How do APIs and IoT relate? The answer is not as simple as merely adding an API on top of a dumb device, but rather about understanding the architectural patterns for implementing an IoT fabric. There are typically two or three trends:
Exposing the device to a management framework
Exposing that management framework to a business centric logic
Exposing that business layer and data to end users.
This last trend is the IoT stack, which involves a new shift in the separation of what stuff happens, where data lives and where the interface lies. For instance, it's a mix of architectural styles between cloud, APIs and native hardware/software configurations.
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
IoT architecture
1. Planning an architecture for the
Internet of Things
IoT Expo , Nov 5, 2014
Sumit Sharma
Director, API Solutions
sumit.sharma@mulesoft.com
2. Leading connectivity platform for
enterprise applications, mobile and IoT
HQ in San Francisco with offices in New York, Atlanta, London, Rotterdam, Munich,
Sydney, Singapore, Hong Kong, Buenos Aires, Rio De Janiero
2
3,500+ on-premise enterprise deployments
25,000+ cloud deployments
50% of the Global 500
www.mulesoft.com
10. At a high level this is the general IoT stack
App
Data Processing and
Platform
Edge
Thing / Device
11. Breaking down the
IoT stack
MuleSoft Confidential - please do not share/distribute 11
12. The IoT Stack
Mobile apps
Mobile aPaaS
Application PaaS ( aPaaS )
Data Management and Intelligence
Device
Management
Hardware / Firmware
API
Design / Build
Sensors
Device
Hub/Gateway
API runtime
management
iPaaS
Middle-ware
Websites
Industry specific
( e.g., appliances, touch
console etc.)
21. Reference capabilities for a gateway
Connectivity
Routing
Enable scalable, real-time, dependable, high-performance
and interoperable data and
device management related exchanges
between publishers and subscribers
Registry
Software mgmt
Control Events Actuator
Aggregation Transformation Provisioning
22. Device, and Device gateway sprawl is going to be a challenge
Too many disparate
ecosystems. Too
many gateways,
hubs, protocols, apps.
23. Solution to the sprawl: A hub of all hubs
Need interoperability
between devices/
machines so they can
all talk to each other.
26. Capabilities required for Data Management and
Intelligence
• Data collection, storage, and analysis of sensor data
• Run rules on data streams
• Trigger alerts
• Advanced analytics/machine learning
• Expose HTTP (REST) APIs
Data, HTTP,
connectivity
Real time event
processing
Batch processing
Data enrichment
Routing and
Orchestration
BigData solution
connectivity
Pattern Discovery/
Model re-training
Driving Forces
Identification
Predictive Analysis
28. API lifecycle tooling can be split between
design time and runtime
Rapidly design, deploy and publish APIs
API
API runtime
Design / Build
management
29. API lifecycle: Design time capabilities
Rapidly design, deploy and publish APIs
API spec
creation
API design
lifecycle
API mocking/
modelling
Reusable API
patterns
Deployment
automation
API
Design / Build
API runtime
management
32. API lifecycle: Runtime capabilities
Rapidly design, deploy and publish APIs
Rate limiting /
Throttling
API SLA
management
Custom policy
engine
Multi-tenant org /
RBAC support
Deployment
automation
API and data
security
API
Design / Build
API runtime
management
42. One final thought: the stack as it exists today is also
converging…
App
Data Processing and
Platform
Edge
Thing / Device
43. Scenarios where the middleware and edge have converged
( i.e., MuleSoft Anypoint Edge )
AApppsp
Data Processing
and Platform
Edge
Thing / Device
44. And there are also scenarios where the app layer is directly
connected to the Thing/Device layer ( i.e., embedded
Android, Java, Javascript etc. )
Data Processing
and Platform
Edge
Apps
Thing / Device