Dal blocco di Tik Tok per gli utenti la cui età non è verificata,…
Internet is composed of 54,000 “autonomous systems” or “networks” (ASs). These should not be confused with telco’s networks. ASs are built on top of telco’s physical networks. ASs complement each other. An isolated AS would resemble an intranet and would not allow a general access to information available on the internet. This means that ASs have an incentive to reach interconnection with other networks through peering or transit.
Peering & Transit
Peering means that two networks exchange traffic directly between each other without going through a third AS. Peering can be physically implemented through connecting two networks with a cable/fibre. Peering agreement can be:
- settlement free: when there is broad symmetry of exchange of traffic between networks. In such case, the two operators simply share the costs of the cable/fibre connection; or
- with a settlement fee: operators peer for economic efficiency reasons but the gains are not split evenly due to traffic asymmetry between the two networks. In such case, “Small Network B” pays a settlement fee to “Big Network A”.
There is transit, on the other hand, when a network buys connectivity from the rest of the internet. The network pays the capacity of the volume of traffic that can cross an interconnection (e.g. at an exchange point).
Study on IP Interconnection in the Netherlands
During the seminar, Jan Tichem – member of Chief Economist Team of the Netherlands regulatory and competition authority (ACM) – presented an analysis of IP Interconnection in the Netherlands. Following the 2014 Netflix vs. Comcast dispute in the US, the Dutch Ministry of Economic Affairs requested ACM to investigate the IP interconnection market in the Netherlands to determine:
- whether restrictive IP Interconnection behavioursexist in the Netherlands; and
- whether existing regulatory instruments were sufficientto address the problem.
ACM published its study in October 2015. The study is structured as follows:
- formulation of two possible theories of harm
- interviews of playersin the market Contents and Applications Providers (CAPs), Internet Service Providers (ISPs), Internet eXchange Points (IXPs), transit providers and experts.
- Assessment of likelihood of competition problemsin Netherlands
Theories of harm
- Exploitation of a competitive bottleneck
In this first theory of harm, an ISP may extract rents from the exploitation of a competitive bottleneck. The competitive bottleneck theory of harm was applied in ACM’s market analysis decision on mobile and fixed voice termination.
Applied to internet, the competitive bottleneck theory may exist when:
- users are not reachableon other network (to be noted: mobile data is of course available, but is currently subject to data limits);
- users would not or could not switchnetworks in case of congestion;
- transit is not availableas a substitute to the current IP interconnection arrangements;
- no countervailing bargaining power
A possible scenario is that the ISP may exploit the situation by (i) refusing a settlement free peering, (ii) demanding a high settlement fee peering, or (iii) offering only partial access to its network.
- Foreclosure of market for content
The second theory of harm explored by ACM is that ISPs might use their market power on the “market for internet access” to foreclose the “market for content”.
Many ISPs offer both access services and content to end users. The idea is that vertically integrated ISPs may compete with CAPs by favouring their own content and hindering IP interconnection with other CAPs (e.g. ISP demands paid peering or congests interconnections). According to ACM findings, such a foreclosure does not occur in practice for the following reasons:
- First, ISPs face competition. Therefore, a CAP experiencing hindering IP interconnection would just switch provider to access end-users.
- Second, there is no incentivefor an ISP to exclude/foreclose a CAP. Indeed, the volume and quality of content in the ISP’s network (whether produced in-house or by competitor) make its internet access product more valuable for end users. Hence, the end-users willingness to pay a higher price for the ISPs access increases.
Not all restrictions are anticompetitive: possible efficiencies and justifications
Restrictive interconnection behaviour may in some circumstances be motivated by anticompetitive concerns. However, ACM notes that it is equally true that this behaviour can be motivated by pro-competitive concerns or otherwise legitimate reasons. From a commercial and economic perspective, the following efficiencies and objective justifications may apply to restrictive interconnection behaviour:
- Protecting transit business: ISPs with a business model based on transitwill have less incentive to peer because a direct link will undermine their transit requests.
- Paid peering may be an efficient contractual solution. For instance, “Big Network A” assure access to “Small Network B” through transit. Then both parties decide to move to paid peering. The payment of settlement fees to “Big Network A” might be justified by the fact that “Small Network B” enjoys better quality access and lower fee than in its transit agreement.
- Settlement fees can simply reflect bargaining strength, mirroring the asymmetryin traffic exchange between two networks.
- Refusal to peer may be caused by excess capacity on other peering links. Upgrading or initializing a peering link might require huge investment.
Difficulty to distinguish between “anti-competitive toll” and “fair bargaining”
ACM is aware of the difficulty to distinguish between “anti-competitive toll” and “fair bargaining” when peering deals are on the spot. However, as Jan Tichem pointed out: “even if the benefits of peering are unevenly distributed, peering is always an efficient solution when settlement-fee is not higher than savings on transit costs plus possible value of quality improvements”.
Experience of ACM at the time of the report (end 2015)
The likelihood of competition problems such as anti-competitive settlement fees or refusals to peer resulting in consumer harm is currently very low in the Netherlands. Paid peering is rare and the vast majority of peering arrangements are closed via a handshake. Even in cases where paid peering would have been more efficient, CAPs in the Netherlands reverted to transit access because they did not want to set a precedent by paying fees to a ISPs. According to interviewed CAPs, no degradation of quality due to IP interconnection conflicts was observed. Transit seems to be efficiently priced and there was sufficient transit capacity anyway.
Bargaining peering agreements might sometimes lead to temporary congestion, but parties seem to find the most efficient way of interconnection anyway. It is not clear if settlement fees sometimes paid by CAPs are too high. Competition law seems sufficient to address potential problems.
7 marzo 2017