Please enable JavaScript.
Coggle requires JavaScript to display documents.
Camoflage Zebra ( Hot failover redundancy) - WALK (SCOPE (Datacenter or…
Camoflage Zebra ( Hot failover redundancy) - WALK
FAILOVER
What does it look like ?
TS Backup - Concept
2 of everything
Streaming data ( or DB )
SLA - 15 MINUTES or less
Trigger off of a new health service (or backup triggering to live service)
Automation
Monitoring / Triggering
Circuit breaker
3rd party open source, i.e. netflix
alter dev / ops of failover occurance
VALUE
KISS ( system refactoring)
depracate MBTX TS
deprecate - MBTX Handler
Service - Switch all backend applications too
SCOPE
Invest ONLY
What
Datacenter or Application - BOX
Focust: DataCenter
be able to support - setup in more than one data center
Support Partioned
No Backup/Backup after failover event
Requirements
CI/CD
Auto Build
Push Button Deploy
Orchestration
Wish list - over time
POC and start to build in unit Test Cases
Automated regression testing against QA
MONITORING
system is getting very complicated - need auto monitoring
GCC support
Dashboard Mgmt
#
display of stats that we see now in gui
handling Management commands that are in the gui
Tradedesk or other ability to have a gui(admin) to manage Order routing
route.ini
options routing
wheel changes
SQL - deployment Standards
RedGate
Source Control
deploy scripts
Evo - db design
DESIGN
HOW TO KEEP hot hot in sync
Msg Queue
database ORM
COMPLICATIONS
SPLITTING MW to more than one TS (FIX
can Mitigate: better Design : Add FIX to Frontliner - 50%done
HANDLERS - REMOTE
WALK
Move Half of handlers to Chicago
Both systems route to all same handlers
Router takes care of dest redunancy w/backup routing
RUN
Keep handler in backup state as well
Run 1- Connect to same dest IP;port:compID
SPRINT
Dependant on destination - everyone does it different way
How do you move people
Getting new Open source includes - included by Ally
RUN phase - hot hot
GCC spin
where do they live.. Can't duplicate
Stop & triggered orders