AWS Unveils Mithra: A Cutting-Edge Security System for Enhanced Cloud Protection
Amazon Web Services (AWS) has launched Mithra, an advanced large-scale security system designed to bolster the security of its cloud services. Mithra leverages an internal neural network graph model comprising 3.5 billion nodes and 48 billion edges to assess and rank the trustworthiness of domains and detect potential threats.
Given that a single AWS region can handle up to 200 trillion DNS requests daily, the scale of this security system is unsurprising. Mithra is capable of identifying 182,000 new malicious domains each day.
Mithra’s system assigns a trustworthiness score to each domain queried within AWS, eliminating the need for third-party reliance and mitigating potential supply chain threats. This process helps AWS maintain a comprehensive and high-quality list of malicious domains to track and understand cybercriminal behavior.
AWS claims that Mithra’s massive neural network graph might be one of the largest in existence, offering high accuracy in detecting malicious domains. It can even predict malicious domains well before they appear on the threat feeds of third-party security providers.
Mithra continuously updates a list of malicious domain names for security providers to monitor and block. This list can also be shared with third-party threat feeds to reduce false positives, while Mithra’s detection of millions of security events daily provides AWS security analysts with valuable data for investigations and additional context.
Organizations receive notifications from Mithra about potential cyber-attacks from malicious domains, along with response recommendations. These recommendations include checking security logs for activity from specific domains, blocking them, and moving infrastructure behind a firewall.
