2018-01-25 Engineering Meeting

Date/Time: 2018-01-25 @ 14:00 UTC / 09:00 EST / 22:00 HKT
Meeting link: https://zoom.us/j/432646691

Attendees

@larry
@aaron
@yukan.id
@jackzampolin
@muneeb
@jude
@xan

Agenda

Should we increase the zonefile limit?

  • Opportunity when roll out atlas update to increase zonefile limit
  • 20-30 subdomains today, 10x the zone file
  • What is the current
  • @aaron will take the lead of creating a pro-con document

Schedule Atlas network upgrade

  • Need ability to auto-ban peers that are misbehaving

Subdomain Creation/Update Rate-Limiting

  • Related to the above — @jude will describe in a forum post

Typescript Definitions for blockstack.js

  • @larry will add feedback to the pull request

Infrastructure Discussion:

  • We’ve decided to go with google cloud because they’ve offered credits
  • Goal: shut down everything on azure and aws
  • Muneeb: We should keep security and access control in mind
  • Jack: automated deployments and moving over to Google Cloud will be ~2 sprint process

Technical Debt

  • Muneeb: something we can work on during Q1 in parallel with other items

Shipping subdomains

  • subdomains should be processed by blockstackd so that linux & mac
  • browser needs to updated and tested to use the subdomain registrar
  • support in windows

Authentication Review and Spec Creation

See Review & Spec creation for Blockstack Authentication

Desired outcome: Does the process make sense and are we missing anything? Can we recommend this for a sprint item for next week?

Yes - it makes sense to address this sooner rather than later.

Community Comments/Questions

Wanted to flesh this out a little bit —

  • blockstackd support – @jude is working on moving the subdomain processing out of blockstack_client and into blockstackd. This should allow mac/linux clients to query subdomain names (currently, the clients are default configured not to try and resolve subdomains, because it would require the clients to download all the zonefiles necessary to process subdomain ownership.
    • I have a question about the timing of this — @jude is also currently working on other changes (atlas-related) in this branch as well – my understanding is that the progress here is really good, and that it looks like the consensus hashes are converging, but Jude should chime in to let us now whether or not bundling these two features together and shipping them is doable next sprint (as it would need to be released before support can be turned on in clients) – I can also help out with this.
  • registrar deployment – the subdomain_registrar code we currently have is very much a prototype.
    • This could just beef up the python code that’s there now and make it easier to deploy in a container. Really, it only uses the blockstack_client to submit it’s updates (though it calls functions out of subdomains.py.
    • Ideally, I’d spin it out into another repo, and have it function somewhat independent of a local blockstack_client instance. This should be pretty doable, though, I’d guess it’d need to know how to generate its own transactions, so it could use my branch of blockstack.js
  • browser support –
    • this shouldn’t be too different from the code @larry already has in there, however, we would need dialogs to select whether you’re registering a domain name or a subdomain.
    • choosing which domain you are registering your subdomain on should be possible, but can probably be pushed to a later release.

I’m happy to work with you on the deployment side of the registrar @aaron. This is something we are going to want to make easy to deploy and operate for other users/orgs.