News
Cloud Expert Details Game-Changing Announcements from AWS re:Invent 2025
Cloud security architect and advisor Steven Smiley walked attendees through a slate of notable AWS announcements in a Virtualization & Cloud Review webcast today. The presentation held by Smiley, an AWS Ambassador, focused on changes to serverless computing and practical improvements in container and networking setup that developers and cloud operators should be aware of.
His session, titled "Top 10 Game-Changing AWS Announcements from re:Invent 2025 -- What You Need to Know Now," is being made available for on-demand replay thanks to the sponsor, Nutanix, which also presented a session at the multi-event summit.
"What is our first really game-changing announcement? This is AWS Lambdas, Durable Functions. So this seriously is a game changer at its core."
Steven Smiley, AWS Ambassador
AWS Lambda Durable Functions: A New Chapter in Serverless Workflows
[Click on image for larger view.] AWS Lambda Durable Functions
One of the most significant AWS announcements at re:Invent was Lambda Durable Functions, which Smiley described as a fundamental shift in how developers can use Lambda for real-world workflows.
Lambda has traditionally been optimized for short-lived execution. While functions can technically run for up to 15 minutes, Smiley noted that this has never been an ideal or economical pattern for most workloads.
"What is our first really game-changing announcement? This is AWS Lambdas, Durable Functions. So this seriously is a game changer at its core," Smiley said. "Lambda is for short-lived functions, usually only a second or less. It can run for 15 minutes, but you really do not ever want it running for 15 minutes."
The challenge, he explained, has always been how to handle workflows that need to pause while waiting for something external, such as a manual approval or a downstream system response.
He noted that the issue is especially obvious when a Lambda workflow needs to pause. "Using a sleep inside your lambda function is just a really expensive mistake until durable functions," Smiley said.
"So durable functions gives us the ability to suspend the execution of this function for either a defined period of time or waiting for some callback from some other event, and then when that trigger comes back, that kind of durable step can continue, and so you don't pay for that whole period of time while you're waiting."
Durable Functions addresses this by allowing Lambda executions to suspend and resume without charging for idle time. State is preserved automatically, enabling developers to write workflows that pause naturally and continue later without external orchestration.
Smiley emphasized that this opens Lambda to use cases that were previously awkward or cost-prohibitive.
"Check it out if you're using Step Functions or if you have workloads that might benefit from this -- this may change your usage, like things that you may have done in an ECS task, for example, where you wanted to have a full container, you may now decide to do in a Lambda durable function," Smiley said.
He also contrasted the new model with AWS Step Functions, noting that Durable Functions removes the need to define workflows in a separate service for many scenarios.
"You are writing normal code, and AWS is handling the state and the resumption for you, instead of you having to define everything in Amazon States Language," Smiley said.
Lambda Expansion: Managed Instances and Tenant Isolation
[Click on image for larger view.] AWS Lambda Managed Instances
Alongside Durable Functions, Smiley highlighted additional Lambda enhancements that further expand where serverless fits in AWS architectures.
Lambda Managed Instances allows functions to run on dedicated EC2-backed infrastructure, giving customers more predictable performance characteristics for steady workloads.
"If you have something that is basically running all the time, Lambda pricing does not always make sense, and this gives you another option that looks a lot more like traditional infrastructure," Smiley said.
[Click on image for larger view.] AWS Lambda Tenant Isolation
Tenant Isolation addresses long-standing concerns around shared execution environments in multi-tenant Lambda applications. Smiley explained that while developers have historically worked around this with duplication and cleanup logic, AWS is now providing a built-in mechanism.
"Now it's as easy as clicking a button to enable the isolation mode, and then when you execute that function, you just tell it what the tenant ID is, and it will manage separate execution environments for you," Smiley said.
Smiley framed these updates as part of a broader trend in which Lambda continues to blur the line between pure serverless and more traditional compute models.
Making Container and Networking Setup Less Painful
[Click on image for larger view.] ECS Express Mode
Smiley also discussed improvements aimed at simplifying container deployments, starting with ECS Express Mode.
Traditionally, getting an ECS service running requires configuring clusters, task definitions, networking, security groups, and load balancers. Express Mode automates much of that work.
"If you've got an image and you've maybe put it on your container registry, all you really have to do is say to express mode, 'Here's my image,'" Smiley said.
He positioned Express Mode as particularly useful for development environments, demonstrations, and proof-of-concept deployments, where speed and simplicity matter more than deep customization.
[Click on image for larger view.] Regional NAT Gateway
Networking configuration was another area where AWS focused on reducing operational friction. The new Regional NAT Gateway removes the need to manually manage NAT gateways across availability zones.
Smiley explained that the regional model operates at the VPC level and automatically scales as needed.
"Now what we can do is just create a regional NAT gateway that doesn't even require a public subnet. It's just attached to the VPC itself, and it will automatically expand and contract across availability zones," Smiley said.
He framed these efforts as part of AWS's broader push to reduce operational overhead and allow teams to spend more time building rather than managing infrastructure and remediation tasks.
And More
As noted, for those who missed the live event, the session will be available for on-demand replay thanks to Nutanix.
Smiley also touched on additional announcements, including preview security and DevOps agents and AWS Transform capabilities aimed at automating code modernization tasks.
And, although replays are fine -- this was just today, after all, so timeliness isn't an issue -- there are benefits of attending such summits and webcasts from Virtualization & Cloud Review and sister sites in person. Paramount among these is the ability to ask questions of the presenters, a rare chance to get one-on-one advice from bona fide subject matter experts.
Additional upcoming Virtualization & Cloud Review webcasts can be found here.
With all that in mind, here are some upcoming summits and webcasts coming up:
About the Author
David Ramel is an editor and writer at Converge 360.