Bedrock AgentCore Runtime vs. Lambda for Hosting Agents: Isolation, State and the Idle Bill

The same agent runs on AWS Lambda and on Amazon Bedrock AgentCore Runtime with two thin adapters. The differences are isolation, where conversation state lives and what is billed while the agent waits. AgentCore is cheaper for I/O-bound agents only if sessions are ended deliberately: left at defaults, idle memory makes it the more expensive option.

Read more →

Zero-Trust Microservice Networking: AWS VPC Lattice vs. Private API Gateway

In modern enterprise cloud architectures, applications rarely live in a single AWS Virtual Private Cloud (VPC) or even a single AWS account. Production environments are partitioned across dozens or hundreds of accounts for blast-radius isolation, compliance boundaries, and autonomous team ownership. However, securely interconnecting these distributed microservices has long been one of the most frustrating […]

Read more →

Declarative Stream Processing: EventBridge Pipes vs. Custom Lambda Glue

EventBridge Pipes is usually sold as cheaper and lighter on concurrency than a Lambda stream consumer. Checked against the AWS documentation, both claims mostly disappear once the consumer filters on its event source mapping. This article compares the two for a DynamoDB Streams to EventBridge integration, with tested code, a reproducible cost model and a decision framework.

Read more →

ECS Express Mode, App Runner or Lambda for a Small API

App Runner closes to new customers on 30 April 2026 and ECS Express Mode is the named alternative. A design comparison and a tested cost model for a small API on Lambda, App Runner and ECS Express Mode: the monthly floor of each, where the money goes, and when a container is cheaper than Lambda. Nothing deployed.

Read more →

SMART Backend Services: System-to-System FHIR Authorization in .NET 10

SMART Backend Services end to end in .NET 10: a client that signs a one-time assertion, a token endpoint that checks it, and a gateway that enforces system scopes in front of HAPI FHIR, including _revinclude, chained search, transactions and paging links. Tested with 22 token requests that must be refused and a second client in Python.

Read more →