← AWS Podcast
AWS Podcast
#703: AWS News Update, January 13, 2025
00:21:04
It is the first show of 2025! Simon walks you through all the updates that happened during the holid
Loading summary
Transcript1 lines
- [00:00]
A
This is episode 703 of the AWS podcast released on January 13, 2025. Hello everyone and welcome back to the AWS Podcast. I'm Lee Shub with you. Great to have you back for 2025, a new year and a new set of hopefully really cool podcasts for you to listen to today. Now both of my co hosts are away today. One is on vacation, one is affected by the fires, so we hope that everyone's but the show must go on. So here it is. Let's start with the topic of the AWS Marketplace. There is a new look for AWS Marketplace Professional Services product details pages so you can get an updated and streamlined layout that lets you understand more about the particular services. And the AWS Marketplace now also supports self service promotional media on seller product detail pages so you can have videos, you can have images, lots more information available to you if you want to access the AWS Marketplace. A lot of people have a lot of success finding different solutions, different capabilities there and it also makes it easier to add it to your account. Let's talk about analytics. Amazon MSK Connect has added support for Apache Kafka Connect version 3.7, so several bug fixes and performance improvements. AWS Glue Data Catalog is offering advanced automatic optimization for Apache Iceberg tables, so this includes supporting compaction of deleted files, nested data types, partial progress commits and partition evolution support. AWS Glue has expanded connectivity to 16 native connectors for applications so things like Adobe Analytics, Asana, Datadog, Facebook Page, Insights, Zoom Meetings, Twilio, you name it, it's in there. And Amazon Quicksight has launched unique key for data set so this enables users to define additional aspects of their data semantics. The unique key will be used to improve performance for quicksight visuals, especially unaggregated table charts. So previously to maintain table pagination stability, all columns in the table visual were sorted which was an expensive query causing performance latency. Now with a unique key defined in the data set, once the column is used in the visual, users will automatically experience improved sorting performance without compromising user behavior. For some cases, the new approach can increase the performance by up to 60% so you can go faster to get your render quick update. For the topic of application integration, Amazon Eventbridge announces API Destinations proactive OAuth token refresh, so this allows you to preemptively mitigate delays and errors during invocations when OAuth access token is closed to expiration. So basically you get to renew it before it goes out of date. Now of course lots of updates in the world of artificial intelligence, Amazon Bedrock agents, flows and knowledge bases now offer support for the recently announced in Preview latency optimizer models via the SDK. This enhancement brings faster response times and improved responsiveness to AI applications built with Amazon Bedrock tooling. Currently this optimization is available for Anthropic's Claude 3.5 Haiku model and Meta's Llama 3.14,05B and 70B models, delivering reduced latency compared to standard models without compromising accuracy. Now this update is particularly beneficial for customers developing latency sensitive applications. Things like real time customer service, chatbots and interactive coding time is frustrating if you're waiting for it. By leveraging purpose built AI chips like AWS Trainium 2 and advanced software optimizations in Amazon Bedrock, customers can now access some more options to optimize their inference for specific use cases. Importantly, these capabilities can be integrated immediately into existing applications without additional setup or model fine tuning so you can get enhanced performance and faster response times. We're also happy to introduce stable diffusion 3.5 large in Amazon bedrock. So this is a powerful model that lets you generate high quality 1 megapixel images from text descriptions with superior accuracy and creative control. Meta's Llama 3.370B model is now available on Amazon Bedrock, so again continuously advancing the model efficiency and performance. And Llama 3.370 B is also now available on AWS via SageMaker, Jumpstart and Amazon Q Business Analytics Dashboard is enhanced with Conversation insights. So now you get new analytics dashboards and the administrators can view and analyze various aspects of the conversations like no answer found queries, blocked queries and end user feedback. So it helps you really create that feedback loop about how your customers are using Amazon q business and AWS. Neuron has introduced support for Trainium 2 and NXD inference. Now what this allows you to do is to basically simplify the deployment of large language models and multimodality models and it lets Pytorch model onboarding be a lot easier and simpler. And The Neuron Profiler 2.0 beta is a new profiler that enhances capabilities and usability which includes support for distributed workloads. Also new inference Support including llama 3.1405 et cetera. Lots of updates in AWS Neuron. Check it out. Let's talk about compute. We're happy to introduce Amazon EC2 High Memory U7INH instance. This is a new addition to the EC2 High Memory family built in collaboration with HP Enterprise. The instance runs on 16 socket HP compute scale up server 3200 and are built on the AWS Nitro system to give you a fully integrated experience consistent with other EC2 instances. These are powered by the fourth generation Intel Xeon scalable processor so Sapphire Rapids and you get 32 terabytes of memory and 1,920 VCPUs. Amazing. This is the highest performing high memory instance and it's really suitable for compute intensive in memory database workloads like SAP, HANA Amazon EC2 instances support bandwidth configurations for VPC and EBS so this is called instance bandwidth configurations or IBC and this is generally available for 8th generation graviton based instances. With IBC you can adjust the VPC and EBS service bandwidth by up to 25% on your EC2 instances. So when increasing the bandwidth for one of the services say VPC the available for the other service like EBS is reduced. It doesn't increase your instances burst bandwidth packets per second or IOPS capability, it just means you have more flexibility in output allocating your bandwidth. Bottlerocket now supports the Elastic Fabric adapter for AI, ML and HPC workloads, so this ties into the Elastic Fabric adapter and allowing it to support Bottlerocket which is the Linux based operating system purpose built for hosting containers. So this gives you more flexibility about how you set things up. Amazon EKS has introduced programmatic access to Kubernetes version availability. It has also expanded the catalog of upgrade insight checks as well. We're also happy to announce a node health monitoring and auto repair for eks so this allows you to be more resilient to errors whilst running and understand which nodes need to be replaced and fixing them automatically. An update for the Amazon Elastic Container Registry Amazon ECR we're happy to announce registry policy V2 which now supports managing IAM permissions for all ECR API applications. This new registry policy makes it easier for customers to control usage of ECR capabilities within their accounts and Amazon ECR now supports network fault injection experiments on AWS Fargate so this helps you improve your application performance, observability and of course resilience. There are now six actions supported for ECS on both EC2 and Fargate network latency, network black hole, network packet loss, CPU stress, I O stress and kill processes so you can do correct experiments to make make sure your application is resilient. Amazon LightSail API endpoints now support connectivity over IPv6 so if you want to use IPv6 you can AWS Parallel Computing Service PCs now supports Slurm version 24.05 so fixes security updates, improved auditing, some hardened security and enhancements as well, and aws Parallel Cluster 3.12 is now available with custom build enhancements, so this makes it possible to include Lustre and Nvidia software components in your parallel cluster. Custom images and AWS Elastic Beanstalk now updates the environment status when invalid, so it allows you to manage it more easily if things go weird in an update failure Quick update On the topic of cost management, we're happy to announce the general availability of Custom Billing View, which is a new feature within AWS billing and cost management that lets customers grant member accounts access to cost management data spanning multiple accounts within their organization. Custom BillingView allows you to provide application and business unit owners with access to relevant cost management data across multiple AWS accounts using a single view in the AWS Cost Explorer without granting access to the Management console. So with custom billingView you can create filtered views of cost management data based on your cost allocation tags or specific AWS accounts and then you can share those views with member accounts using Resource Access Manager, which means that you can allow teams to easily easily see what their costs are. Let's move on to customer engagement. The simple email service SES Mail Manager now supports full lifecycle logging, so this is for ingress endpoints and rules engine actions. You can configure a wide range of monitoring options to CloudWatch, to S3 and of course to Firehose as well a lot of updates. For Amazon Connect it has now launched support for routing to a specific range of agent proficiencies, the ability to exclude certain proficiencies during routing. It has built in capabilities for customer authentication within chats. It now supports the deletion of queues and routing profiles from the admin website. It now supports multi party chat so you can have four additional agents join an ongoing chat conversation to bring the right expertise to bear. It now supports durations of up to 30 days for tasks, so this is up from the default of seven days. And it now provides an agent schedule data in Analytics Data lake so you can understand how your agents are being allocated. So lots of updates there. Let's move on to the topic of databases. Amazon RDS for MySQL supports innovation release version 9.1 in the Amazon RDS database preview environment. So again this is an environment that you can experiment with and understand what new database features make sense for your particular application and this helps you be more confident in your rollout plans. Now this particular release includes bug fixes, security patches as well as a whole bunch of new features, so it's definitely something to have a look at and it's really easy to do in that environment. Speaking of updates, Amazon aurora now supports PostgreSQL 16.6, 15, 10, 14, 15, 13, 18 and 12.22 and you'd be disappointed in me if I didn't remind you to patch your stuff. So always keep your databases up to date because you get the best of features, improvements, bug fixes and performance enhancements. Amazon RDS for DB2 now supports multiple databases within an instance, so this gives you more flexibility in managing your databases and optimizing your licenses for various enterprise workloads. In fact, now you can manage up to 50 DB2 instances within a single RDS for DB2 instance, so you don't have to deploy separate licenses for each EC2 instance, et cetera, et cetera. Amazon RDS for MariaDB now supports minor versions 1144, 10, 11, 10, 10, 6, 20 and 10527 and Amazon RDS proxy has announced caching SHA2 password authentication support for MySQL on Aurora and RDS and we're happy to announce NVMe backend R6GD instances for Amazon DocumentDB with MongoDB compatibility. This gives you up to 7 times faster query performance for applications with large that exceed regular instance memory, and these instances can expand the data caching capability by up to seven times compared to a regular instance. So by using that non volatile memory express based SSD storage you get some good performance. AWS DMS the Database Migration service now supports Kerberos authentication for Oracle and SQL Server source endpoints, and Amazon timestream for InfluxDB now supports IPv6 connectivity. Let's talk about developer tools. The Kinesis Producer library now supports the AWS SDK for Java 2. X for improved performance and security. This new version you can fully remove the dependency on the old Java 1. X SDK and you can upgrade all of your applications running the previous Kinesis Producer library versions to the latest library without a change in your core data processing logic. Let's have some updates around the world of end user computing. Amazon AppStream 2.0 has introduced rocky Linux application and desktop streaming, and Amazon AppStream 2.0 has also introduced a client for macOS and AppStream wasn't done with their updates. Users can now save their user preferences between streaming sessions, so this gives you a nice consistent experience between sessions when you connect and reconnect. And Amazon Workspaces Personal now supports the AWS Global accelerator, so this works to enhance workspaces connection performance by optimizing streaming traffic through the AWS global network and Edge locations. So this is really useful if you're accessing your workspace across long distances. A quick update in the world of front end and web and mobile Amazon Location Service now supports IPv6 Lots of IPv6 support coming this episode, so if you're an IPv6 head you'll be excited onto the Internet of Things AWS IoT Device Management introduces high throughput device connectivity status queries so this lets you query the latest connectivity State of IoT devices for monitoring and managed purposes. Now understanding the connectivity status is crucial for understanding device failures and executing remote commands. The new Connectivity Status API which will be available to IoT Device Management Fleet indexing customers gives you a high throughput solution so 350 plus requests per second to understand device connectivity to the cloud. So this is really useful in terms of understanding what's going on in your fleet, particularly really large fleets. AWS IT Greengrass version 2.14 now supports a new lightweight Edge runtime software which uses less than 5 meg of memory. So this is the new Nucleus Light feature that lets you have access for resource constrained devices which operate on embedded Linux. Now this is offered alongside the original AWS IoT greengrass nucleus so you can have the choice of what you want to use, but this is really focused on that low RAM configuration environment. Let's talk about management and governance. AWS Systems Manager for SAP now supports SAP Netweaver based ABAP APPLIC deployed in distributed and HA topologies so you can get up and running faster. AWS CloudTrail now supports IPv6. There's another one and AWS config now supports three new resource types. We're also happy to announce cloudformation support for the AWS Parallel Computing service and the AWS Resilience Hub now supports Amazon CloudWatch alarm detection for application resilience. So this new capability combines the Resilience Hub recommendations with your current monitoring setup to streamline alarm management and enhance assessment accuracy. It's all about figuring out what's happened as quickly as possible. And the AWS Resilient Hub has also now provided tailored AWS fault injection service recommendations. So when you onboard an application to the Resilience Hub, you'll get customized experiment recommendations based on the specific application context. Which is pretty cool. Means you're not going from scratch. You have a starting point. We're also happy to announce notification actions in the AWS Console mobile app for iOS. You can now take actions like view logs, start, stop and reboot an EC2 instance on an event notification to shorten incident response, diagnosis and resolution time so you can get moving quicker when you're on the go. And the AWS Resource Explorer now supports 59 new resource types. And no, I won't read them out to you, you can check them in the show. Notes one Update for Media Services AWS Elemental Media Connect Gateway now supports source specific multicast, so this allows you to specify both a multicast group address and an optional source IP address when creating or updating your Media Connect Gateway bridge sources. It's all about flexibility. Let's talk about migration and modernization. AWS Application Discovery Service now supports guess what? IPv6 endpoints AWS DataSync now supports configuration updates for all supported storage locations. So these are things like S3EFs, FSX for lustre, ET cetera, et cetera. With this enhancement you can update storage locations directly, removing the need to delete and create new locations in order to update key parameters. And the AWS Transfer family provides additional controls over AS two message attributes, so these are inbound and outbound applicability statement to AS2 messages, which helps you integrate and be more compatible with a broader range of your trading partners AS two implementations without having to ask them to make any changes. Let's talk about networking and content delivery. There is now IPv6 compatibility for AWS Secrets Manager VPC endpoints, so another tick there for those IPv6 folks. And AWS Cloud Map now supports Service Level attributes so you can now store and access information directly at the service level. A few updates in the world of security, identity and compliance, AWS WAF Console has added new Top Insights visualizations. So with this feature, customers with CloudWatch logging destinations can view a new Top Insight sections within the All Traffic Dashboard so you can quickly get information about what's going on in your environment. The IAM Roles Anywhere Credential Helper now supports TPM 2.0 and the AWS Network Firewall now supports you guessed it, IPv6 service endpoints. And let's have a few updates. In the world of storage, AWS Backup now supports dual stack environments so IPv4 and IPv6. And AWS backup has also launched support for Search and Item Level Recovery, so this allows you to search the metadata of your backups so you can find specific files or objects across your backups and recover up to five items at a time, which means you get faster recovery times. And Amazon EBS has launched Local Snapshots for AWS dedicated local zones, so now you can create those in your local zone environment. And finally, if you use the Recycle Bin, which is a data recovery feature that enables restoration of accidentally deleted Amazon EBS snapshots and EBS backed amis. Not that you'd ever make a mistake, you now have the option to use, you guessed it, IPv6 addresses for new and existing endpoints. So a whole bunch of updates have happened since we took our little break over the holiday season. I hope you had a time to catch up. If you haven't caught up on the Re Invent episodes, I recommend you do. It's a great summary of everything that happened over there and we have some great episodes coming up. Lots of deep dives, lots of new information coming this year, so keep an eye out for it. As always, we do love to get your feedback. AWS podcast is the place to do it and until next time, keep on building.