/
2023-06-29 TSC Meeting Notes

2023-06-29 TSC Meeting Notes

Jun 29, 2023

Date

Jun 29, 2023

ZOOM Meeting Information:

Thursday, June 29, 2023, at  9am PT/12pm ET

 

Join Zoom Meeting

https://zoom.us/j/7904999331

Meeting ID: 790 499 9331

Attendees:

  • Sean Bohan (openIDL)

  • Jeff Braswell (openIDL)

  • Peter Antley (AAIS)

  • Josh Hershman (openIDL)

  • Yanko Zhelyazkov (Senofi)

  • Tsvetan Georgiev (Senofi)

TSC Voting Members Attendance:

  • Ken Sayers (AAIS)

Meeting Agenda:

  • Opening

    1. Call to Order 

    2. Anti-Trust, Review of TSC Meeting format and Participation by TSC Chair

  • TSC Activity Desk

    1. Architecture Working Group (SeanB)

    2. RRDMWG Update (PeterA)

      1. AAIS Stat Reporting using openIDL internal Project (PeterA)

    3. MS Hurricane Zeta POC (KenS) 

    4. NodeBuilder Workshops (JeffB/SeanB)

    5. Infrastructure WG (SeanB)

    6. AAIS / AWS Discussion regarding Node Delivery Partnership

  • Discussion:

  • AOB

FOLLOW UP:

  •  

Recording/Meeting Minutes

RRDMWG Update

  • OLGA application (loading data)

  • Internal POCs for unique functionality

  • best practices for uploads of data, chunking data, internal POCs

  • Grid components Weds next week, show stat data, talk Errors

  • Internal stat reporting team

    • getting faster at it

    • Crime being worked on

    • finishing up commercial property

    • Complete - place in the DB to handle putting stat data, way to decode the stat records to make them human readable

      • homeowners, dwelling property, personal and commercial auto

  • OLGA data quality checking - component - loading of positional data records into DB, thought process was more straightforward - loading of data isn't a major task as managing data quality?

    • no small tasks

    • simple one line of stat record into one table, simple as 50 attribute table is to load

    • intricate and time consuming, stat plan, ref tables

    • whats code 3 code 4 code 2

    • paid loss vs premium - ref table and analysis work more challenging than tech side of it is

    • Ref tables to expand codes, anticipated based on definitions, set in advance? Ref data tables "this code means this description for this line of business" - they wouldn't change  - per line effort -  the loading of the data, relates to description of what carrier needs to do to load data - loading of table is a first step

    • ETL on load, auto three digit date implies month/year and convert to ISO data, some of that, scripts for that

    • Sourcing harder than loading

    • can we get carrier involved for MS POC to use it?

    • Team talking to had the right permissions and could produce the dataset, for short term using BOX for data shipping and Peter loads into via OLGA

    • Carrier a reporting team and sharing data but if scripting, diff engineering team and more complexity

    • Data team vs IT team

    • need to inc an IT capability, will need to map those codes

    • level of engagement, participation

  • MS POC

    • talked to carrier, getting sense of what the work is

    • open and another meeting on the 10th

    • do loss codes need to be transformed? maybe, not like stat reporting but will be some work

  • NodeBuilder

    • current plan - varying degrees of completion

    • House calls - 1:1 sessions with carriers to get them thru completion, get each to have a node on the openIDL Network

  • Dependency killer - dont think need working nodes to spin up HDSs, use OLGA as a test harness - figure out baseline for setting up node 

  • establishing PRs process and DCO approvals

  • AAIS / AWS Discussion regarding Node Delivery Partnership

    • good if AWS could set up nodes like an IP, marketplace component

    • doesn't do anything or Azure support

    • call with AWS

    • experts in Fabric and Blockchain

    • General Marketplace

    • overall - to make a marketplace component, AAIS will create and manage the component, no skin in the game

    • learned - perspective - first statement, trim down all the stuff that doesn't work for all and deliver that

    • gets KS thinking - mingle with nodebuilder workshops, what probs did Hartford have (Cognito?)

    • do we need to rethink the partitioning of the systems to make it easier to install 

    • how do we make this as simple as possible

    • Who at AWS was discussed ? (not LF contact)

    • Only anecdotal 

    • Cognito issue

    • Based on exp T&Y had, use their experience to shed some light on the distinctions

    • Not just Cognito, also secrets manager, applications need to be decoupled - needs to be pluggable approach

    • another iteration to make it more deployable

    • could use for open source deployments (there is a F/OSS competitor to cognito)

Discussion Items:

 

 

Time

Item

Who

Notes

 

 

 

 

Goals:

 

Action Items: