Enterprise Architecture Leader, Amazon Web Services
Andrew is a innovation driven Enterprise Architecture leader with 27 years of delivering solutions across Industrial, ISP, Telecoms and large scale utilities networks focusing on process and technology validation.
Working for many of the world’s leading companies to create solutions that make a difference globally.
“AWS IoT Defender ML Detect”
ML Detect is a feature of AWS IoT Device Defender, a services for auditing and monitoring devices connected to AWS IoT. Today, Device Defender customers can create static alarms with an existing feature named rule-based Detect to identify operational and security anomalies across seven cloud-side metrics (e.g. authorization failures, messages sent) and ten device-side metrics (e.g. packets-out, bytes-out).
Customers have told us that existing static alarms doesn’t work all issues because:
1. result in too many false positive alerts since they don’t adjust for seasonality and other changing factors and
2. are difficult to configure effectively because customers must understand how their fleet behaves across a variety of metrics.
ML detact solves this by using data from a trailing 14 day period to automatically set threshold for customers. It retrains the model each day to refresh the threshold vased on the latest trailing 14 days.
This training will go through the setup with an example to show its value andd how it can be used at scale.