A Consumer-Centric Architecture for Energy Data Analytics Rayman - PowerPoint PPT Presentation

A Consumer-Centric Architecture for Energy Data Analytics Rayman Preet Singh, S. Keshav, and Tim Brecht 1

Home Energy Data 2

Smart Meter Deployments Trials Water Electricity Gas 3 Projects Energy Retail Association

Energy Data Use E-mail “ Y our decade-old furnace is in-efficient and is costing you an additional $400 a year. Click here to see replacement options.” Text Message “It is unusual for your oven to be on at this time of day. Would you like to turn it off ? Click Yes/No” 4

Energy Data Privacy Power (W) 0 1 2 3 4 5 6 7 8 9 10 11 12 5 Time of Day

Current Situation • Utility websites "Our constitution allows us not to • Google Powermeter have that kind of intrusion on our personal privacy” • Microsoft Hohm "They’ll be able to tell if you are • Green Button running your computer, air 6 conditioner, whatever it is”

Problems Data privacy loss Frozen innovation in analytics No data ownership or control 7

Goals Consolidation Durability Data ownership Portability Integrity Data privacy Data access Universal access Flexibility Scalability Extensibility Data analytics Performance 8

Approach 1 Privacy, portability, ... App Runtime Scalability* Extensibility* Energy Consolidation Data Durability 9

Approach 2 Scalability, ... App Runtime Privacy* Consolidation Energy Durability Data 10

Approach 3 Universal Access, … App Energy Runtime Data Privacy SaaS Extensibility Flexibility 11

Proposed Architecture Data Control Cloud Based App Native App 12

Proposed Architecture Smart Meter Data Utility-owned Server Native App Web Access APIs Cloud-Based Services Control App App Privacy Data Runtime Protection Solar VHome Panel Public Internet Cloud Appliances Gateway Remote Energy Storage UIs Home EV 13

Implementation Gateway: Microsoft HomeOS Dixon, C., Mahajan, R., Agarwal, S., Brush, A., Lee, B., Saroiu, S., & Bahl, V. (2012). An operating system for the home. Proc. NSDI 2012. 14

Implementation Gateway: Microsoft HomeOS § Driver modules § Communication module § Coordinator module <setStatus classID=1 objectID=2> <power>0.0</power> </setStatus> 15

VHome Suite of web applications using JAX-RS Access Privacy Web Native Control Protection APIs Services App Mechanisms Mechanisms Apache Tomcat (Web App Container) Java è Portability across clouds (Amazon EC2, Root BSD, Windows Azure) 16

VHome Implementation Details Vhome DB using cloud datastores Datastream : (Class ID, Object ID) specific time series e.g., class – heating, object – space heaters RESTful APIs to access datastreams Token-based fine-grained access control mechanism - OAuth 2.0 17

VHome Implementation Details Access scope r/w + datastream(s) + value-based and/or timestamp-based Progress Authorization Request Prompt Allow One Time Authorization Grant One Time Authorization Grant Access Token + Refresh Token CBA ACM WS Access Token API Return Values . APIs . Access Token Invalid Token Error Refresh Token Access Token + Optional Refresh Token API access for a CBA 18

Example Applications Data scraper Energy data analytics 19

Example Applications Interactive monitoring and control 20

Comparison with Existing Solutions Green Microsoft Green Button Utility Hohm, Button OPower Web (Third Party) VHome Google (Self) Portals Powermeter 21

Related Work • Sandboxing native applications – Language based, e.g. Transmute (Griffins et al.) – System based , e.g. xBook (Singh et al.), OSN (Sariou et al.) • Dataware manifesto (McAuley et al.) • Privacy Analytics (Haddadi et al.) 22

Conclusions • Data privacy v/s data analytics – Existing solutions provide just one • Application ecosystem for home energy – Apple App Store, Google Play • Leveraging modern clouds preserves privacy , fosters application development 23