π BREAKING: My First Research Paper Survived Peer Review!
(And so did my hair! βοΈ) From palm wine in Cameroon to peer-reviewed publication β here's the story of AdaptiFlow!
Achievement Unlocked
First Paper Accepted!
Peer-reviewed and ready to share with the world
π The Secret Writing Sessions
Hey network! π Remember when I said I was "fully disconnecting" in Cameroon during the holidays? π
Surprise! I may have fibbed a bitβ¦
While sipping palm wine π₯₯ and eating okok π€€ with family, my brain was low-key brewing something else: my first-ever research paper! πβ‘οΈ
"The best ideas come when you're supposed to be relaxing. Something about palm wine and Cameroonian sunshine unlocks creativity."
π Presenting: AdaptiFlow
After months of research, coding, experimenting, writing, rewriting, and probably too much coffee, I'm thrilled to introduce:
Workshop on Autonomic and Self-Adaptive Systems (WACA) @ DisCoTec 2025
AdaptiFlow: An Extensible Framework for Event-Driven Autonomy in Cloud Microservices
A framework that enables cloud microservices to self-heal, self-protect, and self-optimize through abstraction layers and event-driven architecture.
π± The Problem: Herding Cats
Modern clouds π©οΈ need to self-heal, self-protect, and self-optimize β but doing this in microservices architectures is like herding cats π±π₯
Why? Because microservices are:
- Distributed across multiple nodes and data centers
- Independently deployable (and independently breakable!)
- Communicating through complex networks of APIs and events
- Scaling dynamically based on load
Traditional centralized approaches to system management don't work well here. We needed something different.
π‘ The Solution: AdaptiFlow's Three Layers
AdaptiFlow introduces abstraction layers that let microservices become autonomous while remaining loosely coupled:
Perception Layer
Unified metrics collection β CPU usage, latency, cache hits, error rates, and more. One interface to rule them all.
Execution Layer
Declarative actions β scale instances, toggle feature flags, reroute traffic, adjust resources. Tell the system what to do, not how.
Logic Layer
Event-driven rules for adaptation β without centralized chaos! Each service can reason about its own adaptation needs.
π¬ The MAPE-K Connection
If you've read my post about the glass frog's adaptation strategy, you'll recognize the pattern here. AdaptiFlow implements the MAPE-K loop:
- Monitor β Perception layer collects metrics
- Analyze β Logic layer processes events and detects patterns
- Plan β Logic layer determines appropriate responses
- Execute β Execution layer carries out adaptations
- Knowledge β Shared understanding across the system
Nature figured this out millions of years ago. We're just implementing it in code! πΈ
π£ Resources
Want to dive deeper? Here's everything you need:
Paper Preprint
Read the full paper on HAL Science
AdaptiFlow GitLab
Explore the source code and documentation
Docker Hub
Ready-to-use Adaptable TeaStore images
π Shoutouts That Saved My Sanity
This paper wouldn't exist without some incredible people:
- My surgical-strike supervisors β your guidance was the GPS πΊοΈ through my paper-writing maze!
- The Spirals crew β y'all rock! π€
- The reviewers β your feedback made this paper SO much better
- My family in Cameroon β for tolerating my "vacation" laptop sessions
- Coffee β β the unsung hero of academia
π What's Next?
This is just the beginning! The paper opens doors to:
- Presenting at DisCoTec 2025 in Lille
- Expanding AdaptiFlow's capabilities
- More research, more papers, more discoveries
- Building a community around self-adaptive microservices
First paper down. Many more to come! ππ₯
1st
Publication
3
Layers
β
Coffee Cups
100%
Hair Intact
ArlΓ©on Zemtsop
PhD Student at Inria Lille / University of Lille β’ First-time Published Author! π