LFU:online

Learning to develop, deploy, optimize, and run serverless distributed applications across the edge-cloud infrastructures.

Introduction to Edge-Cloud Continuum. FaaSification. Development and Deployment of Portable Serverless Distributed Applications. Runtime System Optimization for Distributed Applications. Federation (FaaS, Storage, Backend-as-a-Service). Portable Execution of Distributed Applications. 

Students will apply the theory that covers a specific part of distributed application life-cycle (facilitated development, automatized deployment and optimal execution in federated serverless infrastructures) in exercises.

During the semester, students will work on the final project. The design principle code-once run-everywhere, with every service will be explored.

In-class exercises; project development with milestones; presentation of initial, advanced, and final project prototype; and final report.

For each topic, a set of scientific papers will be given that have a huge impact in the topic.

Java, JS, or Python. Basic knowledge of distributed systems.

Planned technologies:

1. FaaS. Students can select any programming language (even multiple) to code serverless functions.

2. FaaS Orchestration. Any serverless workflow management system for orchestration, which allows to develop and run distributed applications across multiple cloud providers. 

3. BaaS. Students can freely select two BaaS services (object recognition, text2speech, speech2text, etc). 

4. BaaS interoperability. Students may use existing libraries for interoperability (dynamically select the specific provider) or code their own within the project.

5. Scheduling. Students may use existing schedulers or code their own within the project.

6. Cloud providers. At least two public clouds will be used. Recommended are AWS and Google, but students can select based on their preferences. Open source are also OK.

All cloud services from AWS, Google, and IBM can be used for free. Students will use the AWS Academy program with a vaucher (100$) and access to many cloud services on two AWS cloud regions. Students will use GCP vaucher ($50) for GCP cloud services.