2023-01-09 Architecture WG Meeting Notes

Date

09 Jan 2023

ZOOM Meeting Information:

Monday, Jan 9, 2023 at  11:30am PT/2:30pm ET

Join Zoom Meeting

https://zoom.us/j/7904999331

Meeting ID: 790 499 9331

Attendees:

• Sean Bohan (openIDL)
• Dale Harris (Travelers)
• Nathan Southern (openIDL)
• Jeff Braswell (openIDL)
• Ken Sayers (AAIS)
• Peter Antley (AAIS)
• Mason Wagoner (AAIS)
• David Reale (Travelers)
• Faheem Zakaria (Hanover)
• Allen Thomspon (Hanover)
• Brian Hoffman (Travelers)
• James Madison (Hartford)
• Tsvetan Georgiev (Senofi)
• Ash Naik (AAIS)
• Yanko Zhelyazkov (Senofi)

Agenda:

1. Welcome Back
2. Ken - Update on ND
3. Jeff - Update on Testnet
4. PeterA - Update on end-to-end demo
5. Sean - Request for AWG Chairs & IWG meeting
6. Continue discussion - Architecture& Scenarios
   1. openIDL - Architecture Definition Workspace (NEW)
   2. Drawings here: https://lucid.app/lucidchart/ac20d4e1-50ad-4367-b5cf-247ed9bad667/edit?viewport_loc=-301%2C-37%2C3200%2C1833%2CkgLSxXlpoGmM&invitationId=inv_de1a5e61-8edc-488a-90ee-8312f8c69cd4#

Notes:

• PA - spent time developing auto EOY report, Homeowners was done in 5 weeks, needs formatting work, believes would satisfy requirements, compliant with what HO would ask for
• PA - generated from homeowners data on testnet, not gen with UI, gen with PA's HDS on the testnet, Mason got a converter, taking in raw stat data, decoding, loading into HDS and running report
• PA - initial guess, once thru Auto got thru HO quick, will continue to accelerate, maybe do 1 line a month, make us more functional - met last week, doing commercial auto next in RRDMWG
• PA - how do we start to codify some stuff, JM discussion on permissions and secure system, upside down L chart
• PA - resources aross the top - screen

• PA - until now mentioning 2 tables, words on stat plans "prem record, cancellation record" - decided Prem and Loss not right words, changed tables to match what is going on
• Prem records in Policy table, and loss records in Claim table
• first user role - user doing load to HDS = ETL User = Read, Create, Update, Delete - doesnt need access to EP schema
• KS - full access? Why does ETL user need Delete? 
• JB on carrier side?
• PA - ETL or DBA?
• KS - sync jobs, want ETL to do it not adhoc dropping of db to be part of ETL process
• PA - haven't needed to delete yet, maybe right?
• JM - phys deletes are reality - CRUD for ETL, CRUD-S for DBA (manages schema, schema modifications), multigrain table - openIDL having R makes sense, Developer is tricky, need to see whats going on, issue is PII and personal info, where do we care? 
• JB - devs on carrier side? policy set by carriers
• JM - grey area - all base schema stuff on carrier side, argue this is a req doc, "heres what we ask carriers to do" - only one really mandatory is openIDL line 
• PA - 2 part - what you have to do and Reference Implementation on testnet
• JM - star by the R, sensitive data add views, etc. - put star and "read if you want to hide cert fields from devs use standard internal policies
• FZ - dev vs support? Ops?
• PA - some level of dev, there is a dev vs ops person, hasn't defined
• JM - dev ops? more oprtations, but L2, L3 in ops support  - not true development, still ops role even though dev comes in
• PA - R- - hide fields if critical
• FZ - devops as deploy/releases - this IS an ops role? change to operations
• KS - get to soon - startt poking at moving the HDS off the node and onto sep thing, PA is doing something similar in testnet, hopefully discuss something a carrier can do on the testnet to push forward (testnet or POC), next Arch proof point - can we get DB out of node, have an Arch that separates it from its own adapter  (vague but directional)
• JB - discussing roles, whats requireed, several on the call eager to do something - looking forward to reaching out in the near future
• PA - def with USAA in mind and where we are right now, working on comprehensive plan for ETL to present to AWG, do something serverless with ref implementation - let carriers handle their stuff, looking into have better conversation about etl at the end of the month
• FZ - is there a code repo, 
• JM - code management early and often is wonderful - need to be elegant with source code, mult worries them
• PA - github - get to place where network code is sep from app code
• PA - working in - how to align this
•    JM - rule 1, when it doubt go fine grained, bias to more not fewer repos, how do you get finer grain than one monolithic place
• repo by layer
• rate of change

Documentation:

Notes: (Notes taken live in Requirements document)

Recording: