AWS beefs up IoT services with AI, management, and security tools
摘要：Above: AWS CEO Andy Jassy speaks at the company's re:Invent customer conference in Las Vegas, Nevada on November 29, 2017.
Above: AWS CEO Andy Jassy speaks at the company's re:Invent customer conference in Las Vegas, Nevada on November 29, 2017.
Amazon Web Services continued to expand its capabilities for powering internet of things applications today with a suite of new services on its cloud platform. Customers should be able to send machine learning models down to edge devices, secure those devices, manage them, and analyze the data coming off of them more readily as a result.
These services help Amazon capitalize on a growing set of workloads that are well-suited for integration with cloud computing. IoT workloads often require the management of large fleets of devices, plus the storage of massive (and constantly growing) amounts of data. Those requirements make cloud computing platforms like AWS well-suited to the task.
A new Greengrass ML Inference service makes it easier for customers to deploy trained machine learning models to edge devices, so that it’s possible to make intelligent decisions without access to the cloud. Greengrass is AWS’s platform for running code on remote devices, while managing that process through the cloud, and this new feature makes it easier to deploy intelligent algorithms to the edge.
That’s important for applications that require low-latency decision making, or situations where developers expect inconsistent network connectivity. For example, writing a machine learning model to run on an oil rig wouldn’t be as useful if it requires a persistent network connection. The new Greengrass ML Inference feature should make that more possible.
AWS IoT Analytics helps tackle the flip side of that problem: bringing data from edge devices into the cloud for analysis and understanding. It’s a fully managed service that processes, cleans and stores data, making it available for further processing including by machine learning algorithms.
When it comes to managing and securing IoT devices, there are a trio of new services for customers to work with. The IoT Device Management offering will allow people to
In addition, a new AWS IoT 1-Click service will let developers connect a device to the cloud and load it up with a pre-defined code snippet built using the cloud provider’s Lambda service, with a minimum amount of effort. It’s a lighter weight form of deployment that’s designed for customers who want to get going fast.
A forthcoming IoT Device Defender service will watch over the edge devices that have been registered with AWS, and help keep them safe from attackers. It will enable continuous auditing to ensure devices are complying with security policies, real-time detection of anomalous behavior, and tools to mitigate and remediate attacks.
As part of the service, AWS will offer customers a set of premade best practice security policies to drive security. IoT devices have been a favorite target for hostile hackers, and this service could help customers secure what they have deployed.
All of these announcements come as part of AWS’s re:Invent conference for customers, taking place this week in Las Vegas. The company made a slew of announcements, including a set of new services designed to help democratize machine learning.