Research & Innovation Laboratory
An independent research laboratory dedicated to advancing software development through rigorous experimentation, technical analysis, and innovative solutions to complex engineering challenges.
TM Dev Lab is an independent research laboratory focused on advancing the state of software engineering through systematic investigation, rigorous experimentation, and open publication of findings. We explore emerging technologies, design patterns, and architectural approaches with scientific methodology, documenting both successes and failures to contribute to the broader engineering community.
Controlled investigations of emerging technologies and architectural patterns using empirical methods
In-depth technical analysis, comparative studies, and exploration of complex software engineering challenges
Development of novel tools and frameworks based on research findings and experimental results
A rigorous, reproducible benchmark of Microsoft's jaz, the Azure Command Launcher for Java, against the plain java defaults across container sizes. Where auto-tuning the JVM from cgroup limits helps, and where it does not.
Read more →The launch of Marreta Lang, a focused DSL for REST APIs interpreted by a compiled native runtime. Why we built it, what the language looks like, and the developer experience it is built around.
Read more →Enterprise systems demand predictable performance under sustained high-concurrency load, and the JVM was engineered exactly for that. This study walks through every major HotSpot subsystem to show where that robustness comes from and what it costs.
Read more →Expanded benchmark covering 15 MCP server implementations across Rust, Java (Spring Boot, WebFlux, Virtual Threads, GraalVM native), Quarkus, Micronaut, Go, Bun, Node.js, and Python. 39.9 million requests across three independent runs with I/O-bound workloads (Redis + HTTP API), 0% error rate.
Read more →Comprehensive performance analysis of MCP server implementations across Java, Go, Node.js, and Python. Testing 3.9 million requests over three benchmark rounds to measure latency, throughput, resource efficiency, and production-readiness characteristics under controlled load conditions.
Read more →A self-hosted gateway for the Model Context Protocol (MCP), enabling organizations to securely connect AI agents to MCP servers with tool governance, multi-tenant isolation, credential injection, and full observability. Built with Rust and Vue 3.
Read more →Third installment in the Production-Ready MCP series. Comprehensive analysis of Zero Trust architecture for MCP ecosystems, covering threat models, OAuth 2.1 flows, policy engines (OPA/Cedar), gateway security controls, supply chain verification, and enterprise implementation patterns for securing autonomous AI agents.
Read more →Second installment in the Production-Ready MCP series. Comprehensive examination of gateway patterns for agentic systems, federated registry architectures for service discovery, distributed systems patterns (Ambassador, BFF), and comparative analysis of production implementations from Microsoft, Docker, Kong, and Cloudflare.
Read more →First post in the Production-Ready MCP series. Technical analysis of Model Context Protocol server scalability, examining the transition from SSE to Streamable HTTP and distributed session management with Redis in Kubernetes environments.
Read more →