December 10, 2024
In our previous issue, Aurora Serverless v2 finally scaled to zero, AWS prepared for re:Invent, and I shared some exciting personal news about joining CloudZero. This week, we look back at a very busy AWS re:Invent, I share some thoughts on the biggest announcements, and Elastic Cloud Serverless goes GA. Plus, we have lots of community reactions and some amazing content for you!
Okay, I'm warning you right now. This is a long one!
Last week was AWS re:Invent, aka, the superbowl for cloud nerds like me (and you 🤓 - it's okay to admit it)! There were a lot of announcements, and since we didn't publish a newsletter last week, a few announcements from the week before are going to sneak in here as well.
Let's start with the "big one", Amazon Aurora DSQL (in preview). I've known about this service for a few months and have spent some time with it. It is not perfect (yet), it doesn't support Foreign Keys (yet), and it isn't fully compatible with Postgres (yet). However, the fundamentals of this thing are rock solid and could completely change how we think about distributed databases moving forward. I mean, we're talking using satellite synced atomic clocks and new types of fiber optics. I might be overthinking it, but I wouldn't want to be a database company that didn't own my datacenters and networks right now. 😬
There are a few posts about DSQL (which we've all agreed to pronounce DEE-SQUEAL - not a typo), and I'm sure there will be many more to come. Check out Unlocking Aurora DSQL with AWS Lambda: A Seamless Solution for Serverless, Scalable, and Event-Driven Architectures by Jatin Mehrotra, First Impressions of AWS DSQL with Lambda and Rust by Benjamen Pyle, and the official Introducing Amazon Aurora DSQL blog post.
There was little doubt that GenAI would feature prominently this year, but honestly, I was pleasantly surprised by many of the announcements. The focus seemed squarely on solving most of the problems that GenAI created as well as launching additional tools to make integrations easier.
For example, Knowledge Bases in Bedrock were already really useful, but now you can use Amazon Aurora as a vector store (read more here), they can process multimodal data, they now support custom connectors and ingestion of streaming data as well as streaming responses, and they now support GraphRAG (in preview). Knowledge Bases also now support structured data retrieval which allows you to query data directly from RedShift and Amazon Sagemaker Lakehouse, and a new Rerank API to improve accuracy of RAG applications by applying relevance to retrieved records.
Amazon Bedrock also had announcements for multi-agent collaboration, Intelligent Prompt Routing, Model Evaluation with LLM-as-a-judge, and Data Automation to apply intelligent document processing, media analysis, and other multimodal data-centric automation solutions.
AWS announced the preview of prompt caching which can reduce costs by up to 90% and latency by up to 85% by caching system prompts instead of burning tokens on them every time. Super useful if you're storing lots of context about a user and/or need to provide a large set of instructions to the LLM. Looking forward to this one.
And Guardrails! These are lifesavers that can massively reduce costs and malicious user behavior. Amazon Bedrock Guardrails now supports Automated Reasoning checks (Preview), which automatically reviews its answers to make sure they are correct (or at least as correct as possible). And Amazon Bedrock Guardrails supports multimodal toxicity detection for image content (Preview) to prevent users from uploading all those inappropriate images.
Some other interesting Bedrock announcements include introducing latency-optimized inference for foundation models in Amazon Bedrock. They are a little more expensive, but super fast (which might satisfy certain use cases). And Amazon Bedrock Model Distillation which helps automate the fine-tuning of faster, more cost-effective models for specific use cases.
There's so much more! The Amazon Bedrock Marketplace now brings over 100 models to Amazon Bedrock that you can deploy to SageMaker endpoints, and they released the new Amazon Nova foundation models, which include video generation (example here).
There were also a number of announcements regarding Amazon Q Developer (formerly known as Amazon CodeWhisperer). Amazon Q Developer can now generate documentation within your source code, automate code reviews, and automatically generate unit tests. It also now provides natural language cost analysis which lets you ask questions about your bill and then deep links you to the cost reports.
And there are now several new "transformation capabilities" for mainframe modernization, .NET porting, and VMware. Interesting times we live in. 🤖
I should also mention the next generation of Amazon SageMaker and the preview of Amazon SageMaker Unified Studio. I appreciated that Werner mentioned "old ai" in his keynote last year, so it's good to see that AWS is investing there as well.
It was also a big year for S3. They announced Amazon S3 Tables, which are fully managed Apache Iceberg tables optimized for analytics workloads. And Amazon S3 Metadata (in preview) that lets you search and manage S3 metadata. These could be very big.
And finally, two honorable mentions include Amazon DynamoDB global tables previewing multi-Region strong consistency and Amazon EventBridge and AWS Step Functions integrating with private APIs.
I'm not the only one recapping these announcements, so be sure to check out AWS's top announcements of AWS re:Invent 2024 as well as Ran Isenberg's AWS re:Invent 2024: My Serverless Takeaways and Yan Cui's Biggest re:Invent 2024 serverless announcements.
With all the AWS re:Invent excitement, it's easy to forget that some other companies out there are making significant strides in the serverless space, including Cloudfare. They've accomplished quite a bit in 2024, all captured here in their Cloudflare 2024 Year in Review.
Also, Elastic Cloud Serverless is now GA! According to the announcement, they've completely re-architectured Elasticsearch for the cloud to quickly start and scale security, observability, and search, without the usual hassles. I've had no time to dig into this, but given that Ajay Nair is involved, I'm going to be setting aside plenty of time.
Marc Brooker wrote a series of excellent deep dives into the new Amazon Aurora DSQL service that are definitely worth checking out:
No, ‘Serverless’ and ‘Cloud’ Are Not ‘Scams’
I like a good rebuttal. 😉 Mark Tinderholt shares his thoughts on the recent "Why Serverless Is A Scam" article floating around Medium. This is behind the paywall, but maybe you'll find a way around it for a good counterpoint argument.
Re:Invent 2024 - A New Hope
Andreas Casen thinks AWS is back on the right track (as opposed to the heavy focus on GenAI). I tend to agree, but there's more work to be done.
Serverless Under Fire: Complexity, Costs, and Cloud Confusion
Bran Kop shares a "balanced" exploration of serverless' value. Seems quite reasonable to me.
Help! I Think I Broke DynamoDB – A Tale of Three Wishes 🧞♂
Sebastian Zakłada has some beef with DynamoDB, some of which is justified. Sure we all wish for full-text search capabilities, but I wouldn't hold my breath.
I haven't had a chance to watch the amazing sessions from AWS re:Invent just yet, but here are a few places to get you started. The CEO Keynote with Matt Garman (with a nice shoutout to AWS Heroes), the Keynote with Dr. Swami Sivasubramanian, and, of course, Dr. Werner Vogels' Keynote (with a reminder to "listen to the heroes") are all loaded with great overviews of the main announcements. They also dig in a bit more to some of the strategy behind these announcements and how they align with customer needs.
There was also the Reimagining the developer experience at AWS (DOP220) session that is worth becoming familiar with. Does AWS finally understand DevEx? It still doesn't seem like it, but perhaps this session will give you some insight as to where their thinking is.
Two quick shoutouts to Best practices for serverless developers (SVS401) and Asynchronous frontends: Building seamless event-driven experiences (API305). I did watch these sessions and hope they haven't set the bar too high for all the others to follow. 😉
I was lucky enough to join Farrah Campbell and AM Grobelny on AWS onAir for a segment talking about 10 years of AWS Lambda. It was fun and informative (hopefully). I think you'll enjoy it.
Also, Corey Quinn joined Sebastien Stormacq for an AWS re:Invent re:Cap on the AWS Developer Podcast. I've known Corey for a long time and have been following his work for even longer. This might be the most insightful and genuinely thoughtful interview I've ever heard him give. Definitely worth the listen.
There were some other conferences besides AWS re:Invent this year that recently released videos.
From QCon London:
And from GOTO Conf:
Also, SiliconANGLE caught up with Deepak Singh before re:Invent to talk about the future of AI, especially surrounding developer productivity. The first comment on the YouTube video made me laugh. I do believe that AI helps developers do things faster, but I'm not 100% convinced that translates to productivity gains.
AWS Bites #136. 20 Amazing New AWS Features
Eoin and Luciano share a long list of their favorite pre:Invent announcements. I always appreciate their take on this stuff.
I know, I know, this is ridiculously long, but here are several more AWS announcements from pre/re:Invent that might interest you:
Introducing Hono OpenAPI: Simplifying API Documentation for HonoJS
Not really a "serverless" tool, but I use Hono for all my serverless APIs, so 🤷♂️. Aditya Mathur has put together a new tool for generating OpenAI specs from your Hono API definitions. It looks very cool.
Serverless Hosting pricing
Online calculator that still needs a bit of work, but gives you a general idea of how pricing compares across different providers. Still missing Lambda as of the time of this writing, but still useful.
December 13, 2024 - ServerlessDays Rome
February 20, 2025 - ServerlessDays Manchester 2025
Please send me your serverless events!
AWS announced Amazon Aurora DSQL, a new serverless, distributed SQL database with active-active high availability. Which database engine is it compatible with?
Click an answer below to start the quiz.
I've honored all of them separately in the past, but this week’s stars are Farrah Campbell (@FarrahC32), Taylor Jacobsen (@taylorjacobsen), and Ross Barich (@rossbarich). No one understands community like they do, and I am so thankful for everything they’ve done for AWS developer communities. Whether you’re a community builder, user group leader, or AWS Hero, they have no doubt made your experience with AWS remarkably better. Farrah, Taylor, and Ross, thank you again for your commitment to all of us.
Well, that was a lot. If you’ve made it this far, congratulations—you might just be as big a cloud nerd as I am. 🙂
For those of you who were at re:Invent that I didn't get a chance to connect with, let's set up a time to chat. Also, for anyone suffering from a post-re:Invent illness (aka re:Invent's revenge or re:Invent re:Infect), I hope you get some rest and re:Cover soon. I'm right there with you. 🤒
Until next week,
Jeremy
I hope you enjoyed this newsletter. We're always looking for ideas and feedback to make it better and more inclusive, so please feel free to reach out to me via Bluesky, LinkedIn, X, or email.
Stay up to date on using serverless to build modern applications in the cloud. Get insights from experts, product releases, industry happenings, tutorials and much more, every week!
Check out all of our amazing sponsors and find out how you can help spread the #serverless word by sponsoring an issue.
Jeremy is the CEO and Founder of Ampt and an AWS Serverless Hero that has a soft spot for helping people solve problems using serverless. He frequently consults with companies and developers transitioning away from the traditional “server-full” approach. You can find him ranting about serverless on Twitter, in several forums and Slack groups, hosting the Serverless Chats podcast, and at conferences around the world.
Off-by-none is committed to celebrating the diversity of the serverless community and recognizing the people who make it awesome. If you know of someone doing amazing things with serverless, please nominate them to be a Serverless Star ⭐️!