# DoC paper meeting 2022-01-19

## Meeting venue
Online [[WebEx](https://fu-berlin.webex.com/fu-berlin-en/j.php?MTID=m31855a21c0a057fa48642d9cbe84a69d)]

## Agenda

- Short introduction
- Important dates:
- 2022-02-07: early bird registration!
- [2022-03-07 23:59 UTC](https://datatracker.ietf.org/meeting/important-dates/): IETF 113 draft cut-off
- [2022-03-19 - 03-25](https://www.ietf.org/how/meetings/113/): IETF 113 in Vienna
- [2022-04-13](https://www.usenix.org/conference/nsdi23) (CfP not out yet): Usenix NSDI'23 paper titles and abstracts
- [2022-04-20](https://www.usenix.org/conference/nsdi23): Usenix NSDI'23 full paper submissions (12 pages without refs, appendixes, ...)
- Discuss outline / goal
- Discussion: conclusion/take-aways so far
- Further evaluation
- Implementation section rework
- Go through paper and outstanding TODOs section-wise
- Further steps
- additional experiments
- additional data sets
- a new Content-Type?

## Notes

## Diskussion: Zielsetzung
- Presentation of DoC
- DoC: provides comparibly lean means to transport DNS encrypted and blockwise (mode)
- Design Decision Documentation
- Caching: TTL vs. Max-Age
- FETCH/GET/POST
- CBOR-based Content-Format
- ACK-timeout at stub resolver

## Discussion: conclusion/take-aways so far
- Reading out takeaways from 1178516435d429224df09cfd45774baf2b26eec5 version of paper
- Discussion, Fig. 8:
- Discussing scraping of bytes of OSCORE AAAA-Response
- „Getting below the red line“ needs to be a focus of the evaluation
- Different records in same query:
- How to compress? Fallback possible? Wireformat, CBOR, CoAP options...
- Multiple responses for same DNS TID possible in DNS?
- First in design decisions standard compliant proposal
- Then show in eval that non-standard-compliant might be feasible

## Further evaluation
### Empiric DNS data model
- How do queries generally look like? Talk with Marcin about IXP analysis.
- For responses Afnic + find typical IoT names
- TTN for their names?
- What does the IoT-device resolve?
- Talking of devices among each other => core-rd
- DNS in the IoT: IMC paper von [Choffnes](https://david.choffnes.com/) et al.
- Further reading:
- https://dl.acm.org/doi/10.1145/3419394.3423650
- Maybe https://dl.acm.org/doi/10.1145/3355369.3355577
- Maybe also in part https://dl.acm.org/doi/10.1145/3485983.3494869 for classification approach

### Other things to look into
- Blockwise: Wait for data model, first data argumentation then evaluation
- DNSSEC? Discuss, but not evaluate. Most likely not relevant for IoT

### Blocking evaluation
- Move to appendix, "in Nebensatz abtun"

### A record evaluation necessary (IPv4 vs IoT)
- Yes, but mention why. Also BLE allows IPv4 without 6LoWPAN

## Implementation section
- Matthias: System perspective lacking for a systems conference
- Thomas: Change of perspective? Rather than how to integrate DoC into a specific OS (combination of components), discuss what components are needed, how to combine => this is why we use RIOT, since these components are there
- Modular deployment
- How much ROM / RAM needed should already be discussed
- What's the difference in approach to a non-IoT system

# Next steps
See https://github.com/anr-bmbf-pivot/pub-dns-over-coap/projects/1