Let us have a look at this week's AWS updates:
AWS Step Functions now support Private Link
This feature will allow you to access SEP functions without using the Public Internet. Step Functions enable you to build applications by stitching together services like Lambda, Fargate, and SageMaker to create workflows where the output of one step becomes the input to the next. So by using a private link, you can now access Step-Functions while keeping your network traffic within the Amazon network using VPC endpoints. Doing so, your network architecture significantly simplifies because you no longer need to use an Internet Gateway or NAT Gateway to connect these functions.
Deep Learning containers include TensorFlow 2.0 support.
AWS Deep Learning containers are now available with TensorFlow to support. Deep learning containers are docker images which come pre-installed with deep learning framework to make it easy to deploy machine learning applications using SageMaker, Elastic Container Service or Elastic Kubernetes services. By this new release, the TensorFlow team has taken on feedback from the machine-learning community as a whole. They have worked to make TensorFlow much easier to use for both beginners and experts with simple API's as well as better Python integration.
Enhancement for moving data between Amazon FSX for Lustre and S3
The FSX is a high-performance file system suitable for machine learning and high-performance computing workloads and the file system is backed by S3. So there are two announcements- firstly, you can now synchronise file data permissions and even symbolic links for files transferred from FSX to S3. Secondly, they have quadrupled at the speed at which file metadata is imported from S3. This means you can now launch an FSX file system four times faster than before.
AWS System Manager Change Calendar
AWS System Manager is a service which gives you visibility and control of your infrastructure in AWS. AWS has introduced the AWS System Manager Change Calendar, a feature which lets you schedule calendar events to ensure changes are allowed or blocked at the right time. If there are huge business promotions or events where you expect a high demand for your resources, then you can automatically block any changes from occurring until that time has passed and it can be done programmatically. For example, before an AWS Lambda function is invoked, which may change some resources, it can automatically check to see whether this is the appropriate time to perform so. If not, it won't execute.