The term cloud computing first appeared in October 1996. Now, 22 years later, it has become the default mode for software delivery. But users in aid and development organisations often work in places lacking cloud cover, with limited or no internet access.
“Offline capabilities are a big priority for us,” says Amy O’Donnell, digital in-programme lead for charity Oxfam. “We’re working in rural Ethiopia and the middle of DRC [Democratic Republic of the Congo] where we simply can’t take connectivity for granted.”
Mobile networks are the main provider of telephony in many developing countries, and in some cases data collectors can use SIM card-equipped devices to send information from remote locations at a low bandwidth. “But oftentimes, we find it more reliable and cost-effective to collect offline and sync when it’s possible,” says O’Donnell.
Oxfam uses software from Mobenzi, a specialist South African provider, and Survey CTO, the latter based on the open source Open Data Kit.
Similarly, World Vision International uses Open Data Kit-based software, including KoBoToolbox and Smap FieldTask, through which workers can collect data offline on Android mobile devices then upload it using their office wireless network.
The office Wi-Fi may not be that reliable either, which means the software needs to be good at uploading or synchronising. “You need to be really careful, thinking about how it is all being logged, what has synched and what hasn’t,” says Amos Doornbos, the charity’s disaster strategy and systems director. “It’s all possible – it’s just stuff you have to pay attention to more than in other environments.”
The organisation also makes use of standard software such as Microsoft Office, due to its ubiquity, interoperability and because, despite Microsoft’s enthusiasm to move users to cloud computing versions, Office works well offline.
MSF teams up with Red Cross and Open Street Map
This doesn’t mean that aid charities eschew cloud computing techniques entirely.
Médecins Sans Frontières (MSF), which provides emergency medical care in more than 70 countries, collaborates with the UK and US Red Cross and Open Street Map on Missing Maps, which provides up-to-date mapping of disaster-hit areas, particularly to fill gaps.
“There’s nobody who knows where certain villages are in the heart of Congo,” says Jorieke Vyncke, MSF’s Missing Maps coordinator. Governments may be unable to collect information, and commercial services have no reason to – a deficiency that can be seen in Google Maps’ lack of coverage of some refugee camps.
“Offline capabilities are a big priority for us. We’re working in rural Ethiopia and the middle of DRC where we simply can’t take connectivity for granted”
Amy O’Donnell, Oxfam
On 8 May 2018, an Ebola outbreak was declared in the Democratic Republic of the Congo. To help fight it more effectively, requests by email and social media led to more than 850 Missing Maps volunteers using satellite images to map all the houses and buildings in a 23,500km2 area – equivalent to south-west England, from Cornwall to Gloucestershire, Wiltshire and Dorset – using Open Street Map’s Tasking Manager to allocate work.
Local volunteers, as well as staff from MSF and DRC’s health ministry, added hospitals and places where people were at high risk of spreading infections, such as markets. They held a workshop in Mbandaka, one of the two places central to the outbreak.
Although compiled digitally, the first maps took an old-fashioned route to their destination. The charity printed poster-sized versions in Geneva, which were taken by a geographical information system (GIS) officer on a flight to DRC.
“Two weeks later, the team on the ground had those maps hanging on their wall in the heart of Congo,” says Vyncke, allowing effective planning of the vaccination campaign.
Later, all the available Open Street Map data was physically brought in by another GIS officer on a small server, to avoid downloading it over expensive and sometimes unreliable local connections.
Cisco emergency networks
When aid work needs an internet connection, software can be used to prioritise its use. Cisco provides temporary emergency communications free of charge to areas hit by disasters, usually through a satellite link with 10-20Mbps capacity – equivalent to the download speed of a standard broadband service in the UK.
As mobile networks are often out of action, such a link will often carry voice calls, which will normally be prioritised. But Ron Snyder, solutions architect for Cisco tactical operations, says decisions over traffic prioritisation have to be made in discussion with those using the system.
Cisco set up an emergency link for a multifunction centre used by police and provincial government in Borongan in the central Philippines following a typhoon.
“We saw that a lot of traffic was going to YouTube, so we let the centre coordinator know,” says Snyder. “What was happening was that most of the people had not seen news of what had happened in other parts of the country.” What looked like a waste of bandwidth was actually a useful source of information, and as a result the team did not place specific limits on it.
Software design important
Software can be used to research the needs of aid recipients, but it needs to be designed for the task.
Elrha, a UK-based charity which funds innovation in humanitarian work, approached Pivotal Act to better understand the principles of user-centric software design. It then commissioned research on why many children in refugee camps avoided using latrines, despite this increasing their risk of illness. Eclipse designed pictorial diagrams for tablet computer screens, allowing refugee children to point at where problems occur regardless of language.
The software, designed to work offline, has been used by aid organisations including Save the Children in Bangladesh with Rohingya refugees who have been driven out of Myanmar by that country’s military.
This allowed children to communicate their fears. “They couldn’t lock the toilet doors, so they were really scared when they were inside the latrines,” says Cecilie Hestbaek, innovation manager at Elrha. “The [latrine] hole was very big, so smaller children were really scared of falling into the holes. That’s not an unsubstantiated fear, as it does happen.”
The research led to locks being placed lower on doors, within a child’s reach, and adjustments to latrine slabs to reduce the size of the holes, as well as moving some latrines so children had less distance to walk to use them.
What works in London may not work in South Sudan
Hestbaek says technologists working in development have to collaborate and be willing to learn.
“We see a lot of tech innovation, especially apps, being brought to the field as if it can solve all the problems. We have a real issue with apps for refugees,” she says, with developers designing software with little or no input from the intended users.
In some cases, the best answer is not to develop an app. Elrha funded a project that aimed to build software to help refugees in Kenyan camps access training and education. However, preparatory research found that people already knew about online resources, but the bigger problem was that local mobile connections were expensive and of poor quality.
“Building that app would have been a waste of resources,” says Hestbaek. Instead, the project shifted to focus on improving mobile reception and prices.
“There’s a massive amount of technology out there, and a huge amount developed by people from a tech background wanting to do good, and yet they tend not to be developed by humanitarians or development workers in the context in which they are working,” says World Vision’s Doornbos.
What works in London may not work in South Sudan, and not just because of bandwidth limitations. A privacy policy that is only available by clicking on a link and is written in legalese English is not a good fit for charities such as World Vision and Oxfam, which aim to treat beneficiaries with respect and dignity.
“Value for money is very important to us as a charity. Free software is like free puppies – they start costing money very quickly”
Amy O’Donnell, Oxfam
Specialist requirements and limited budgets
As well as strong data protection standards, with Oxfam avoiding the use of biometric technology, O’Donnell says the charity looks for software with a simple and clean interface.
“There are some free open source tools we’ve struggled with in the past, because of maybe a harder user interface, where it requires some coding skills to get these things set up,” she says, which could require training and cost them time if they break. “Value for money is very important to us as a charity. Free software is like free puppies – they start costing money very quickly.”
Aid and development organisations demand a lot from software, with another requirement being the need to support multiple languages, including those using different character sets such as Arabic and Burmese.
Aid organisations and government departments, including the UK’s Department for International Development (DfID), support the Principles for Digital Development, which summarise the sector’s needs.
“Any digital solution should take into account local factors,” says a DfID spokesperson. “Common factors to consider are connectivity (mobile coverage, bandwidth, latency, power); devices (smartphones, feature phones or laptops); language; and education levels.”
The department says it is happy for organisations to select their own software, and recommends the use of open standards to enable data sharing.
Despite the specialist requirements, organisations are always under pressure to spend money efficiently, making it hard for them to justify the cost of developing their own applications. “Oxfam is not in a good position to be a tech provider. There are many people out there much better equipped than us to develop this sort of tech more efficiently, more effectively, more appropriately to people’s needs,” says O’Donnell.
Collaboration can provide an answer. Oxfam is one of more than 30 organisations, including Save the Children and Unicef, that use Last Mile Mobile Solution (LMMS), a suite developed and supported by World Vision International. World Vision Canada began developing the suite a decade ago, with support from the Canadian government, and it is now run as a self-funding business within World Vision International.
Initially intended as the aid beneficiary equivalent of an airline frequent flier scheme, LMMS’s functionality includes creating and managing digital identities; the distribution of goods and monetary aid, including cash and vouchers, both physical and digital; analytics and dashboards; and data sharing for multi-agency work.
Doornbos adds that LMMS is strongly focused on its customers’ needs, with – unsurprisingly – strong capabilities for working offline, even when managing the likes of digital vouchers. “From day one, it’s been built by humanitarians for humanitarians,” he says.
Read more about software used for international aid
Path health organisation is fuelling its anti-malaria programme in Zambia, and other African countries, with data analytics from Alteryx, Tableau and others.
