Emergence of MongoDB as an Enterprise Data Hub, presented by Dylan Tong, Sr. Solutions Architect, MongoDB at MongoDB Evenings Seattle at the Seattle Public Library on October 6, 2015.
4. 4
70 Different Policy
Schemas
How can we translate
this into a Customer
View?
ETL 70 applications
into a Dimensional
Model? Integrating a
few is hard…
5. 5
Strategy: All documents can have variable schemas
db.policies.find(
{
first:”Dylan”,
last:”Tong”,
type:{
$in[“Healthcare”,”PPO”,”HMO”,”Auto”]
}
)
6. 6
Problem Emerges in Many Forms
• Redundant systems due to M&A or lack of
governance across LOB
• Mature Product Suite of silo applications,
and a need for integrated Solutions
• Operationalizing Data Lake for Online
Systems
• Data-mart proliferation and need for more
real-time sharing of data
18. 18
APP 1 APP 2 APP N
DATAHUB
CONNECTOR / ETL TOOL
DATA HUB API / SERVICES
MULTICAST UPDATES
APP DATABASES
19. 19
APP 1 APP 2 APP N
DATAHUB
CONNECTOR / ETL TOOL
DATA HUB API / SERVICESCDC
IMPLEMENT CHANGE DATA CAPTURE AND RESYNC UPDATES
1. What’s the last successful sync time?
2. Query data modified
since last success sync
3. On failure, do idempotent retry (unique index: ID, SourceSystemID)
20. 20
Solution Improves with 3.2
• Document Validation
• Mongo Compass (Schema Visualization)
• Partial Indexes
• BI Connector