You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 7 Next »

Goal

Hyperledger Umbra (https://github.com/hyperledger-labs/umbra) is a project created during the 2019 mentorship period as a network simulation tool to run the blockchain environment. The plan for this year is to add network fuzzing capabilities such as packet drops, delay, loss, etc.

Proposed Timeline

Week 1 - 3 (June 1 - June 19)

  • Understand blockchain, docker, containernet, SDN, etc concepts
  • Setup and get familiar with the development environment - run tests, read logs, etc
  • Get high level understanding of the current umbra code - advanced Python ramp-up
  • From above learnings, improve umbra docs, setup scripts, etc if needed
  • Outcome: setup dev environment and run the project. Familiarized with the code, run the sample fabric configs, understand the output.

Week 4 - 6 (June 22 - July 10)

  • Pathfinding: ability to save and replay packet. E.g. each peer saves each packet sends so umbra-broker can instruct all the nodes to replay the packet
    • Outcome: current maintainer has the sample implementation (umbra/umbra-agent, umbra/umbra-monitor) that will be integrated to umbra-broker. Need to study and understand how it works
      • umbra/umbra-agent: Uses third-party tools like Linux `tcpreplay` and `tcpdump` to save and replay packet
      • umbra/umbra-monitor: Monitor host platform (via `platform` Python standard lib), Linux processes (via `psutil` thirdy-party library), and Docker container (via Docker Python API)

      • Learned a lot about python asyncio which is used extensively to schedule commands in both umbra-agent/monitor to simulate the environment
      • Found issue with scheduler.py where it cannot run command repeatedly. Raphael Rosa fixed it with this commit
  • Implement the ability to remove container(s), remove/update link behavior (bandwidth, delay, loss, etc), and change node resources (cpu/mem) in umbra-scenario component
    • Test feature -  refer umbra/examples/fabric/build_configs.py
    • Outcome: pending, moved to next quarter. But learned a lot about the internals design and how the umbra-scenario <--> umbra-broker works together.

Week 7 - 9 (July 13 - July 31)

  • Continue working on packet save/replay feature. 
  • Add events to umbra-broker based on previous quarter link/node behavior changes in umbra-scenario
  • Review implementation details with Umbra maintainer
  • Outcome: rough implementation of event that can update environments like stopping container, updating cpu/mem resources, and link interface status (UP or DOWN)

Week 10 - 12 (Aug 3 - Aug 21)

  • Add the ability in broker to parse events from config (e.g., event with “category” field set to “scenario”), containing the proper params to call the scenario modifications (e.g., node: name, operation: delete, etc)
  • Generate a report for this fuzzing activity - list out all the events, status, etc
  • Document and wrap-up work (scripts to automate stuff, etc)
  • Prepare for project presentation
  • Outcome: Two PRs for code changes PR#9 (integrate umbra-agent + monitor + environment plugin), PR#10 (Refactors + Simulation report) and another PR for docs. Documentation update for new components might be pushed to another one or two more weeks.

Process

  • Weekly meeting with mentor and other mentees that use Umbra. Other meetings scheduled as required
  • Chat via Signal
  • No labels