2. How we used to work
Infrastructure
1000’s Servers in 3 Data
centers.
Monolithic Apps
Very little flexibility and
customization options.
Security
Never ending patch cycles!
Athlete Impact
Features and bug fixes
affected by operational and
long deployment cycles.
3. How we work now!
Multi-Cloud
Working with VSphere,
Azure and GCP.
Automated
Deployment
Deploying in minutes
not months.
Product Focused
Breaking silos and
empowering our teams
through labs.
Security
Patching is done on
demand and without athlete
or teammate impact.
4. Why Is Search Important?
Fundamental
Product search is
fundamental to ecommerce.
Influential
Search influences athletes
shopping behavior.
Engaging
More than 50% of people
that visit DICKS.com
engage with search.
6. The DSG Search Solution Journey
May 2018 Oct 2018 May 2019
Cloud Native Application Architecture | Spring Framework | Highly Scalable Search Engine |
MySQL as Backing Service
Batch Process
Source of truth
Search Cluster
MySQL Service
7. The DSG Search Solution Journey
May 2018 Oct 2018 May 2019
Legacy
Solution
3rd
Party
Search
Engine
DSG
Search
Solution
DSG search solution
● Outperformed the other search engines
● Handled 100% of holiday traffic on
Cyber 5
● 250 requests/sec on search
● 2500 requests/sec on search
autocomplete suggestions
8. The DSG Search Solution Journey
May 2018 Oct 2018 May 2019
Multi-cloud failover in place
● From: 2-3 hours to
failover manually
● To: Automated failover
with zero human
intervention
9. Our Current Successes
CI/CD
Complete automation of
deployments.
Better Page
Performance
Reduction of 4+
seconds in page
load time!
Innovation
Creative solutions that
makes business impact
Real Time
Updates
Business has the ability
to change based on
market trends
Resiliency
Multi-Cloud
Strategy
17. Minimize manual toil for data
Move more workloads to the
platform
Reduce the manual overhead of
managing, upgrading, and
monitoring data services. Pivotal Platform
18. Pivotal Platform
Real-time updates to search results
Batch process → Event streaming
Continuously stream new search
data so in order to more quickly
add or remove products from the
website.
19. Pivotal Platform
Personalized search results
Use data to derive conclusions
Implement machine learning
algorithms to detect patterns and
provide personalized results to
Athletes.
23. Automate wherever possible, deploy fast and often, and go multi-cloud.
Componentize your architecture to allow for iterative development.
Lessons Learned
24. Automate wherever possible, deploy fast and often, and go multi-cloud.
Componentize your architecture to allow for iterative development.
Rethinking your data strategy is vital to continue delivering customer value.
Lessons Learned