Lumada IoT Architecture Today and Tomorrow Rob Tiffany CTO and Global Product Manager/Lumada IoT 9/18/2017
Session Objectives and Takeaways Session Objectives: ▪ Discover the baseline architecture found in most IoT platforms. ▪ Learn about Lumada’s high-performance architecture that integrates with a customer’s existing systems. ▪ Get a glimpse of the highly-distributed, global future of Lumada. Key Takeaway: You should be able to describe a secure, global IoT system designed to deliver customer value at unprecedented scale.
Common IoT Architectures
Generic IoT Platform Architecture IoT IoT Cloud Message Device Auth Gateway Queue SDK Do Unique Device ID Something IoT Device Registry Security Token
Let’s Build it in the Cloud! Analytics Device Registry Cloud Gateway Thing Thing Thing Thing Thing Thing Thing Thing Thing Thing
Lumada Architecture
Lumada at a Glance Intelligent Composable Secure Flexible OT Systems IT Systems Studio Machine Data Alerts / Application Dashboard Notifications Enablement Business Data Edge Core Analytics Asset Artificial Data Collection Integration Intelligence Asset Batch / Stream Human Data Real time Management Analytics Control Data Blending Data Caching / Asset Avatar / Orchestration Filtering Edge Foundry Analytics
First We Start with an Asset Avatar Type ▪ An asset avatar type is a digital blueprint for a class of physical assets of the same type. Asset Asset Avatar Asset Asset Asset Avatar Avatar Type Asset Asset Avatar ▪ Asset avatars are an instance of an asset avatar type and inherit a subset of properties and metadata.
Then We Derive Asset Avatars ▪ An asset avatar is a digital representation of a physical asset. Physical Digital Sensor Sensor Value Asset Sensor Sensor Sensor Sensor Asset Value Value Avatar Data Sensor State Asset Asset Avatar Query Asset Avatar Asset Service Avatar Sensor Sensor Sensor Sensor Avatar Value Value Sensor Sensor Value ▪ It serves as a digital proxy for physical assets with continuously updated sensor values to meet the needs of data consumers.
Portability to Meet Customers Where They Are ▪ Not all of our Industrial IoT customers are prepared to move their data and operations to the cloud, so Lumada runs everywhere. Core Core Assets Assets Assets Assets Edge Core Assets Assets Edge ▪ Lumada runs fully on-prem, in a hybrid on-prem to cloud configuration or completely in the cloud as needed.
Flexibility to Work with Customer Systems ▪ Many customers have already made a sizable investment in their own apps, analytics, data management and visualization systems. Visualization Databases Asset Data Avatar Integration Analytics Cloud CRM Apps Data Consumer Message Queue Data Query Connectors Service Real Time Subscription Alerts ERP ▪ Customers can integrate with Lumada via the asset avatar data query service, the real-time data consumer queue and data integration connectors.
Lumada 2.0 Architecture Private or Public Cloud (IaaS) APIs APIs OT Data Core Analytics Studio Active Directory Interactive Assets Machine Real Time Telemetry Asset Avatar Data Dashboards MQTT AMQP HTTP/REST Learning Asset SDK Reporting Python, R, Weka Query Service Batch File Telemetry for Securable IP Asset Registry Spark Execution Security Token TLS Endpoint Devices Assets + Gateways + People Edge Engines w/ Compute + Security Tokens Enterprise Edge SDK Storage Integration for Securable IP Edge Devices w/ Data Pipelines Asset Avatars Initiate Actions Serialization of Compute + Storage Properties + State + History Workflow Engine Sensor Values Data Sink in Customer KPIs + Thresholds + Rules Business Solution Cores Protocol Systems via Update Avatars Adapter Framework Queue Subscription Data Integration Producer Message Queue Connectors OPC UA HTTP Data Replay CAN Bus. Proprietary Update Queue TCP UDP Streaming Consumer Message Queue Real-Time Alerts Analytics via SMS, Email Data Producers and Mobile Push Unstructured Storage Notifications Data Blending Data Lake Business Data CRM Data Consumers Cloud Data Integration Connectors CSV Data JSON ERP Microservices + Clustering + SSO + Authentication Foundry Authorization + Multi-Cloud Deployment + Logging IT Data
The Future of Lumada
The Future of Asset Avatar Types If it’s a Child of a complex Parent Asset Avatar Type, it will also contain the Unique ID of the Parent Asset Avatar Type to maintain the Parent/Child relationship Baseline metadata about this type of asset Asset configuration information including network interface, protocols, data formats, parsing and data transformation information required to initiate telemetry ingestion Asset Names, data types, and units of measure for every sensor and actuator exposed by the asset Avatar User-defined KPIs, thresholds, and rate of change velocity per sampling for each sensor Type Code to be executed against incoming telemetry since data structure is known in advance Analytic engines or runtimes used to execute code Latest versions of Software and firmware plus the path to retrieve updates New or updated metadata contributed by one asset avatar to be shared with all asset avatars of the same type
The Future of Asset Avatars URL of authoritative upstream server node The Unique ID of the Asset Avatar Type it is derived from Location (GPS or other) If it’s a Child of a complex Parent Asset Avatar, it will also contain the Unique ID of the Parent Asset Avatar to maintain the Parent/Child relationship The Unique IDs of one or more groups the asset belongs to Asset Baseline metadata about this particular asset Avatar Activated + Retired Flag + associated dates and data retention policy Ongoing maintenance records Names of every asset sensor along with current and historical values Names of every asset actuator to receive commands from authorized services and relay them to the actual asset Current version of Software and firmware to compare with Asset Avatar Type version New or updated metadata to be shared with the Asset Avatar Type it’s derived from
Moving to a Global Offering ▪ Lumada will move to a distributed, global cloud model that looks like the Internet using data replication services that work like the Internet Domain Name Servers (DNS). ▪ Administrators can create and maintain groups, roles, asset avatars, analytics, ML models, and global instances of Core, Fog, Edge and Assets with a single SaaS system. ▪ Hitachi and partners will connect Assets to Edge and Fog infrastructure. Hitachi and partner domain experts will create Asset Avatar Types. Hitachi and partner data scientists will pick Machine Learning algorithms and create/train models. Hitachi, partners or the customer will manage and monitor the entire environment from a single cloud location.
Solving for Global Scale Global Company Machine Two Machine Machine Three One Factory One Japan Machine Four Assembly Line Tokyo One Osaka Machine Factory Five Two Kyoto Assembly Factory Line Three Two Assembly Line Three
Smart Connected Assets Analytics Analytics Smart Connected Assets Sync Identity Producer Identity Studio APIs APIs Sec Sec Fog Consumer Avatars Avatars Edge Sync Identity APIs Sec Data Sync Avatars Japan Core ML ML Model Model Training Training Analytics Americas Core ASIA Core Fog Fog Avatars Identity Avatars Identity Edge Edge Producer Producer Sync Identity Sync Identity Sync Identity Sync APIs Identity APIs Sec APIs Sec APIs Sec Sec Consumer Consumer Avatars Avatars Avatars Avatars Sec Sec APIs Analytics Studio Analytics Studio APIs ML ML Analytics Model Model Training Training Smart Connected Assets EMEA Core Identity APIs Sec Avatars Sync Edge Analytics Analytics Sync Producer Identity Identity Studio APIs APIs Sec Sec Fog Consumer Avatars Avatars Smart Connected Assets
In Review: Session Objectives and Takeaways Session Objectives: ▪ Discover the baseline architecture found in most IoT platforms. ▪ Learn about Lumada’s high-performance architecture that integrates with a customer’s existing systems. ▪ Get a glimpse of the highly-distributed, global future of Lumada. Key Takeaway: You should be able to describe a secure, global IoT system designed to deliver customer value at unprecedented scale.
Q & A
Thank You
Recommend
More recommend
