Historically, humanity has not been great at protecting our oceans. One out of every five fish sold – worth a staggering $23.5B USD – is illegally caught, and nearly 80 percent of our fish stocks are being fished at their maximum sustainable level or beyond. Not only is this behavior putting entire species of fish at risk, it threatens the entire oceanic ecosystem – an ecosystem that’s critical for producing the oxygen we breathe every day.
But, it’s the ocean – it’s vast. Its incredible size makes it very difficult for the authorities to detect and identify suspicious fishing activity fast enough to stop it. At least, it was, until OceanMind came along.
Protecting the oceans with technology
OceanMind is a UK-based non-profit dedicated to ending illegal fishing and protecting the oceans. Unsurprisingly, many illegal fishing operations also engage in other criminal activities such as human trafficking and the plundering of marine heritage sites, so the more illegal fishing operations OceanMind can help stop, the better.
OceanMind’s approach is simple on paper but complex in practice: use AI and billions of data points from around the world to identify illegal fishing operations for the authorities. By developing AI models to replicate the work of human intelligence analysts, OceanMind could scale their impact while keeping costs manageable. With such a huge goal and a very small team, OceanMind knew they had to be efficient, but scaling their on-premises solution to meet the growing global challenge would have required a significant investment in hardware and personnel to manage it. Additionally, OceanMind’s on-premises servers had their own limitations – OceanMind had to process data in batches, could only analyze data for specified locations, and had to pay very high operating and maintenance costs to keep the servers running. While OceanMind’s on-premises IT was an impressive solution with many valuable capabilities, it was not an efficient platform to scale their impact globally – to reach their goal of near-real-time, world-wide analysis, OceanMind knew they would have to migrate their operation to the cloud.
Selecting a cloud and finding a partner
When it came time to select a cloud provider, OceanMind carefully assessed their options and then selected Microsoft Azure due to its reputation for scale, reliability, and security. Because OceanMind works closely with governments and law enforcement agencies from around the world, maintaining compliance with data regulation laws was critical. Thankfully, Microsoft Azure boasts over 90 compliance certifications, over 50 of which are specific to global regions and countries, making it the ideal cloud provider for international operations like OceanMind.
To facilitate their migration to Microsoft Azure, OceanMind partnered with endjin: a UK-based consultancy with deep expertise in Microsoft Azure, data, AI, and solving complex software engineering problems. As a Microsoft partner with gold competencies in the Microsoft Partner Network for Cloud Platform, Data Platform, Data Analytics, and DevOps, endjin helps small teams achieve big things. Naturally, with this blend of skills and experience, they were an ideal partner for OceanMind’s global ambitions and cloud migration.
Making the transition to the cloud
To kickstart OceanMind’s digital transformation, endjin applied their usual iterative delivery process, identifying and addressing potential risks early on. They also provided the evidence, documentation, and other materials needed for OceanMind’s development team to ramp up and participate in the migration. Because endjin works with their partners remotely, they were easily able to coordinate OceanMind’s various stakeholders, ranging from data scientists in Cambridge to analysts and developers in Oxford.
Because of their sizable data processing platform and numerous machine learning models, OceanMind’s migration was challenging. To meet their needs, endjin helped them reengineer their 65-server, on-premises solution for Microsoft Azure. By using the latest zero allocation memory features of the C# programing languages, endjin was able to increase OceanMind’s data processing from seven thousand messages per second to over seven million. By deploying that into Microsoft Azure PaaS and other serverless products, OceanMind was able to improve reliability and significantly reduce capital costs.
OceanMind’s new solution utilizes numerous Microsoft Azure services including Azure Data Factory, Azure Data Lake, Azure Functions, Azure API Management, Azure DevOps, Azure Application Insights, Azure Key Vault, and more. Alongside endjin, Microsoft also supported this migration via the AI for Earth Grant, a dedicated Customer Success Manager, and support from additional product groups including Microsoft Azure Functions and Cosmos DB.