[FIRST DRAFT; QUERIES/NOTES IN CAPS AND SQUARE BRACKETS]
Sahana, a free and open source (FOSS) disaster-management system, is a web-based collaboration tool that grew out of the 2004 Asian tsunami disaster. It addresses common coordination problems experienced during a disaster, including finding missing people, managing aid, managing volunteers and tracking camps. Sahana facilitates communication and coordination between government groups, the civil society (NGOs) and the victims themselves.
This case study is based in part on a keynote address delivered at ApacheCon US 2008 by Shahani Markus Weerawarana PhD who, as a lead developer on Sahana Phase I, 'witnessed the birth of a new global community rising above the depths of despair'.
The start of the Sahana story
The Indian Ocean Tsunami struck on 26 December 2004. The first wave thundered into the coastline of 11 Indian Ocean countries at 8.30 local time; a second wave hit 20 minutes later. With two-thirds of its coastline affected, Sri Lanka was one of the countries hardest hit. In that country alone, 35,000 people died, 100,000 houses were destroyed and a further 50,000 houses were considerably damaged. In total, 5 per cent of the Sri Lankan population became instantly homeless, and more than 500 million kilos of rubble and waste was dumped on the island.
The rest of the world responded within hours and donations started flooding into aid agencies. In the affected regions, local people set up refugee camps, aid collection points and medical camps in frenzied, chaotic attempts to alleviate the suffering. Within days, foreign aid workers and aid had begun to arrive in the area. It soon became clear that a system to coordinate relief efforts and the information generated by the disaster was required, and that any existing systems were unsuitable for a disaster of this scale. In Sri Lanka, a small but growing group of volunteers, predominantly from the Sri Lankan IT industry, began to build software to help coordinate the relief efforts. They did so by creating an open development community. The majority of the early work came from Sri Lanka itself, but critical support was forthcoming from other significant players around the world.
The birth of Sahana
On 29 December 2004, Sahana (a Sri Lankan word for ‘relief’) was born. A call to the Sri Lankan prime minister, with a single question, had given the IT team the direction it needed. The question was, ‘What is the greatest need right now?’ An answer came immediately: a missing persons registry, to help reunite people separated in the evacuation efforts, and help rescue workers identify the dead.
Development efforts were frantic. With more than 80 people from around the world active on the project, work continued around the clock, and the first useful release was made within about seven days. After that, major releases were being made almost daily, and the first implementation of Sahana Phase I was produced in three to four weeks. During this period, over 26,000 families were tracked in the missing persons registry.
It was quickly realised that Sahana had mobilised enough effort to go beyond the missing persons registry. People out in the field had reported that there were refugee camps that had more equipment, food and medical supplies than they required, while other camps needed more supplies. Facilitated by the open development structure, individual effort was immediately directed (by the individuals themselves) to areas in which they could make the most impact. Sahana Phase I addressed these needs in the following key modules: the People Registry; the Organisation Registry; the Camp Management System, and the Request/Assistance Management System. [CAN WE ADD SOMETHING HERE ON HOW THE OPEN SOURCE COMMUNITY WAS BUILT?
Sahana Phase II
Since 2004, Sahana has been completely rebuilt to provide a more maintainable architecture. Known as Sahana Phase II, it has an impressive list of additional features, including a messaging module and a child protection module. It is more flexible and powerful than its predecessor, and can cater to any type of disaster, from a tsunami to an earthquake. The system has been successfully deployed in China, Peru, New York, Indonesia, Philippines, Pakistan, Myanmar and Sri Lanka.
Sahana II is being developed and maintained by a team of six full-time developers, led by the Sahana Project Management Committee. The core developers are assisted by a global community of over 200 participants and contributors, including emergency management experts, humanitarian consultants, academic and FOSS developers, with sponsorship from the Swedish International Development Corporation and formal support at the organisation level from the IBM Crisis Team and LSF (Lanka Software Foundation).
Sustainability through open development
Propelled into existence by the humanitarian demands following the 2004 disaster, Sahana was perhaps necessarily an open development project. With very little time to plan, volunteers worked wherever necessary to get the system up and running, with little in the way of formal project structure or managed development roles. Despite this, it is clear that open development represents a good choice for the project for a number of reasons [WHICH ARE DISCUSSED BELOW?]
Customisation and localisation
Clearly disaster can strike anywhere and equally help can come from anywhere. For this reason it is vital for a project such as Sahana to be readily adapted to suit pre-existent local IT systems and readily localised into a wide variety of languages. FOSS licensing and open development methodologies decrease the overheads associated with achieving these goals, making the software's source code immediately available and legally adaptable without requiring additional permission from its owners and maintainers.
Benefits to local and global community
As open development and FOSS licensing permits anyone to customise and localise the software it covers, it is entirely practical to build IT support networks among local people, minimising the need to send money for support abroad. Commercial disaster management software tends to be created in developed countries, and money spent on licensing it and obtaining support for it will tend to leave the local economy and not result in a significant amelioration of the local skills base.
If you can see the source, you can more accurately judge what actions the software will perform. While governmental organisations may at first be wary of deploying code that originates from outside traditional large software development firms, the ability to examine and change any aspect of the software's structure and operation can effectively mitigate this reluctance.
Increased rate of development
While open development does not necessarily imply rapid development (in fact, a project to create a FOSS disaster management application called OpenEOC was started in 2003 and failed to get a single line of code committed), it does mean that when a widespread will to contribute does exist, barriers to submission are low in comparison to a closed development project. Thus open development helps to harness the will to contribute where it exists.
Building on existing tools
The flexibility and global reach of free and open source licensing makes it easier to incorporate other free and open source tools into your project. No formal licence negotiation or even additional overt permission will usually need to be sought. This can greatly expedite the creation of software to answer an immediate need.
Spread of costs
As open development allows effort to be parallelised across as many willing volunteers as you have, it consequently allows costs to me met from an equally wide variety of sources.
Outcomes and lessons learned
[FLESH OUT INTO A PROPER INTRO]Valuable lessons have been learned since Sahana was first deployed. These include:
- Use of disaster-management system should be approved by authorised relief coordinators (usually from the government), to ensure that everyone works together.
- Open systems should be built, so that all parties can have access to data.
- Dedicated professional teams to provide localisation, customisation and help-desk services should be set up immediately.
- System should be flexible, permitting vaguer data recording standards for times of high uncertainty and chaos while allowing more precise data management when easing of the situation allows.
The efforts of the Sahana team have also spawned a concept and community known as Humanitarian-FOSS, founded by humanitarian consultant Paul Currion and the Sahana project lead Chamindra da Silva. Motivated by broader humanitarian ideals than disaster management, Humanitarian-FOSS promotes the use of IT to improve human welfare.
[A FEW SKETCHY IDEAS] The Sahana team’s long-term goal is to ensure that communities, especially those in regions prone to disaster, are prepared for and therefore better able to manage their own disasters when they occur. To achieve this, the Sahana team, working with government bodies, NGOs and civil society groups, is developing tools that will make it possible to incorporate plans and material such as communication directories in advance, and promoting other techniques to encourage a state of preparedness.
More immediate challenges include: [MENTION THESE FIRST?]
- Securing ongoing funding
- Getting professional people involved at strategic level (by providing financial incentive?)
- Gaining access to public data
- Ensuring project awareness/raising profile/understanding of FOSS and therefore engagement
- Directing volunteer contributions in a constructive way
[ANYTHING TO ADD? PERHAPS ALONG THE LINES OF HOW SAHANA CAN BE APPLIED TO SITUATIONS?] The success of Sahana is in part a result of the open development model, which when applied correctly and managed effectively, is highly appropriate for producing a disaster-management system. Not only can the software be obtained free of charge, but it can also be developed at far less cost than a commercial system, and more quickly, if there is a small dedicated team assisted by the global FOSS community. This encourages worldwide volunteer and community involvement and, therefore, a global solution to a problem, benefiting the community at large.
Sahana and its availability as FOSS has made a single cohesive disaster management system possible, which will not only facilitate more efficient disaster management and relief efforts in the future, but also help those regions most at risk to better prepare for and deal with disasters when they occur.
Gavin Threadgold, Sahana - engineering a sustainable ICT solution for disaster management, paper published for Digital Earth '06, Summit on Sustainability, Auckland, New Zealand;
UNDP IOSN case study;
National Geographic website