Grant Allen, CTO Chief Product Officer at Dow Jones explains how to deploy Flowable at scale in AWS.
It was presented at the Flowfest 2018 in Barcelona, Spain
1. FLOWFEST - NOVEMBER 2018PROFESSIONAL INFORMATION BUSINESS
Professional Information Business
Deploying Flowable at scale in AWS
1
2. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 2
Introduction
● Dow Jones Professional Information Business - what does that really mean?
● Dealing with content, human research, and automated processing at scale
● Our historic approach, and why we needed BPM
● Choosing Flowable - as easy as ABC (anything but closed-source :) )
● Our architecture, how we manage and evolve it
● Challenges
● Future Expansion and Long term goals
3. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 3
Dow Jones Professional Information Business
A Brief Overview
Millions of articles flow into our
research engine Factiva.
4. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 4
Dow Jones Professional Information Business
A Brief Overview (cont.)
A range of processing steps occur in
our content pipeline, variously using
tools like rules-based coding engines,
normalisation/transformation, ML
models, etc.
5. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 5
Dow Jones Professional Information Business
A Brief Overview (cont.)
Historically, we scaled with people. If
we needed to cover more content, or
extract more structured data, we’d just
add more people!
6. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 6
Dow Jones Professional Information Business
A Brief Overview (cont.)
We knew BPM was part of the
solution. While not magically doing all
of the work, it helps us standardise,
decide what a “task” really is, highlights
best areas for automation, unlocks
insight, and more!
7. PROFESSIONAL INFORMATION BUSINESS FLOWFEST - NOVEMBER 2018 7
Choosing Flowable
● Light vs Heavy BPM
● Ease of Adoption
● Other features, e.g. Forms
● Support model(s)
12. 12
Challenges
● Decoupling BPM engine from custom
workflow application
● User/Groups integration with the external
user management application
13. 13
Future Expansion and Long Term Goals
● History / archive solution
● Deeper analytics, share data with
operational and customer data lake
● Parallel deployments across other major
businesses within Dow Jones - Ad Tech
● Using endpoints from AWS Sage Maker to
use ML models