The main obstacle to building broadband Internet networks in rural areas is the low economic viability of constructing such networks when expensive commercial technologies to reach relatively few and dispersed users are used. The high cost of Internet upstream bandwidth, particularly in rural areas, where providers tend to be few, adds significant challenges. Further, frequent long hours of total service downtime and high congestion are common, often resulting in slow, limiting, and erratic service.
One of the necessary conditions for ISPs operating in such areas is in finding ways to ensure uninterrupted Internet connectivity and maximizing the utilization of available bandwidth with minimal impact to the user.
The AirJaldi team worked on the development of a unique set of solutions for this challenge, collectively referred to as the “Bandwidth Maximizer” (BwM). Essentially, AirJaldi’s BwM will combine tools for identification and prioritization of traffic with local content caching and filtering with peering-optimization tools. This will enable the use and load-balancing of multiple upstream connections based on traffic prioritization and bandwidth costs*.
In this proposal we seek funding for streamlining and sharing our achievements in this field to date, as well as for laying the foundation for further essential work in this field, to be carried out with future additional funding.
The expected results of this proposal are:
(i) A prototype bandwidth maximizer integrated and operational on a working network – the AirJaldi Network in Dharamsala.
(ii) Integration and component configuration of the working prototype documented and summarized in manuals to be made publically available along with the source-code.
(iii) The technical and financial benefits of the BwM clearly demonstrated and documented.
(iv) Outlining further work needed for improving the product and making it suitable for easy replication.
* A more detailed paper on the BwM concept is available upon request.
Project background and justification:
The general quality standards of Internet service providers in developing countries tend to be low as compared to developed ones. Total service downtime is frequent; when service is available, high congestion is common, resulting in slow, limiting, and erratic services and overall low user experience.
Moreover, bandwidth costs in developing countries tend to be high – in fact much higher than in developed countries. In rural areas, the situation worsens with costs for service increasing as service quality decreases.
The success and viability of any Internet network operating under these realities hinges on finding ways to successfully maximize the number of users per available bandwidth with minimal impact on their experience. The erratic and problematic nature of many upstream connections makes it difficult to provide good Internet service to subscribers while being dependent on a single upstream provider for the whole network.
Finding ways to effectively and economically increase redundancy through use of multiple providers is therefore essential. Where Internet connectivity is available, usage patterns have a profound impact on bandwidth congestion and hence user experience. While some of the most-often used applications and protocols (for example, email, web browsing, VoIP, social networks) tend to consume relatively little bandwidth, services such as peer-to-peer (P2P, for example, Kaaza, emule) applications can easily lead to available bandwidth being “hogged” by a few users. (Robert Pepper, at a presentation in the IGF in Hyderabad in December 2008, suggested that “The top 1 per cent users use about 35 per cent of the bandwidth; the top 10 per cent use around 90 per cent of the bandwidth”). Developing ways to successfully address such “hogging” will ensure better overall user experience as well as increase the number of users per available bandwidth (commonly referred to as “oversubscription ratio”).
Information Society Innovation Fund
Comments