Planned Architecture Changes

Goals

High-availability covered at architecture level
- Asynchroneous processing of Master and Agent components
- Outages of components/connections are covered at architecture level

Distributed architecture replaces component clustering
- No single point of failure
- Automated reconciliation and recovery

Database independent architecture
- Recovery files used for restart of Master and Agent, consistency is provided at application level
- No dependency from database availability, common DBMS products are used for reporting purposes only
- No costly database management products for clustering required
- No administration and on-going management of database required

Master/Agent do not require a permanent connection
- each component works asynchronously and independently
- a connection should be established at least once per day
Master
- control a number of Agents
- act as the central access point to Agents, e.g. for the JOC Cockpit GUI and Web Services
- can be terminated and restarted during ongoing operations
- control the daily plan, what to run, when and where (but not how)
- operate as a singular instance
  - multiple Master instances are operated independently from each other
  - multiple Agent instances can be shared across a number of Master instances
Agent
- execute tasks independently from the daily plan
  - with the daily plan being forwarded to an Agent Cluster the included job chains can be executed independently from a Master being available
  - availability of a Master is required to perform pre- and post-processing scripts that e.g. check the job history
- operate in a high-availability cluster