IXPs in Latin America and the Caribbean: From traffic exchange to the sharing of knowledge
Internet users are looking for a positive user experience with the least possible latency and the lowest possible cost; IXPs play a key role in achieving this.
By Alejandro Prince and Lucas Jolías*
In order for the Internet to actually work as a “network of networks,” an Internet exchange point (IXP) structure is needed that will allow interconnecting the traffic flowing through the various providers’ networks (ISPs). In this sense, an IXP is similar to a local airport where individual airlines exchange passengers instead of doing so at airports abroad. Just as tourists wish to reach their destination having to deal with as little congestion as possible, Internet users wish to have a positive user experience with the least possible latency and the lowest possible cost, and IXPs play a key role in achieving this. Traffic Exchange Points allow information to be routed locally, thus reducing unnecessary traffic to devices abroad. Exchanging traffic locally also results in increased Internet speed and quality of service.
According to the IXP Toolkit, a project of the Internet Society that aims at developing a global map of Internet exchange points, there are 448 IXPs worldwide. In Latin America and the Caribbean, a first wave of IXPs began operating during the second half of the 90s, led by Panama, Argentina, Ecuador, Peru, Chile, and Brazil, and following a slow and uneven liberalization of the telecommunications sector. In recent years, new IXPs have been installed in countries such as Mexico, Costa Rica, Bolivia, Haiti, and Curaçao. At the Internet Society’s request, during 2014 Prince Consulting prepared a study on the economic, social and organizational impacts of IXPs in nine different Latin American and Caribbean countries, with a focus on best practices, the historical characteristics of each IXP, and the benefits they have generated.
There are different organizational models among IXPs. On the one hand, some IXPs are operated by non-profit, neutral business chambers or associations, such as in the case of Argentina, Peru, Panama, Ecuador, and Haiti. On the other hand, some IXPs are managed and administered by universities or academic centers, such as in the case of Mexico or Costa Rica. Finally, some Internet exchange points are operated by private companies such as in the case of Brazil, Chile, and Curaçao, or the state itself has significant participation, as in the case of Bolivia. As can be seen, there is no single successful model. Instead, each IXP depends largely on the historical characteristics of the specific country’s telecommunications market, the evolution of the regulatory framework, the stakeholders involved and their interests, as well as the existence of “champions” or leaders willing to implement these initiatives.
IXPs have had many direct and indirect benefits and impacts. The scope and intensity of these impacts varies whether one considers the short, medium or long term. Roughly, the impacts of an IXP can be divided into three categories: connectivity and quality, economic and commercial, and cognitive and organizational. Clearly, maintaining local traffic reduces latency and increases speed, thus improving users’ experience. These effects are strongly enhanced when CDNs are also hosted locally, which is also one of the greatest incentives for CDN deployment. In addition, IXP presence promotes local content development, which in turn helps reduce international traffic and consequently lowers system costs and improves traffic. Greater backup governance completes the list of benefits resulting from better traffic management. The interconnectivity and redundancy of this shared traffic exchange format can allow effective network availability in case of emergency situations or disasters, which was the case of the Haitian IXP during the 2010 earthquake.
From an economic point of view, the shared exchange model reduces operating and maintenance costs directly in the short term, while also generating a long-term reduction of international traffic. Reflecting these two benefits, costs and barriers to entry for new players and services are lowered, end-user prices are reduced, and quality of service is improved. The above is clearly a key factor for greater digital inclusion. Where the IXP administrator is a non-profit and/or tax-exempt organization, the impact on the lowering of costs is much higher and equitable.
All respondents in the study (individuals responsible for the region’s IXPs) confirmed IXPs’ technical and economic short- and medium-term impacts. IXPs with greater experience and development valued the long-term benefits deriving from the mere fact of their association. Though intangible and emerging, working collaboratively and sharing costs, resources, and management develops knowledge and a skill set that are difficult to acquire individually. Sometimes called coopetition, cooperative competition makes it easier to set up spam filtering, peering and other types of agreements from the start; at the same time, it allows optimizing the Internet ecosystem and making it more transparent. Cost and equipment sharing as well as the sharing of knowledge and problem solving abilities mean that an IXP and its participants are part of a community that is constantly adapting and evolving. With its regulatory changes, its rapid growth and unpredictable technological disruptions, the dynamics and balance of the Internet ecosystem require new, smarter and more effective conflict solving forms of organization.
In times when neutrality and freedom of expression are central topics in the debates on governance and the future of the Internet, we believe that the IXP model promotes greater fairness and de facto neutrality.
*Directors of Prince Consulting