architecting iot for the cloud - a case study
DESCRIPTION
This presentation describes how FramTack architected their IoT product to send and receive data between IoT devices and cloud databases using an open source REST API platform called DreamFactory. This was presented at the Bay Area Open Source Software meetup at SAP on October 29th, 2014 http://www.meetup.com/Bay-Area-Open-Source-Meetup/events/211202322/TRANSCRIPT
About DreamFactory
Open Source Software Apache licenseQ4 2013 - Version 1.0 Q1 2015 - Version 2.0 Strong developer and cloud vendor adoption
REST API PlatformRun-time server softwareAuto-generates APIs for SQL, NoSQL, file storage
Use CasesREST APIs and server-side security for enterprise mobile apps REST APIs for IoT data
Development Process
DreamFactory provides REST API Services to your
data
Build apps for phone, tablet, desktop or IoT
device
Install Connect Develop
+ =Install
DreamFactory on IaaS cloud, PaaS cloud, or server
Unified REST Interface
SQ
LN
oS
QL
Files
SQL
NoSQL
Files
Fragmented APIsDreamFactory
FramTack IoT Case Study
Software Vendor
Solution Family Product for IoT Solution Engine for processing IoT dataSolution Builder for configuring data collectors,
rules, and statistics Reduces cost and time required to build IoT engine
yourself
Building Automation Use Case
6
Solution Family Suite
Edge
SolutionEngine®
Data Model
Clouds
Storage
Analytics
Appliances
IoT Data Flow
2. Analyze Data
Solution Builder®
1. Get Data
3. Send Data to/from Cloud
4. Control the Appliance5. Build Dashboards
Building Automation Example
Intel Gateway
Intel Gateway + PLC
Temperatures Pressures
Solution Engine®
Steam Room
Intel Gateway + PLC
Temperatures Pressures
APT1 Lobby
Pump Room Space Temps
Electric Meter
Analytics
From Sensor to End User
Solution Builder
Solution Engine
Mobile App Dashboard
DreamFactory Admin Console
Service Platform
Solution Family Products
IOT Data to Cloud via REST
Alerts and Analytics via RESTDB Connection, Schema, Data
10
Dashboard Builder
7/10/13
Discussion
Data explosionWhat data is actually useful for end users?
• Transactional vs aggregated data• Tolerance thresholds for alerts• Learning from false positives and false negatives
Where does data processing occur (e.g. gateway vs cloud)?• Complexity of analysis• How transient is the data (e.g. one day vs one month)?
IoT trade-offs Business Requirements – e.g. what data matters, what
frequency?Cost – e.g. API calls, bandwidth, storageSpeed – e.g. how “real-time” must the data be?Scalability – related to data explosion considerations
above