Microsoft Azure Helps Academic Researchers Accelerate Discovery
It’s just as important to analyze the past as it is to understand the causes of natural occurrences. That’s because insights can lead to knowledge and actions that help reduce the risk of destruction. Two academic researchers are using the power of Microsoft Azure to process and dissect this critical data.
Investigating climate change in the cloud
Jeff Dozier, professor of environmental science and management at the University of California, Santa Barbara, is ensuring that the next generation of scientists grasps how climate has changed and what has driven those deviations.
Studying those variations — both recent and those from long, long ago — during the course of a semester is challenging. Not only are there massive amounts of data, many charts, copious videos, numerous illustrations and a variety of other materials to cover, but presenting it across such an expansive timeline is difficult without the context of the big picture.
To present these materials in an understandable and flexible manner, Dozier paired ChronoZoom, an open-source development devoted to visualizing history, and Azure. In using these two tools, an extensive timeline was created, enabling users to explore the volume and depth of the information by zooming in and viewing important details in one period or another, and moving through the topic’s history at their own pace.
With both tools being cloud-based, Dozier’s graduate students can access the information from anywhere — they just need a Web browser. The data is processed in the cloud through the use of Windows Azure. By using Azure, they also have the ability to scale up or down.
“Climate changes for reasons, and the reasons can be different at different time scales,” says Dozier. Between ChronoZoom and Azure, he says, “The real innovation is in the way they’ve been able to visually depict and explore long periods of time as well as short durations within that same context of the big picture.”
How have the students benefitted from this approach? Dozier says that using this style of teaching ensures that they can educate others in an impactful manner. “Don’t take positions on an ideology. Use statistics and evidence to address disputes instead,” he teaches.
Assessing wildfire risk with Azure
When seconds can mean the difference between saving life and property, or the loss of it when wildfires erupt, data can help us predict the threat and spread of such disasters. Developed by the University of the Aegean in Greece, the VENUS-C Fire app, which leverages Windows Azure, Bing Maps and Microsoft Silverlight, assesses the day-to-day risk and propagation of wildfires on the susceptible island of Lesvos throughout the dry season.
In using this application, Lesvos’ fire management team can predict and optimize their resource allocation for the day based on emerging or potential fires. To process all the data that helps predict risk, a comprehensive cloud infrastructure was needed to allow users to access the VENUS-C Fire app from multiple devices and from anywhere with an Internet connection. The team chose Azure for its elasticity, scalability and efficiency.
“The algorithms of forest fire risk and fire behavior procedures in the VENUS-C Fire application were automated by using Microsoft Azure to speed up the calculations and promptly provide the outcomes to end users through the Web platform,” explains Kostas Kalabokidis, associate professor at the University of the Aegean. “In wildfire emergency situations with high workloads, Microsoft Azure can provide the means for high availability, high throughput and high accessibility before a disaster strikes.”
He says that Azure has been a critical component for them, as it offers the needed processing power and storage at a lower cost than on-premise solutions. Kalabokidis says the department is able to invest more in personnel, equipment and further research rather than in IT overhead.
Not only is this solution saving the department money, but it’s also saving valuable time. Between the new application and the old process, Kalabokidis says they’ve shaved two hours off the response time. “The response time for the creation of the 112 fire risk maps was 20 minutes with nine virtual machines available, while the response time for the serial execution was 140 minutes,” he says.
Numerous academic researchers are accelerating our ability to understand and leverage data with the help of Microsoft Azure. Is your institution ready to comprehend big data in the cloud? Talk to a specialist to learn how this solution will address your specific needs.