My Big Campus

Tamarind Tree School and Nomad India Network are two non-profit organizations based in rural Dahanu in the western Indian State of Maharashtra in India. Together, they have seeded a wireless mesh network targeted at extending the school’s educational resources to the homes of the first generation learners. The idea for the project was formed after I attended a workshop organized by the Open Technology Institute (OTI) and Airjaldi in Dharamsala, India, in June 2012 and shared my experiences with people at the Tamarind Tree School. We started building this network after some research and outreach. Founded in 2007, the Tamarind Tree School is committed to building an equitable and just society where individuals and communities can realise their full potential. The school believes that bridging the digital divide can play a significant role in overcoming social exclusion and a lack of voice that are the root causes of poverty. The Nomad India Network, a partner with Tamarind Tree, is an organization based on the principle of social entrepreneurship. In its nine years of existence, it has emerged as a leading organization providing technical know-how to communities wanting to start their own community radio station. With its flagship low cost FM transmitter, Nomad has setup more than 35 community radio stations all over India.

The project, “My Big Campus,” uses the Commotion mesh software on top of standard 2.4GHz WiFi devices. We plan on linking 150 to 200 households in the first phase, with the intention to spread the network across 8 to 10 villages in later stages, predominantly inhabited by the Warli community.

Image: The planned network in Dahanu

Located in the western Indian belt, more than 90 percent of the schoolchildren are Warlis, an indigenous community that resides mostly in the forests of Maharashtra and Gujarat. This community largely remains separate from mainstream society. In a region with very poor indicators of literacy or health, the school is a community driven initiative to provide quality education to the local tribal children.

The school currently runs classes from pre-Kindergarten to Grade 3 with the intention of increasing one class every year to complete secondary education. Teachers are recruited from around the community and trained to follow a plethora of pedagogies using art and craft, open source technology, storytelling, and working the land, among other methods. The curriculum also includes tribal narratives from their tradition with the goal of helping students become balanced individuals with a sense of their cultural identity.

An offline network to host local applications

Over the years, we have discovered that a critical barrier to learning for these students is a non-conducive environment at home, where a conventional school has limited impact. The lack of last mile connectivity means Internet access continues to be elusive for most households in the region. The legal environment in the country is also not conducive for community-based initiatives for Internet distribution as it restricts the use of Wi-Fi for Internet access. As a result, the digital divide in the community is widening.

Against this challenging and growing digital divide, Tamarind Tree is working to create an environment that enables and facilitates learning for the first generation learners in the rural community of Dahanu. The project is an attempt to reduce the digital divide and provide equal and contextually relevant access to the digital world for these students with community-driven local applications running over the mesh network. When the students are at home they can access local applications hosted on the Commotion mesh network.

The school is the primary location for maintaining e-learning servers that host applications on Moodle; a video distribution platform that will support e-learning so that these students can access educational materials/videos even when they are at home; an Icecast server for community radio that will help community members learn about local information through media that is created by people living in the community as well as relay educational programming for these students; and an Asterisk server for free telephony among students, teachers, and parents. This creates an enabling environment for students in their homes to advance their learning.

Building the network

Image: Young community members putting up a Commotion node

The network building team is spearheaded by two young members from the community, Nitin and Anil, who learned to install routers with Commotion and configure them for deployment. They used Commotion Construction Kit (CCK) modules to learn to configure nodes running Commotion firmware. Commotion provides a graphical interface for configuring and visualizing the mesh network, which makes it much easier to discuss wireless networking with youth in the Dahanu community, whose first language is not English.

Image: Mesh visualizer with three nodes

We use Radio Mobile for planning our network buildout. This tool uses digital terrain elevation data to map and determine if the WiFi radio links will encounter obstacles. It provides an automatic extraction of the path profile between the emitter and receiver to predict the performance of a network.

Currently, four nodes run the Commotion firmware and the network provides an e-learning platform in the school premises. Network users access a Moodle and Icecast server hosted at the Nomad Lab (see the close up of the network map) while at the school. Users can also access the network in the dining area, where a lot of children and teachers get together in their free time.

Image: Current network in Dahanu (traced in green)

In the beginning, we tried using consumer grade Wi-Fi devices installed with Commotion as the operating system. The efforts did not produce good results due to non-standard and non-descriptive hardware. In order to jumpstart the project from that stage, we decided to use well-tested Ubiquiti routers even though the high cost of equipment was a major hindrance. The cost of Ubiquiti devices was too high for community members to afford and hence we decided that Tamarind Tree and Nomad will bear the cost of the initial main links of the network. The recent weakening of the Indian Rupee against the global currencies made the cost challenge even greater for us.

Image: Using a sector antenna

We realized a way to reduce the cost of a mesh node while boosting performance was to build antennas ourselves, since the antennas can be more expensive than the radio. For example, a Ubiquiti Rocket M2 for 2.4 Ghz would cost Rs 5,800 (US$ 90) while a sector antenna is available for Rs 9,600 (US$ 154) in India.

In the past three months, we started working on designs for Wi-Fi antennas with active participation from community enthusiasts. As of now, we have an omnidirectional antenna with 16 dB of gain, and we have another workable design for a directional antenna. We have done a comparative analysis of cantenna, Yagi antennas, and narrow beam grid antennas, and we feel that a narrow beam grid antenna may work out better for us. This is to ensure that a long distance link (more than 5 km) remains steady in a region which has very high humidity and very, very poor reflective surface with thick vegetation. We initially decided to use a Yagi antenna for this purpose but now my understanding is that a sectoral antenna may not achieve what a narrow beam directional antenna would in such a situation.

One major challenge in building the network is very thick vegetation in the region. Some parts of the area are covered with dense orchards producing tropical fruits while some parts are forested mountain.

The future of the network in Dahanu

The future of this project depends on further research and development, assistance from domain experts to develop antennas, and training youth to cobble up antennas for their radios. We plan on energizing our arterial network with higher powered Ubiquiti routers such as Rockets, NanoStations, Bullets, and Airgrids for their long-range capabilities.

Even though access to Internet is a great leveler, the project emphasizes on local applications in this phase. Digitizing a community based on local applications could be more contextual, participatory, and less intimidating. The local applications like a local e-learning platform, video distribution server, streaming audio, and telephony could give an immediate impetus to these community members to join the network. Steadily, more parents are beginning to believe in this project and are purchasing laptops for their children, these young first generation learners in the tribal areas of Dahanu, so that they can benefit from the learning environment that this network would extend to them.