Profile: Florida State University

Florida State University is an elite research institution. Its 16 colleges offer 275 degree programs to more than 40,000 students.


The Challenge

Florida State University (FSU) prides itself on being one of the top idea incubators in the US. Its commitment to research is evident in FSU’s $200 million annual research budget and the 50 prominent institutes calling the university their home. It’s also clear in the charter of FSU’s Research Computing Center (RCC), tasked with providing the technology needed to support researchers’ projects.

“Databases are alien to most researchers,” said Paul Van
Der Mark, interim director of the RCC. “Everyone who uses MySQL complains it’s too slow, but they don’t know how to tinker with it and don’t want to learn. Some people resort to analysis-by-spreadsheet, but that’s extremely difficult and even slower.”

RCC’s staff had its own database challenges. Prasad Maddumage, an application specialist, had to wait up to four minutes each time he ran a query on his supercomputer job scheduling application, and he was running queries all the time. “I couldn’t get anything done while I waited around. It was very frustrating,” he said.

FSU needed a database solution that would speed projects without taxing researchers.

The Solution

While searching online for the latest database technologies, Maddumage discovered deepSQL and was intrigued by its performance claims.

deepSQL – the world’s only adaptive, MySQL-compliant database – is purpose-built for data-intensive, large-scale physical, virtual and cloud environments where fast load and accelerated queries are crucial. Leveraging new computer science, machine learning and self-tuning technologies, and using the MySQL interface businesses know, deepSQL delivers blazing speed, unprecedented flexibility and scales to hundreds of billions of rows, with no application changes, no trial-and-error configurations and much-reduced overhead.

Maddumage downloaded deepSQL and, he said, “The tests lived up to Deep’s promises. deepSQL’s performance was much better than MySQL with InnoDB.”

The RCC is now offering deepSQL to FSU researchers, targeting complex projects such as data mining business information for economic modelling and an astrophysics cry-electron microscope used to create 3D images.

The Results

Accelerates complex queries by 1,000%

With deepSQL processing queries faster than any other solution on the market, time-to-answer is expedited as never before. Using deepSQL, Maddumage ran queries on his 200GB database 100x faster. Complex queries that took 200 seconds with InnoDB were completed in just two seconds with deepSQL. Even simple queries were accelerated by 20 percent.

Eases analysis for researchers

Although RCC staff supported researchers who used databases when they needed help, they didn’t widely promote the use of databases, until deepSQL. According to Van Der Mark, “deepSQL is the one solution we’ve found that’s ideal for researchers who work with complex data sets but aren’t database experts. It’s incredibly easy to use and its auto-tuning does all the heavy lifting without going offline – all of which streamlines and speeds analyses, and thrills our researchers.”

Eliminates need for RCC training

deepSQL’s ease of use was also important for the RCC’s staff. If FSU had chosen a different database solution, they would have had to train RCC employees on it, which there wasn’t time or budget to do. That’s not the case with deepSQL, as Maddumage found. “Because I know MySQL, there was no learning curve, no new commands to learn. deepSQL just worked,” he said.

Enables database-as-a-service

FSU currently uses deepSQL on bare metal machines, but has the ability to run databases wherever and however it makes most sense. “We’re confident deepSQL can be a great help to our researchers – whether it’s used in physical, virtual or cloud environments. Thanks to deepSQL, we can now offer high-performing database-as-a-service at minimal cost,” said Van Der Mark.

 

“Not only does deepSQL run all types of queries much faster than InnoDB, it’s easy to use. If you know MySQL, you know deepSQL.”
— PRASAD MADDUMAGE
Application Specialist,
Florida State University

 
Back to Case Studies