An elevator pitch to executives from health care sector to get them to migrate their data to the AWS cloud. This was an academic exercise and the focus was on making a case in a very short amount of time (less than 20 minutes)
2. The top 3 key driving factors
▪ Healthcare reforms and compliance (HIPAA, FedRAMP, etc.)
▪ Exponential data growth (cost and scalability)
▪ Security risks (cyber attacks, data theft, etc.)
3. Key design criteria
▪ Flexible ingestion of data
▪ Data resiliency including backup and DR
▪ Ability to protect and secure critical data
4. Our four steps approach to data migration
Ingest Store Protect Optimize
5. Ingest
We have three modes
available for data ingestion
▪ Batch (via snowball)
▪ Continuous (via DMS)
▪ Hybrid (via SGW)
Our
Data
AWS
Cloud
AWS Database Migration
Service
AWS Snowball
AWS Storage Gateway
9. Store
Once the data is ingested, we
have various options to store it
on AWS.
▪ Block storage (EBS)
▪ Object storage (S3)
▪ Databases (DDB, Aurora)
Amazon Elastic Block
Store (EBS)
Amazon Simple Storage
Service (S3)
Amazon Aurora
Amazon DynamoDB
10. Protect
To protect the data stored on AWS, we will utilize multiple security features
on AWS.
▪ Encryption: AWS already provides encryption for data at rest. For
encryption of data in transit, we will use using IPSec and/or SSL/TLS
▪ Monitoring: We will set up monitoring of our data usage via Amazon
Cloud-watch metrics and utilize Cloud-watch events for notifying
stakeholders of unwarranted activity.
▪ Controlled access: We will utilize AWS Identity and Access Management
(IAM) roles and policies to grant controlled access to our AWS
resources.
11. Optimize
Once the data is stored and protected on AWS, we can utilize AWS
analytical services to optimize the data and build analytics.
▪ Scaling: S3 scales automatically for us. For Dynamo DB, we will set up
provisioned IOPS such that we can always handle growing customer
traffic. For any EC2 instances we will set up auto-scaling.
▪ Optimizing: For structured data, we will use Redshift to build robust
analytical software and for unstructured data, we will use Hadoop
clusters and build machine learning applications.
12. Conclusion
We have catered to multiple needs by migrating our on premise data to the
AWS cloud.
▪ Scaling: We have used S3 and Dynamo DB to seamlessly scale to meet
the exponential growth of our data.
▪ Cost: We have significantly reduced our operational and storage costs
by moving on premise data to the AWS cloud.
▪ Security and Compliance: We have met HIPAA compliance and provided
robust security to our data by adopting AWS secure technologies.
▪ Optimizing: We have integrated with Redshift to enable future use of
technologies like machine learning, business analytics, etc.
Editor's Notes
Cloud providers like AWS make it easier to become HIPAA and FedRAMP compliant
Every 73 days, the amount of health care data doubles. We need a sound strategy to manage the scale and cost effectiveness (https://www.foxbusiness.com/healthcare/how-big-tech-could-revolutionize-the-healthcare-industry)
Ingest – collect / gather data from our sources
Store – transfer to AWS cloud storage
Protect – Secure the data, make it compliant
Optimize – scale the storage, visualize, business analytics etc
Use AWS Snowball to transfer large datasets into AWS in a batch mode. Very cost effective but not real time.
DMS can perform continuous replication from the source to destination. We pay for network costs apart from the storage and transfer costs.
SGW is installed on premise and it continuously keeps in sync with the AWS storage
Block storage is good for RAID arrays, VMs etc
Object storage is cheap and fast
Databases are best for storing relational data
IPSec is a network layer encryption
SSL/TLS is application layer