Welcome to episode 238 of the Cloud Pod Podcast – where the forecast is always cloudy! This week we’re bringing you a preview of Amazon re:Invent 2023. We’re talking all things AWS, Bedrock, Q, and frugal architecture, and – you guessed it – AI.
Titles we almost went with this week:
🪨 Amazon Builds on Bedrock with Q
🏗️ You Need to Be All Frugal Architects
A big thanks to this week’s sponsor:
Foghorn Consulting provides top-notch cloud and DevOps engineers to the world’s most innovative companies. Initiatives stalled because you have trouble hiring? Foghorn can be burning down your DevOps and Cloud backlogs as soon as next week.
“Pre”:Invent
Is it just us, or is a lot of the stuff released during pre-invent stuff that would have been main stage just a few years ago?
01:48 Major Items
Introducing Amazon CloudFront KeyValueStore: A low-latency datastore for CloudFront Functions
03:43 📢 Ryan – “I found this being announced pre-invent to be kind of shocking, because this is one of those announcements where you could re-architect your entire app for better performance using this type of solution, and it’s not even big enough for the main stage. But there’s huge potential in doing that edge transformation so that you can directly serve at the edge at much lower latency. So it’s awesome.”
Announcing AWS Console-to-Code (Preview) to generate code for console actions
*No Terraform yet, but hopefully that will come soon!
05:18 📢 Jonathan – “I think it’s great for learning too, actually. I mean, I use this in the Google console all the time because I try and put together a command line to do something and it fails miserably. And so I go and do it in the console and it generates the command line coding thing. Ah, I missed that thing, which isn’t documented anywhere.”
07:23 Storage
Optimize your storage costs for rarely-accessed files with Amazon EFS Archive
FlexGroup Volume Management for Amazon FSx for NetApp ONTAP is now available
New – Scale-out file systems for Amazon FSx for NetApp ONTAP
Introducing shared VPC support for Amazon FSx for NetApp ONTAP
Announcing on-demand data replication for Amazon FSx for OpenZFS
New – Amazon EBS Snapshot Lock
Automatic restore testing and validation now available in AWS Backup RL(Maybe?)
08:56 📢 Ryan – “that’s the main reason why I flagged this is that I’ve just done so many tabletop exercises and so many, you know, compliance evidence sessions where you’re screen sharing this restore process and it’s just so painful. And so it’s, I love the fact that this can be automated and sort of just, you know. completed and at that point it’s just sort of monitored. You have this part of your release process where you run through this process and it’s a pass fail on your application and you can respond to it there. And I think that’s a fantastic way of sort of signifying your compliance and your ability to be able to survive a disaster or unintentional failure.”
Amazon EBS Snapshots Archive is now available with AWS Backup – MK My PFR from 2 years ago
10:08 📢 Matthew – I think it was right when they implemented the – they had it for EFS to go down to archive and a few other things. So it was something that I requested for, and we tried to set it up when AWS backup first came out. And I think it was like my PFR about two years ago for a client. So as soon as I saw this release, I pinged my old client and was like, hey, go do this. This will save you lots of money.”
Replication failback and increased IOPS are new for Amazon EFS
11:06 Serverless & Step Functions
AWS Lambda functions now scale 12 times faster when handling high-volume requests JB
*This definitely would have been on the main stage not too long ago.
10:08 📢 Jonathan – “For me, scaling up to a tone faster is great, but I think there’s probably very few customers that would really be impacted by not being able to scale up faster than 3,000 every minute or whatever it was. I think for me, what stood out as being an important change was isolating the functions in their own scaling patterns so that the 3,000 per minute, whatever it was, is not across all the functions you have deployed in an account, it’s per function.”
External endpoints and testing of task states now available in AWS Step Functions
14:26 Finops
New Cost Optimization Hub centralizes recommended actions to save you money
17:15 AI/ML
18:28 📢 Ryan- “I really like the infrastructure’s code support for this. I’ve been using Code Whisperer, my personal projects for coding things, to make up for my shoddy code ability. And now it can make up for my shoddy infrastructure’s code ability, which I really enjoy. So this is awesome.”
Use natural language to query Amazon CloudWatch logs and metrics (preview) – MK
19:18 📢 Matthew – “That’s why I’m so excited for this is, you know, trying to figure out the exact syntax of log of cloud watch log insights. And you know, all these other things, you know, is always a pain in the neck. And you’re like, okay, what are the fields called and everything else to make sure I get it right. And okay, exactly all the things here, at least, you know, you can just tell it what you want and get me 80% of the way there. I can tweak from there to get me what I want. And having that in cloud watch logs is great because you forget that debug mode and leave it on into production. Now you’re like searching for the needle in the haystack, which definitely has never happened to any one of us.”
Amazon Transcribe Call Analytics adds new generative AI-powered call summaries (preview)
Build generative AI apps using AWS Step Functions and Amazon Bedrock
Build AI apps with PartyRock and Amazon Bedrock JPB
22:54 📢 Ryan – “It’s so much redemption. They took the visual aesthetic of 1980s Macintosh OS and they put it into a modern web application and I just feel like every one of my BS little things that I’ve done is just completely validated now and it didn’t matter that it was terrible and hacky and annoying. Thank you, Amazon.”
22:52 Observability
New Amazon CloudWatch log class for infrequent access logs at a reduced price
Amazon Managed Service for Prometheus collector provides agentless metric collection for Amazon EKS
Amazon CloudWatch Logs now offers automated pattern analytics and anomaly detection – RL
Use Amazon CloudWatch to consolidate hybrid, multi-cloud, and on-premises metrics – JPB
25:12 Containers
Amazon EKS Pod Identity simplifies IAM permissions for applications on Amazon EKS clusters
Detect runtime security threats in Amazon ECS and AWS Fargate, new in Amazon GuardDuty
Introducing Amazon GuardDuty ECS Runtime Monitoring, including AWS Fargate
26:23 Security
IAM Access Analyzer updates: Find unused access, check policies before deployment
Introducing IAM Access Analyzer custom policy checks
Amazon Detective adds new capabilities to accelerate and improve your cloud security investigations
Amazon Inspector expands AWS Lambda code scanning with generative AI powered remediation
Amazon Inspector agentless vulnerability assessments for Amazon EC2 now in preview JB
AWS Control Tower adds new controls to help customers meet digital sovereignty requirements
27:51 Misc
Introducing Amazon EC2 high-memory U7i Instances for large in-memory databases (preview)
New Amazon WorkSpaces Thin Client provides cost-effective, secure access to virtual desktops
Announcing cross-region data replication for Amazon WorkSpaces
Increase collaboration and securely share cloud knowledge with AWS re: Post Private
Use anomaly detection with AWS Glue to improve data quality (preview)
Check your AWS Free Tier usage programmatically with a new API
Getting started with new Amazon RDS for Db2 JPB
*If this makes Db2 popular again, please avoid Justin.
Announcing throughput increase and dead letter queue redrive support for Amazon SQS FIFO queues
Manage EDI at scale with new AWS B2B Data Interchange
And that’s it for pre-invent! A handful of years ago this would have been the whole conference, so it will be interesting to see what *actually* made the stage!
32:51 Monday Night Live with Peter Desantis – Infra Keynote
35:10 📢 Justin – “The whole concept of his talk was the journey to serverless and really managed services in general. And he talked too, you know, nice dig at Azure at one point where he said the only true isolation is a hypervisor at level high isolation, which is a dig at Azure, of course. And then, you know, but he basically started talking about RDS and Aurora in particular.”
Amazon ElastiCache Serverless for Redis and Memcached is now available
Join the preview of Amazon Aurora Limitless Database
37:11 📢 Matthew – “The Aurora Limitless is just fascinating. Like they’re literally taking care of so much stuff that you have to deal with, you know, with, okay, cool. This is now too large. And, you know, the concept of sharding databases just goes away, which is phenomenal because I have, I’ve helped so many people through, ‘okay, we’re going to start to shard in order to scale past this because we can’t handle it one database, even with read-only replicas and everything else’. And the fact that it’s all just taken care of. And the technical way of how they solve those problems is really – like Ryan said – just fascinating. The little problems I have versus the amount of stuff that they deal with; it’s amazing that AWS stays functional, you know with the complexity level that they are dealing with.”
Reserve quantum computers, get guidance and cutting-edge capabilities with Amazon Braket Direct
44:30 Announced but not mentioned
Amazon CodeCatalyst introduces custom blueprints and a new enterprise tier
AWS Step Functions Workflow Studio is now available in AWS Application Composer
HashiCorp at re:Invent 2023: A year of collaboration with AWS
45:10 Tuesday’s Keynote – Adam Selipsky
Announcing the new Amazon S3 Express One Zone high performance storage class
37:11 📢 Matthew – “It just amazes me that they keep finding places and ways to innovate on S3. The service has been around for, I don’t want to try to make up a number, but enough years where I’m like, okay, it’s stable, like they’re good. And every year there’s always, you know, one announcement. I feel like, you know, it’s object locking. It’s this, you know, it’s just like these small features that just keep making. This core, crock service to pretty much everything else in AWS, just slightly better.”
Join the preview for new memory-optimized, AWS Graviton4-powered Amazon EC2 instances (R8g)
51:16 AI/Bedrock
Customize models in Amazon Bedrock with your own data using fine-tuning and continued pre-training
54:33 📢 Ryan – “The guardrails is super cool too, just because I think that, um, you know, this is, you know, the first thing I always think of, and I don’t know what, what is wrong with my brain, how I was dropped as a child. But the first thing with new technology is how can I break it? Um, and, and so this is one of those ways that, you know, like putting these guardrails in place so that you can very easily, you know, prevent the, you know, violent content or inappropriate content or what have you, wherever you want to put your guard rails into that, which is just making that an easy button. I think it’s super, super cool. And because it’s kind of a difficult challenge with using just basic AI tools to do that. Right. How do you safeguard against that data? How do you clean your data set enough?”
56:06 Q Continuum – A Hackathon project turned into production
Amazon Q brings generative AI-powered assistance to IT pros and developers (preview)
Improve developer productivity with generative-AI powered Amazon Q in Amazon CodeCatalyst (preview)
Upgrade your Java applications with Amazon Q Code Transformation (preview)
*allegedly
Introducing Amazon Q, a new generative AI-powered assistant (preview)
New Amazon Q in QuickSight uses generative AI assistance for quicker, easier data insights (preview)
Basically, Q has taken over the world – whether you (or us) like it or not.
1:04:10 Zero ETL
Amazon DynamoDB zero-ETL integration with Amazon OpenSearch Service is now available
AWS announces Amazon RDS for MySQL zero-ETL integration with Amazon Redshift (Public Preview)
AWS announces Amazon DynamoDB zero-ETL integration with Amazon Redshift
AWS announces Amazon Aurora PostgreSQL zero-ETL integration with Amazon Redshift (Public Preview)
Amazon Connect provides Zero-ETL analytics data lake to access contact center data (preview)
Apparently this is a theme – we’re going to see Zero-ETL everywhere. It’s either really easy to integrate, or they’ve been working on it for quite some time. I guess we’ll see…
1:07:09 Swami Keynote (AI)
We aren’t spending a lot of time here…feel free to peruse the links.
Vector search for Amazon DocumentDB (with MongoDB compatibility) is now generally available
Vector engine for Amazon OpenSearch Serverless is now available
Amazon Bedrock now provides access to Anthropic’s latest model, Claude 2.1
Announcing Amazon OpenSearch Service zero-ETL integration with Amazon S3 (preview)
Analyze large amounts of graph data to get insights and find trends with Amazon Neptune Analytics
AWS Clean Rooms Differential Privacy enhances privacy protection of your users’ data (preview)
AWS Clean Rooms ML helps customers and partners apply ML models without sharing raw data (preview)
Amazon Redshift adds new AI capabilities, including Amazon Q, to boost efficiency and productivity
54:33 📢 Jonathan – “The guardrails is super cool too, just because I think that, um, you know, this is, you know, the fiThe vector search stuff and the vector engine, I think are really good features that people will find useful applications for because in ML models, vectors are everything, but in static datasets like JSON documents and things, it’s really hard to search for things that are like something else, but not exactly the same. Like if you’ve got pictures of buildings and you say, well, give me all the red buildings. Well, which shade of red exactly? What you’re asking for is what Postgres or SQL server would be saying, you know, what’s the RGB color that you’re looking for? Whereas vector searches will let you say, I found all these results and these are the closest things to what you’re asking for that I can find. Or these are the 10 closest things I can find. And so it lets you search for things that are like other things without having to search for them precisely.”
1:09:48 Not in Keynote:
Package and deploy models faster with new tools and guided workflows in Amazon SageMaker
Use natural language to explore and prepare data with a new capability of Amazon SageMaker Canvas
Leverage foundation models for business analysis at scale with Amazon SageMaker Canvas
Introducing highly durable Amazon OpenSearch Service clusters with 30% price/performance improvement JPB
You can now get a better price for all your OpenSearch needs – and Justin is *very* excited about this one.
Amazon SageMaker Clarify makes it easier to evaluate and select foundation models (preview)
Easily deploy SaaS products with new Quick Launch in AWS Marketplace
1:13:00 Werner’s Keynote
Basically, we got a sustainability and cost management talk – which is better than AI, which is what we were expecting.
https://thefrugalarchitect.com/
Amazon CloudWatch Application Signals for automatic instrumentation of your applications (preview)
New myApplications in the AWS Management Console simplifies managing your application resources
1:28:10 Not mentioned in Keynote
Use AWS Fault Injection Service to demonstrate multi-region and multi-AZ application resilience -MK
54:33 📢 Justin – “When I first read this, I was kind of like, I don’t, I don’t think I understand it very well, but then I kind of clicked in my head. I was like, oh, well if RDS is broken and AZ one, but everything else is working right. I’m still routing traffic in the front end to this thing that’s going to hit a RDS node that’s dead. And that’s not so great. So if I can at least turn it off on a DNS level, then I can shift all the components that are living in a single AZ to the other AZs that I know are fully working and fully operational. And think about it from a full stack health perspective versus a partial.opponent level health perspective. And so it does make a lot of sense why you would want this. And so this is a nice improvement. Now that I understand it, when I first read it, I was like, I didn’t get it. But it clicked with me earlier.”
AWS Predictions
Jonathan
GPU Support for Lambda functionsChat Bot integration for the support portal that pulls from documentationNew Baremetal Instance with more GPU’s for AI Training
Justin
Graviton AI Chip Capabilities- Olympus with a bigger data set than Open AI and publicly available
- Major Improvements to Quicksight
Ryan
AppMesh will support serverless workloadsData Sovereignty on stageJust in time IAM Permissions powered by AI
Matt
- AI Chat feature in the AWS Console
- Carbon Emissions and Green Technology talked about during the keynote.
Predictive typing thing integrated into AWS Shell (cloud 9).
Tie Breaker: Number of times the word Artificial Intelligence and/or AI.
Matt – 72
Ryan – 563
Justin – 142
Jonathan – 90
Results:
MNL – 2
Adam S – 109
Swami – 65
Werner – 29
Total – 204 Times
Honorable Mentions:
Reinvent announcement of Clippy/Mascot (Jonathan)
Chip Fab (Jonathan)
Astro Bot upgrade (Ryan)
Astrobot Robot Wars (Ryan)
Extra effort/hardware on energy usage (Jonathan)
IAM Permissions reducer (Matt)
Security/Guardduty/SOC AI (Justin)
DuckDB (Justin)
AI for Opensearch (Justin)
Werner masterclass on AI (Justin)
Simulated worlds (Jonathan)
Closing
And that is the week in the cloud! We would like to thank our sponsors Foghorn Consulting. Check out our website, the home of the Cloud Pod where you can join our newsletter, slack team, send feedback or ask questions at theCloud Pod.net or tweet at us with hashtag #theCloud Pod
Apple Podcasts
Google Podcasts
Stitcher
Spotify
Overcast