Let us have a look at this week’s key updates from AWS!

  • New EC2 G4 instances which optimise machine learning.
  • Amazon Transcribe now supports KMS encryption.
  • AWS Data Sync now supports all Amazon S3 storage classes.
  • Amazon Redshift now has automatic workload management and query priorities.

New EC2 G4 instances

AWS has just announced new G4 instances which optimise machine learning and graphic-intensive applications. So these instances has GPU enabled in the User Graphics processing unit rather than the regular CPU. In this G4, they use the latest generation NVIDIA T4 Tensor core GPUs which is designed specifically for Machine Learning and Graphics. These are going to be great use cases for activities like image classifications and speech recognition as well as graphic applications like game streaming and video transcoding. It is the fastest instance type that you can use for AWS Machine Learning Workloads.


Amazon Transcribe supports KMS encryption

Transcribe is an automatic speech recognition service which enables you to add speech to text capability to your applications and save the output to a text file. Now onwards, you can use your encryption keys using KMS to encrypt the output from Transcribe when storing in S3. This gives you a lot more flexibility and control around how you are protecting your Transcribe file and also, it is supported in all regions where Transcribe is available at no additional cost.


AWS DataSync supports all S3 classes

DataSync is a data migration service which automates and accelerates the transfer of data between NFS or SMB file service to AWS services like S3 and Elastic file system (EFS). So you can now transfer your data directly to any S3 storage class and select the most effective storage class for your data. This also introduces additional controls to prevent accidental overwrite as well as a data integrity check to verify your data after the transfer has been completed.


Automatic Workload Management for Redshift

Amazon Redshift now has automatic workload management and query priorities. RedShift is a fully managed data warehouse which allows you to analyse your data using standard SQL as well as your existing business intelligence tools. They have now introduced automatic workload management which uses machine learning to dynamically manage memory usage and concurrency to maximize query throughput and performance. Also, you can now easily set the priority of your most important queries, so that the critical workloads will get preferential treatment over lower-priority loads even when hundreds of queries are being submitted.