Skip to content

Meeting Minutes

Jump to bottom
Lukas Röllin edited this page Nov 28, 2018 · 51 revisions

Nov 27

  • The project description will include the initial evaluation
  • DiscoverNetworks was used (addClient), but no ping on findClient since that led to deadlocks
  • Deployed TURN servers are not working yet. We will try another day to fix it. If we can't fix it, we'll write up the reasons

Nov 21

Discussion:

  • If TomP2P is bootstrapped, further continuous bootstrap calls make sure the network stays intact
    • When bootstrapping to a node outside the current DHT (or even to a node inside another DHT), the current DHT is expanded
  • peer.peerAddress does not find external addresses
    • Discover mechanisms should be used for finding these

Nov 14

We skipped this meeting since there were only updates and no questions

Nov 7

Discussion:

  • IEEE citation style of [123 is ok
  • When citing T. Bocek, be consistent about the name
  • TURN P2P doesn't seem feasible, other approach presented (manipulation priority decision based on latency)

Nov 6 - Meeting with AnneMarie O'Neill

(AnneMarie is an English lecturer at HSR Rapperswil)

We approached AnneMarie with the topic of proofreading our study thesis. To see what our intended writing style is and what parts need improval, we sent AnneMarie the current state of our thesis.

She proofread this version until our meeting and suggested detailed improvements. They can be summarized as:

  • Use a consistent tense
  • Do not use contractions in academic writing
  • Check quote marks (start marks seem to be the wrong way around)

October 31

Discussion:

October 24

Present:

  • Started writing
  • Started small refactorings

Discussion:

  • Thomas has some ideas for a Bachelor thesis
  • Send source code/give repository access for first review
  • A best-effort shutdown of the TomP2P peer is ok

October 17

Present:

  • Project is running in principle (but still a lot of details)
  • Demo is shown

Discussion:

  • New task: can TURN servers be decentralized in a similar fashion? Could they employ failover mechanisms?
  • We could use dsl.hsr.ch as a domain, but we would face administrative and technical challenges. Therefore, we will use our own DNS servers and maybe move over to dsl.hsr.ch later
  • It is possible to check for the bootstrap status by reading the status of the Future

October 10

Discussion:

  • multiple listeners on objectDataReply (started by sendDirect are not supported. Need to write our own dispatcher
  • ports(0) should let TomP2P choose its own port
  • TomP2P includes a mechanism to find public/outer addresses. It may not work with multiple levels of NAT, as is the case using Docker
  • Instead of using await, success/failure listeners should be added. Only awaitListeners in cases where it's absolutely required
  • Ideas for next steps:
    • Polish GUI to use as an example app
    • Example app should also work over the Internet

Objective:

  • Signaling for WebRTC works until next week

October 3

Present:

  • Proof of Concept

Discussion:

  • T. Bocek will delete the oral exam part out the project description (only needed for Bachelor thesis)
  • We found examples for sendDirect by other students
  • Use these as examples for "How to do X in TomP2P"
  • Brainstorming about additional features

For next week:

  • Tell if we're on track for creating a more robust solution in 2 weeks

September 26

Present:

  • Project idea => Sounds reasonable

Discussion:

  • We will get the project description later
  • Time tracking categories and granularity is our decision
  • It is ok to use Kotlin instead of Java
  • It is ok to do the thesis in English

For next week:

  • Write some of the project management stuff
  • Design the architecture
  • Write a really basic POC

September 20

  • We checked codecs and we found out that we would rather not implement a (Java) GUI and send/receive media streams etc.
  • We would use use WebRTC for this, but implement the signaling plane ourselves
  • We came to the conclusion, that we can split this up.
    • Priority 1: implement signaling plane (prepare both clients for an upcoming connection, exchange connection details, etc)
    • Priority 2: start the actual connection over this network

September 19

  • Kickoff
  • Discussed the scope of the project, we will use TomP2P
  • "P2P Skype" is still an option and we are evaluating this first
  • Until next week:
    • Check if OPUS and/or AV1 are candidates for a P2P Skype
      • Issue is presumably performance
      • Minimal platform: laptop
    • Set up project
    • Check distribution method: Flatpak, AppImage
Clone this wiki locally