How Adopting Elastic at Bankwest was a Catalyst for Cultural Change (Elastic{ON} Sydney 2016)

Slide 1

Slide 1 text

Rob Crowley, Technical Product Owner - APIs How Adopting Elastic at Bankwest was a Catalyst for Cultural Change Bankwest 29th Nov 2016

Slide 2

Slide 2 text

Digital disruption is driving change across established industries and finance is no exception.

Slide 3

Slide 3 text

Our customers’ adoption of technology is defining our agenda Number of interactions Revenue per interaction Experience Economy Product Economy Store Monthly decreasing Internet Monthly decreasing Mobile Daily increasing Wearables Several times a day increasing Interconnected Always on increasing

Slide 4

Slide 4 text

the species that survives is the one that is able to adapt to and adjust best to the changing environment - Charles Darwin

Slide 5

Slide 5 text

AD Producer Logstash nginx FRONT END Rabbit MQ BACK END Consumer Logstash HA Proxy (Canonical) Use Case: Centralised Logging Push to Rabbit MQ cluster at opposite site via Shovel Shield Watcher ELASTIC Kibana

Slide 6

Slide 6 text

Surprise #1 – Analytics from Kibana dashboards transformed data into information.

Slide 7

Slide 7 text

“The only way to win is to learn faster than anyone else.” - Eric Ries

Slide 8

Slide 8 text

Surprise #2 - Watcher precipitated a shift from reactive to proactive support models.

Slide 9

Slide 9 text

Encryption at rest for PII data was a challenge.

Slide 10

Slide 10 text

How you tackle it depends on your company’s risk appetite. - Elastic Search Plugin (3rd Party Vendor Options Available) - Lucene Plugin (Introduced ~25% overhead) - Disk Level (dm-crypt) - Lightweight governance (& Watcher to keep us honest)

Slide 11

Slide 11 text

Surprise #3 – Elastic use cases expanded from supporting product delivery to powering end user experiences.

Slide 12

Slide 12 text

Idea + Team + Elastic = Business Value

Slide 13

Slide 13 text

(Emergent) Use Case: Customer Search Search API F5 OpenStack Elastic Data Tenant Cloud 2 OpenStack Elastic Data Tenant Cloud 1 Bulk Agent Customer Data Extract

Slide 14

Slide 14 text

(Emergent) Use Case: Customer Search Search API F5 MQ Web API Layer System of Record OpenStack Elastic Data Tenant Cloud 2 OpenStack Elastic Data Tenant Cloud 1 Sync Agent Customer Domain GET /customers/{id} HTTP/1.1

Slide 15

Slide 15 text

Cloud computing enabled a low cost iterative approach to optimising Elastic cluster design.

Slide 16

Slide 16 text

Optimizing full text queries can be non trivial. A robust automated testing process is table stakes.

Slide 17

Slide 17 text

Takeaways