Evolve logo オンデマンド配信 Evolve 2023 Tokyo|AI、データと分析の先進事例を紹介
  • Cloudera Cloudera
  • FireEye logo
    60% increased research productivity

    Key highlights




    Milpitas, California, USA

    Solution highlights

    • Modern Data Platform: Cloudera Enterprise
    • Workloads: Data Warehouse, Data Science and Data Engineering
    • Components: Apache Hive, Apache Impala, Apache Spark™

    Applications supported

    • Dynamic threat intelligence


    • Helps threat hunters obtain responses to queries magnitudes faster
    • Provides access to a wider range of data that wasn’t accessible before
    • Increases researcher productivity by 60 percent

    Big data scale

    • Terabytes of data from over 16 million virtual analyses per hour

    FireEye improved productivity and visibility into available data, both key in its ability to detect zero-day attacks and stop advanced persistent threats earlier, using a modern analytics platform from Cloudera.

    FireEye is an intelligence-led security company that helps organizations better prepare for, prevent, and respond to cyberattacks. FireEye serves more than 5,600 customers across 67 countries.


    In just three years, FireEye’s threat intelligence team watched as their data volumes exploded—growing over 25 times. The team’s dynamic threat intelligence database, which combines malware information collected from FireEye customers with malware analytics, helps FireEye threat hunters identify zero-day attacks and stop advanced persistent threats before they can accomplish their mission.

    As the amount of data grew, the team found its relational database could not easily scale, and response time to queries slowed.  “Our threat intelligence data must be readily available,” said Alex Rivlin, software development manager, FireEye. “We came to the point where data access became a challenge using traditional technologies.”


    FireEye worked with Cloudera to build an analytics platform to improve analysis required for faster detection of zero-day malware and advanced persistent threats. The platform supports terabytes of data collected from more than 16 million virtual analyses per hour.

    Our researchers sometimes need to query data over a year, or run queries across multiple customers. Working with Cloudera, all the data we have is now available and easily accessible.

    -Ganesh Prabhu, Staff Engineer, FireEye

    FireEye threat hunters can also more easily apply machine learning to identify new attacks and campaigns. “Analytics and machine learning are core components of malware protection,” said Rivlin. “The Cloudera platform makes it easier to enable standard machine learning libraries and will free researchers from writing extractors and adapters to collect the data.”


    With cybercriminals continually launching new attacks, FireEye wanted to move quickly in implementing its new platform. “Cloudera helped us meet a very aggressive timeline,” said Prabhu. “We could not have done that without their help.”

    Because Apache Hadoop was new to the team, training was an important component of the company’s implementation plan. “Hadoop offered the best fit due to SQL access patterns and the ability to scale horizontally,” said Rivlin. “However, our team is skilled in relational database technologies and standard software application development. We needed to learn Hadoop tools and best practices. In four short days, Cloudera training gave us the jumpstart we needed and made our implementation easier than it would have been otherwise.”


    The new platform improved productivity and visibility into available data, both key in the organization’s ability to detect zero-day attacks and stop advanced persistent threats earlier. For example, threat hunters obtain responses to queries magnitudes faster, and they can access a wider range of data that simply wasn’t accessible before.

    “We are opening a new door for security research,” said Prabhu. “Before, researchers spent about 50 percent of their time collecting data. Now, more than 80 percent of their time is focused on research. Additionally, we can process two years of data and identify impacts that were practically not feasible with our legacy platform.”