direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Page Content

Master Project: Distributed Systems


Dates / Locations
Event

Date/Time
Format
Kick-off meeting
14.04.2021 (Wednesday)
16:00 - 17:00
Zoom
Regular meetings
Regular project meetings from 28.04
(approx.) bi-weekly on Wednesdays
16:00 - 17:30
Zoom

Registration

  • This project is limited to 12 students.
  • Registration has to be done via the ISIS page.
  • The registration opens on 14.04.2021 at 4pm and closes on 20.04.2021 at 9am. Each successful registration will be confirmed on 21.04.2021.

Topic: Serverless Data Processing on Edge Resources

This summer semester, we offer a master project on serverless data processing in edge computing environments.

The Serverless and Function-as-a-Service cloud computing models describe a paradigm in which the computation and storage is decoupled and code can be executed without the need to manage the underlying resources. For this, event-driven programming is typically combined with cloud features such as auto-scaling, resource scheduling, and virtual networking. Although the serverless model is widely available and popular in public clouds, the new network and computing architectures of the Internet of Things, such as edge and fog computing, require adjustments to the existing FaaS frameworks that realize low-latency, high-throughput and reliable processing at the edge. Specific challenges to work on may include the integration of distributed data storage into serverless architectures and other possible I/O bottlenecks, cold start of functions, auto-scaling, scheduling and resource allocation, and monitoring in order to increase dependability.

In this master project, you will work in small teams and investigate how the existing Open-Source FaaS frameworks can be used in IoT environments.

During an initial task you will get familiar with serverless platforms (e.g. Apache OpenWhisk, Kubeless, OpenFaaS) and orchestration stacks (e.g. Docker, Kubernetes), before starting to develop scientific approaches to tackle one of the aforementioned challenges in the IoT domain. In your team, you will be working collaboratively in the assigned topic area on resources (e.g. Public Cloud credits, Raspberry Pis) provided by DOS. Further, the developed methods will be benchmarked to evaluate their applicability and compared to state-of-the-art approaches. Therefore, existing IoT workloads need to be examined, implemented and simulated. During the course of the project, you will regularly present the current work of your team and finally hand in a deliverable in form of a report.

Target Audience

Master students with solid programming skills, who would like to gain some hands-on experience in a practical project in the area of distributed and operating systems.

Prerequisites

  • Good programming skills in at least one high-level programming language (e.g. Java, Python, Go, Scala).
  • We assume basic knowledge of OS-level Virtualization technologies (e.g. Docker, Kubernetes), Linux/Unix systems and an interest in Edge/Fog Computing, Serverless Architectures and Distributed Processing.

Zusatzinformationen / Extras

Quick Access:

Schnellnavigation zur Seite über Nummerneingabe

Auxiliary Functions

Contact

Sören Becker
+49 30 314-78592
Room TEL 1204

Contact

Jonathan Bader
+49 30 314-22132
Room TEL 1217

Contact

Lauritz Thamsen
+49 30 314-24539
TEL
Room 1210