ePay TPR1

2022-05-17

Attendees

• Adam Cable
• Rex
• Karl Newitt
• Mark Wilmshurst
• Andy Townsend
• Richard Hague
• David Williams
• Edgar Schouten
• James Coates
• Richard Kennard
• James Heaton

Discussion

• TPR schedule
• ePay upgrade

Adam

• old, needs rebuild

  • Oracle -> Postgres
  • Oracle WebLOgic -> ? Boot?

• front door(eCapture) -> payment providers (SecurePay, PayPal, adyen)

  • hold enough info to do repeat payments
  • exposes various web hooks and batch interfaces to common api

• only eCapture needs to be PCI compliant (holding card details)

  • in fact invocation of payment service provider done client side (but still needs to be PCI compliant)

• so epay holds: payment amount, frequency, tokens from PSP

  • short token
  • long token

• all info about how may use token is in ‘commerce store’

  • key is ‘order ref and xToken (either short or long)’

• question over value of Kibana (50-60K / year)

  • purely debugging, expect to use NewRelic

• payment gateway

• Arch level 2 not 3 because potentially impact RX

  • seek to maintain compatability
  • potential issue if go ahead and rotate secrets
    • will have to invite the other leads

• confidential not restricted

• have heard issues with Java on Cortex

  • James don’t exclude cortex
  • Are there any reasons not to go to cortex
  • Stateless MS arch
  • Don’t even know the issues with Java, just rumours
  • Agree to follow up with Adam
  • concern over need to retrain team
  • Neil Burroghs (EIP?) already in Java space

• risk of outage

  • can’t take payments! lose sales
  • taking epay offline necesitates taking ecapture offline too
    • can we make it more losely coupled (add token at end of ecaptrue to q)

• Java issue: resource contention on startup

• Desired timeline:

  • start June (optimisitic)
  • live by year end
  • peak times to avoid (sept and xmas maybe)
  • TPR2 2-3 weeks
  • BAU not ‘project’ means lesser business case
  • necesitates cortex multi-tenancy, but pilot on that soon, not epay