.  We seemed on track for this when global v6 deployment was climbing at a clip of 1% per month between April and July.  But things have slowed since and at this point hitting between 22% and 23% is a more realistic year-end projection.

So far the trajectory of IPv6 deployment in 2017 is mirroring 2016.  Global user connectivity climbed about 4 percentage points during the first 3 quarters of 2016.  The same progress was made during this same period in 2017.  According to Alexa Top 1000 statistics, websites reachable over IPv6 climbed by 3 percentage points over the first 3 quarters of 2016.  In 2017, Alexa statistics show growth of 2.5 percentage points through Q3.

In the U.S., some network operators have made impressive progress in broadening their deployment of IPv6.  For example, Google Fiber, Verizon, T-Mobile, and AT&T show IPv6 deployment based on overall traffic volume in excess of 70% (see http://www.worldipv6launch.org/measurements/).  Users still are not flocking to IPv6, however.  At the end of Q3, Google stats show user IPv6 connection percentages in the U.S. slightly below where they were at the end of 2016.   One analysis of IPv6 enable browsers projects deployment reaching 50% globally and 75% in the U.S. by November 2019.  (See http://www.vynke.org using a logistic s-curve regression formula). Given recent progress, this may be a stretch, but time will tell.

A key takeaway from both the transfer and IPv6 deployment data is that there continues to be little evidence that IPv6 is close to replacing IPv4 as the dominant protocol for Internet routing, and there is still no evidence of any adverse relationship between IPv6 deployment and the IPv4 market.  This could change if the market shortage of large block IPv4 address space supply continues over the next 12-18 months and network operators planning to still dedicate resources to IPv4 instead turn their attention toward accelerating v6 migration.

RIR Transfer Policies

Transfer policies that streamline M&A transfers and relax needs based requirements in the ARIN region were implemented in August.  At the ARIN meeting in San Jose earlier this month, the only active transfer-related policies, prompted by transfer proposals in both AFRINIC and LACNIC, addressed whether ARIN should implement a policy that permits inter-RIR transfers with RIR regions that permit only one-way inter-RIR transfers into their region.  Although the policy in LACNIC is still under consideration, the proposal in AFRINIC was abandoned.  At the ARIN meeting, there was only luke-warm support for such a policy as a pre-emptive measure.   A policy proposal that would close loopholes in the current interRIR policy but would have the effect of prohibiting all APNIC transfers received no visible support from the membership.