GraphConnect 2018

Filling in the gap of our database offerings (e.g. graph) we were inspired by our past experience to bring NoSQL into the financial industry. As with NoSQL, we started with 2 use-cases - one small and one big. Small enough to learn, prove, share and deliver. Big enough to challenge, imagine, impress and expand.

With challenging requirements in availability, scalability and global reach we needed to reconsider our architecture. A strong principle starting point for customer facing services is that we adopt a master-less architecture where possible. In our use-case(s) availability is often more important than consistency however as time goes by and data quality degrades, consistency becomes more problematic.

Imagine, one could design an architecture to remove throughput as a challenge, eliminate migrations and ensure consistency over time by means of re-deployments. We call this the cache cattle pipeline and started viewing our ‘datastore’ as multiple technologies to provide a total solution meeting our ultimate demands for the global reach.

In our talk, we will share the use-case(s), architectural overview and the paradigms adopted to help us bring graph database to life in our organisation.