Advantages and Drawbacks of Developing Mobile Health Technologies in the Cloud Nima Kaviani - @nimak http://nima.magic.ubc.ca Director of Technology Curatio Networks Inc.
Mobile Health in the Cloud Cloud Computing HealthCare in North America, a $6.5 Billion Market by 2018 [1] General acceptance by the community Adaptation to changes in market dynamics [1] Markets-And-Markets Survey: http://www.marketsandmarkets.com/market- Reports/north-america-healthcare-cloud-computing-market-54089599.html 2
Mobile Health in the Cloud Cloud + Mobile + Health 3
Mobile Health in the Cloud Cloud + Mobile + Health Cloud Scalability Affordability 4
Mobile Health in the Cloud Cloud + Mobile + Health Cloud Scalability Affordability Mobile Personalization Accessibility 5
Mobile Health in the Cloud Cloud + Mobile + Health Cloud Scalability Affordability Mobile Personalization Accessibility Health Personal Info. Privacy Regulations US: Health Insurance Portability and Accountability Act (HIPAA) Canada: The Personal Information Protection and Electronic Documents Act (PIPEDA) European Union: Data Protection Directive 6
Mobile Health Requirements Preserve Privacy for Patient Information Security (Data Encryption) Transparency Auditability Privacy of Data 7
Privacy of Data Requirements Data is not stored out of the country (Japan s Privacy Directive) Where to keep the data Where to backup the data Data does not leave the country (Canada s Privacy Directive) How does the data flow? How to track the data? 8
How to deal with Personal Data? Assume you are a Canadian Company 9
Where to Store the Data? Trying to use Amazon Data Centers 10
Where to Store the Data? Trying to use Rackspace Data Centers 11
Where to Store the Data? Using Local Data Centers Issues: Cost Not a real cloud Potential data leak 12
Where to Store the Data? Hybrid Cloud + Application Partitioning Optimization for Deployment Premise (Canadian DC) US Cloud (e.g., Amazon or Rackspace) 13
Example Cross-tier Request Flow Pull up user information and Stock holdings for a given user id Request Query App Tier function function function function Data Tier DB Table(a) a h p DB Table(p) DB Table(h) 14
Example Cross-tier Request Flow Pull up user information and Stock holdings for a given user id Request Query App Tier function function function function Data Tier DB Table(a) a h p DB Table(p) DB Table(h) 15
Application Partitioning (Cross-Tier) Code Profiling EXPLAIN PLAN Join Orders Application-Tier BIP Create Objective Function Create Constraints Combined Partitioning Solver Code Placement Table Placement 16
Cost Modeling Execution Time: Communication Time: 17
Cost Modeling Execution Cost: Communication Cost: 18
Application Partitioning Benefits Up to 54% improvement in Cost Up to 56% improvement in Performance Drawbacks Does not guarantee privacy of data flow Impossible without proper automation 19
Analyzing Data Flow Combining Application Partitioning with Data Flow Analysis Which software elements deal with security-sensitive data Constrain security-sensitive software elements to be on private infrastructure Push everything else to the cloud 20
Analyzing Data Flow Homomorphic Encryption: Encrypt the information on the private premise Let the cloud operate on encrypted data Let the data flow anywhere 21
Conclusion Concerns: Preserving Data Privacy in the Cloud is Challenging Many companies switch back to a traditional data center model of deployment Improvements: Regulations are becoming easier to adopt Cloud technologies are becoming easier to integrate into traditional data centers. 22
Publications Cross-Tier Application & Data Partitioning of Web Applications for Hybrid Cloud Deployment. Nima Kaviani, Eric Wohlstadter, Rodger Lea. In Proceedings of Usenix International Middleware Conference 2013, December 2013, Beijing, China. MANTICORE: a Framework for Partitioning Software Services for Hybrid Cloud. Nima Kaviani, Eric Wohlstadter, Rodger Lea. In Proceedings of the 4th IEEE International Conference on Cloud Computing Technologies and Science, December 2012, Taipei, Taiwan, pages 333-340. Profiling-as-a-Service: Adaptive Scalable Resource Profiling for the Cloud in the Cloud. Nima Kaviani, Eric Wohlstadter, Rodger Lea. In Proceedings of the 9th International Conference on Service Oriented Computing, Paphos, Cyprus, 2011, pp 157-171.
QUESTIONS? WE ARE HIRING nima@curatio.me 24