AWS re:Invent kicked off today, and Amazon has already made a number of announcements, including product updates, performance improvements, and better integrations among services.
Here are some highlights from the event so far:
Cost Optimization Hub provides recommendations for cost saving
This is a new section in the AWS Billing and Cost Management console. It provides recommendations for how customers can optimize their billing and allows them to query how much will be saved by implementing each action.
Currently the tool provides six types of cost optimization actions: stopping idle resources, rightsizing, upgrading to a later-generation product, Graviton migration, saving plans, and reserved instances.
RELATED CONTENT: News from third-party providers out of AWS re:Invent
Amazon WorkSpaces Thin Client provides easy-to-manage, affordable virtual desktops
The Thin Client devices are pre configured devices that can run Amazon WorkSpaces that are shipped directly to an end user. Once received they don’t require additional setup steps and can start being used instantly.
The device itself is a small cube that can be connected to peripherals like a monitor, keyboard, and mouse. It currently costs $195, has several security precautions built in, is easy for administrators to manage, and easy for the end user to use.
Additionally, the company also announced support for cross-region data replication in Amazon WorkSpaces. Snapshots are taken every 12 hours and are replicated to the desired region. This enables a recovery point objective of 12-24 hours, the company explained.
Better integration between Amazon Bedrock and AWS Step Functions
Amazon Bedrock is the company’s generative AI application development platform, while AWS Step Functions is a workflow orchestration tool. Two new API actions were added to AWS Step Functions to be used with Amazon Bedrock.
InvokeModel can be used to “invoke a model and run the inferences with the input provided in the parameters.” CreateModelCustomizationJob launches a fine-tuning job, which can be used to customize a base model.
Amazon Detective adds new features for cloud security investigations
Detective investigations for IAM provides insights into AWS IAM objects — like users and roles — to detect tactics from MITRE ATT&CK, which is a knowledge base of malicious tactics used in security attacks.
Detective finding group summaries analyzes finding groups and provides insights in natural language, helping accelerate security investigations.
The company also announced support for investigations for threats discovered by Amazon GuardDuty ECS Runtime Monitoring, and integration with Amazon Security Lake.
AWS Lambda functions now scale 12 times faster
According to the company, each synchronously invoked function will now scale by 1,000 concurrent executions per 10 seconds, until the concurrency limit of the account is reached.
Previously, functions would scale by 500-3,000 concurrent executions in the first minute and 500 every minute thereafter until the account’s limit was reached.
Functions also now scale independently from other functions, no matter how they are invoked.
Amazon CloudWatch Logs updates
Several new features were added to this product, starting with automated pattern analytics and anomaly detection. Developers can analyze patterns and trends in logs, compare query results in a specified time range, and discover anomalies in logs.
Metrics can now be consolidated across hybrid, multi-cloud, and on-premises data sources, and then be processed in a consistent manner.
Amazon also introduced a new log class for infrequently accessed logs. The new Infrequent Access class is cheaper and allows customers to consolidate their logs in one place. According to the company, this new class is 50% cheaper per GB ingestion than the Standard log class. It also provides a tailored set of capabilities so that customers aren’t paying for features that wouldn’t be used.
Generative AI capabilities were also added, allowing natural language prompts to be used to query CloudWatch logs and metrics. This feature is currently in preview, and at the moment it provides three main capabilities: the ability to generate queries from a description or question, explanation of queries, and the ability to refine queries using guided iterations.
Amazon FSx gets several improvements
The company’s file server also got several upgrades, such as the ability to send snapshots between file systems in Amazon FSx for OpenZFS.
Amazon FSx for NetApp ONTAP was also upgraded with the ability to create Multi-AZ FSx in ONTAP file systems in shared VPC environments. This provides a clear separation of responsibilities between network admins and storage admins. Customers can also now create scale-out file systems, leading to a 9x improvement in performance.
And finally, the company added support for creating, managing, and backing up Amazon FSx for NetApp ONTAP FlexGroup volumes using the AWS Management Console, Amazon FSx CLI, and AWS SDK. Previously they could only be created in the ONTAP CLI or ONTAP REST API. FlexGroups can be up to 20 petabytes large.
