By Gabe Fried, Chief Executive Officer, Hilco Streambank
Recently, I had a conversation with someone whom I would classify as an IPv4 market opponent while I was at the African Network Information Center (AFRINIC) in Nairobi. It was a useful conversation for me, and hopefully for him as well.
Taking a cue from everything I’ve read about productive dialogue, I asked him about his concerns and we proceeded from there. His principal concern was that the presence of IPv4 brokers in the market would cause IPv4 pricing to be higher than it would without the brokers. This is an easy conclusion to reach if you look at the prices that the buyer pays, the proceeds that the seller receives, and the commission that the broker earns in the process. Remove the IPv4 broker commission and there is room for both buyer and seller to improve their outcomes. A larger concern was the existence of a market at all, when Registries already had a mechanism to deal with the allocation of scarce resources that wasn’t purely price. Totally understandable points of view, but it overlooks a few other things that brokers do, which actually improve market outcomes.
But I suspected there was more to his concern, so I asked a probing question: What if you’re right, and IPv4 brokers cause prices to be higher than they would otherwise be? Isn’t IPv4 an old technology at this point? It’s been declared “historic” by the IETF. Who cares if the prices are high?
The answer didn’t surprise me at all, as it’s been at the core of every “soft landing” discussion in each Regional Internet Registry (RIR). He was concerned that new African ISPs who currently did not have access to resources would price the end user out of the market, leaving more end users offline for longer as a result. The lack of access should clearly be a concern to everyone who believes that internet access is socially beneficial and has the potential to enrich lives, further education, enhance standards of living, slow the spread of disease, improve economic efficiency and build communities. For sure, universal internet access is a worthwhile objective, and sooner rather than later is always better. However, it’s not clear to this author that the address transfer market impacts universal access in any way. I share his concern, but think it’s misplaced in the context of the transfer market.
I asked if $5,000 for a /24 would be prohibitively expensive for new entrant, such as a small ISP, in a poor country who wanted to bring a community online. The answer was unambiguously, “yes.” At $4,000? Still yes. At $3,000? Yes.
This makes the case that having an IPv4 broker in the middle of that /24 transaction has no impact on connecting that poor community. When non-profits such as the Internet Society (ISOC) are involved in building networks, some of their funding will undoubtedly go toward acquiring IP addresses, which diverts the money away from other uses and would clearly have a negative impact on new deployments. The right price for that yet-to-be-connected community is as close to $0 as possible. Or IPv6. Those issues need to be addressed by community policy around the last /8 and continued progress to deploy IPv6. I will return to this point later.
I asked this individual if I could offer a perspective on how I believe we create value in the marketplace, and he graciously agreed to listen. I boiled it down to three quick points.
First, the IPv4 brokers do the digging/mining to find allocated but unused blocks and bring them to the market, which increases supply. If you refer to introductory economics, increases in supply reduce price, (holding everything else constant). This work requires retracing the history of defunct companies, cold calling into organizations to find the right person in charge of the network who might be able to find available address ranges, and working with Registry staff to ensure that there wasn’t fraud perpetrated somewhere along the way. We bring clean supply to the IP address marketplace. When the price is zero, no one has an incentive to hunt for inefficiently used space and facilitate a transfer.
Second, we provide price information to the market that’s more robust and reliable than posting to the local ‘NIC or ‘NOG mailing list to ask if a certain offer you’ve received is reasonable or not. By posting price data going back several years and covering hundreds of transactions, we provide support for those people who need to make a case internally for buying or selling IP addresses. We’ve heard that charts of our data on address sales have appeared in numerous investment memos or requests to sell IPs on the basis that the proceeds will be used to fund infrastructure investments.
Third, we simplify the transaction process. Whether it’s creating standardized documentation, developing relationships with RIRs to get answers quickly, or providing free escrow services, we do everything to save buyers and sellers time (which is money) and aggravation. Most organizations will only buy or sell IPv4 once or twice. Developing the expertise in-house is time consuming and wasteful.
We’re looking forward to bringing our IPv4 transfer market expertise to AFRINIC members in the coming years and working with all of the stakeholders in the community to productively address concerns as they arise. It’s been a pleasure participating as an IPv4 broker in several AFRINIC meetings so far and we look forward to more.
Returning to the issue of available “free pool” addresses for new entrants, the communities of each Regional Internet Registry have developed policies aimed at meeting the specific needs of their region. In the ARIN region, no reserve pool was left for new entrants. In the rest of the world there are reserve pools for either new entrants, incumbent members, or both. All of these regions restrict allocations from these reserve pools. The debate typically revolved around the issue of “need” and putting the needs of existing members either ahead or behind the needs of future entrants. What’s clear from each policy is that the “landing period,” whether “soft” or “hard,” triggered an active interest in the trading market for existing RIR members as well as their downstream customers. The existence of the free pools appears to have little impact on market activity in the short run. Hopefully, those not yet online can get online soon, and hopefully on IPv6, where address space is plentiful and therefore inexpensive.
Some marketplaces create externalities that are not fully built into the price. Some of those externalities are positive, and some are negative. What we’ve found in the IPv4 marketplace is that it creates a business case for migrating to IPv6 for sellers (good!) and provides a cost-efficient way to get started for buyers, who presumably are rational and therefore also advantaged by the process. Increasing overall supply into the marketplace has no impact on available free pools and creates choices for buyers and sellers. The existence or absence of the IPv4 marketplace is an indicator of how ready the world is to abandon IPv4.