Learn how to achieve holistic operational visibility into IIoT business environments by correlating the data from Operational Technology and IT, and organizing it as a single pane of glass in accordance with business processes.
Unified Analytics in GE’s Predix for the IIoT: Tying Operational Technology to IT
1.
2. True IoT End-To-End Visibility
Translating
Technology Data
to Business
Meaning
BusinessIT
3. Outline
• Phases of IoT Unification:
Operational Technology
o Challenges: Protocol/Data Model Diversity & Security
IIoT Alignment with IT Infrastructure & Applications
o Challenges: Redundant Tools & Lack of Real-time Correlation
Business Services
o Challenges: Silos of Data & Poor Service-Level Awareness
• Predix Use Case
• Recap: Centerity’s Added Value to Industrial IoT
4. IoT Projects Challenges
IoT projects impose a number of challenges that lead to very complex systems both at first deployment and throughout the
life-cycle of these systems.
• Multiple communication methods, complicate connectivity and interoperability
• Diverse data models increase ecosystem management complexity
• Service use case and ecosystem definitions change after deployment
• Custom made sensors and actuators are expensive. On
the other hand COTS (off-the-shelf) devices are affordable
but are not tailored for ecosystem’s security, service
model and communication requirements
5. Centerity IoT / IIoT End to End solution
device/sensor
connectivity
Agent
system
connectivity
Ecosystem MGMT
• Wide connectivity support for any communication
method, protocols and IT technologies
• A single abstract service model by translating IoT device
data models at runtime into business and operational
service views
• Dynamic ecosystem scalability and maintenance
• Adjusting COTS devices to the ecosystem’s security and
communication requirements with no firmware changes
• Secure device intercommunication across the entire
ecosystem
8. IIoT End-To-End Discovery, Data Collection & Translation
Thread
MQTT
WebSockets
AllJoyn
REST
HTTP
HTTPS
TCP
CoAP
BLE GATT
IoTivity
Thread
MQTT
WebSockets
AllJoyn
REST
HTTP
TCP
UDP
CoAP
BLE GATT
IoTivity
UDP
HTTPS
WiFi
Bluetooth
6LowPAN
Zigbee
Z-Wave
WiFi
Bluetooth
6LowPAN
Zigbee
Z-Wave
Translator
IT Infrastructure
Big Data & Advanced
Technologies
9. IIoT End-To-End Discovery, Data Collection & Translation
Thread
MQTT
WebSockets
AllJoyn
REST
HTTP
HTTPS
TCP
CoAP
BLE GATT
IoTivity
Thread
MQTT
WebSockets
AllJoyn
REST
HTTP
TCP
UDP
CoAP
BLE GATT
IoTivity
UDP
HTTPS
WiFi
Bluetooth
6LowPAN
Zigbee
Z-Wave
WiFi
Bluetooth
6LowPAN
Zigbee
Z-Wave
Translator
IT Infrastructure
Big Data & Advanced
Technologies
10. Outline
• Phases of IoT Unification:
Operational Technology
o Challenges: Protocol/Data Model Diversity & Security
IIoT Alignment with IT Infrastructure & Applications
o Challenges: Redundant Tools & Lack of Real-time Correlation
Business Services
o Challenges: Silos of Data & Poor Service-Level Awareness
• Predix Use Case
• Recap: Centerity’s Added Value to Industrial IoT
11. Challenges to IoT & IT Alignment
• No single solution available to monitor every technology in the environment
• Too many technology-specific tools needed
• No real-time correlation of performance data
• Result: No way of providing a single pane of glass to the entire ecosystem
15. • Big Data – Hadoop, SAP HANA etc.
• Converged / Hyperconverged infrastructure – Nutanix, Vblock, FlexPod etc.
• Cloud Technologies – Pivotal, Openstack, AWS etc.
• Others
Containers (Docker etc.)
Hypervisors (VMware, Citrix etc.)
Applications – Event logs, Syslogs, Rest, etc.
Connectivity / networking
DB (Hana, NoSQL etc.)
Storage
And much more…
Advanced Technologies Out of The Box
17. Outline
• Phases of IoT Unification:
Operational Technology
o Challenges: Protocol/Data Model Diversity & Security
IIoT Alignment with IT Infrastructure & Applications
o Challenges: Redundant Tools & Lack of Real-time Correlation
Business Services
o Challenges: Silos of Data & Poor Service-Level Awareness
• Predix Use Case
• Recap: Centerity’s Added Value to Industrial IoT
18. Challenges to Service-centric Performance
• Gaps in technology coverage
• Lack of correlation between domain-centric tools
• Complexity of enterprise environments breeds resignation
• Result: silos of technology data with no way to accurately perform cross-domain impact
analysis in real-time, and therefore giving an incomplete picture of a business process.
24. Outline
• Phases of IoT Unification:
Operational Technology
o Challenges: Protocol/Data Model Diversity & Security
IT Infrastructure & Applications
o Challenges: Redundant Tools & Lack of Real-time Correlation
Business Services
o Challenges: Siloed Data & Poor Service-Level Awareness
• Predix Use Case
• Recap: Centerity’s Added Value to Industrial IoT
25. Predix Use Case
Real-time coverage and monitoring of the Edge side
Integration with 3rd party products (VCE –Vision)
Provides end-to-end in-depth visibility to GE customers
Allows remote monitoring with managed services option
26. Centerity IoT Added-value
Measuring SLA for IoT platforms
• Business/Process Visibility across all Eco-systems
• Assuring High-level Services
• Single Pane of Glass approach
Improving Business/Process Health with minimal unplanned downtime
• Bridging Operational Technologies (OT) & Information Technologies (IT)
• Decrease Mean-Time-to-Repair (MTTR)
• Improve Troubleshooting process
• Proactive approach
• Improving efficiency
So with that being said, I’ve outlined this discussion to focus on unifying Industrial IoT environments from a performance management perspective in three phases:
1.) The operational technology or connected devices and sensors
2.) The wider IT environment that supports those connected devices and networks
3.) The actual business processes in the service of which all the technologies operate
I will then take a moment to put what you have learned in the context of GE Predix, before a brief recap and Q & A.
Thanks to all of you for attending today. I’m going to talk today about managing and maintaining peak performance for Industrial IoT environments, but to do that I want to step back from the connected devices and sensors we normally focus on and broaden our view to include everything else that allows us to capture and use the data from those devices to drive business: the network and compute infrastructure, the cloud platforms, the applications and databases, etc.
This is the crucial starting point for maintaining performance in any complex technology environment and it guides the philosophy of Centerity. Because only by viewing an IoT business environment holistically can you recognize every potential point of failure and see the true impact of one component on the performance of another.
Centerity makes the impact on your business service performance your starting point, whether it’s your wind farm, Oil Rig, or smart city street lighting, and from there you can dive into specific technologies and KPIs.
So with that being said, I’ve outlined this discussion to focus on unifying Industrial IoT environments from a performance management perspective in three phases:
1.) The operational technology or connected devices and sensors
2.) The wider IT environment that supports those connected devices and networks
3.) The actual business processes in the service of which all the technologies operate
I will then take a moment to put what you have learned in the context of GE Predix, before a brief recap and Q & A.
IoT environments are necessarily complex and that complexity presents operational challenges not just in deployment, but throughout the system’s lifecycle. So what are some of these challenges in connected industrial environments?
1.) The variety of communication protocols and data models used by these devices and sensors complicate connectivity and interoperability and make managing the ecosystem difficult.
2.) Often you might want to switch out devices for a different brand or change protocol methods, so even if you can tune the ecosystem for deployment, an enterprise system is dynamic, so your service use case and ecosystem definitions will be in constant flux.
3.) Finally, depending on the environment, you may have a mix of custom and generic Commercial Off-the-Shelf (COTS) devices, which each bring unique challenges. Whereas custom devices are fine-tuned to the environment but expensive, COTS devices are comparatively cheap, but are not designed for the ecosystem’s security, service models, or communication requirements.
So, let’s look point-by-point at how Centerity meets these challenges.
First of all, Centerity is deployed as a software-only platform comprising a single console for management and monitoring with agents sitting on edge devices or any hardware on the network to pull data from connected sensors, manage the devices, and translate data into various protocols and data models at runtime for device interoperability.
The platform supports nearly any device protocol, be it standard and low-level or non-standard and more complex. If there is ever a communication method that is not already supported in the system, our development team can build a plugin for integration in a matter of hours or days, at most.
And as environments change, Centerity is able to scale dynamically, discovering and adding new devices extremely quickly.
If COTS devices are added to the environment, you can use Centerity to quickly adjust those devices to the security requirements of the ecosystem without having to touch the firmware.
Lets look deeper at this translation process.
Centerity’s abstraction layer is able to overlay a single user-defined abstract data model across a service to manage disparate connected devices via the same model.
Using the Centerity API translator, you can bind your desired common abstract language to the device’s specific language, essentially hiding the lower-level device protocol from the service or business logic manager, and allowing for communication with other devices.
The distributed architecture of the system allows you to connect and make interoperable devices residing on different local networks in an event-driven fashion, as though they were sitting next to one another.
Expanding this virtual network requires simply running the Centerity Agent on a machine in the newly added local network. This can be deployed to connect directly to the device, behind a hardware hub, or via a service cloud.
As this diagram depicts, considering the solution end-to-end, Centerity allows you to automatically scan for discoverable devices on any network, translate their non-standard protocols into standard protocols like REST API, etc. to send data to the gateway and cloud, and apply a single abstract data model for a service across any devices in that service for interoperability and common management.
Thus, with the OT layer standardized, all supporting technologies at the Edge, in the service cloud, and on-prem can be incorporated into the environment to measure overall service availability on a single pane of glass.
With that, lets move on to the second phase of unification: Supporting IT infrastructure & applications
With that, lets move on to the second phase of unification: Supporting IT infrastructure & applications
Here you can see the environment end-to-end as a series of points of failure, all monitored through Centerity.
As I’ve just shown, the Centerity platform is able to monitor the performance data from the connected devices themselves, the connectivity to the edge and the cloud (WiFi, BLE, Zigbee, etc.), and then any IT devices, applications, or databases comprising or connected to the cloud, and ultimately present all that performance data in logical groups according to the industry or business process.
Enabling this functionality is the federated scalability and multi-tenancy of the architecture and the phenomenal extensibility of the platform’s collector engine and methodologies.
The enterprise architecture allows all distributed collector nodes and associated agents to report back to an enterprise node that manages the system. Regardless of where or how the collector nodes are deployed, the user still sees a single pane of glass according to their individual permissions.
The multi-tenancy allows for environments and services to be accessed and displayed according to user permissions, so users only see what is relevant to them.
In addition to the extensive OT device connectivity, Centerity has out-of-the-box integrations in place with all of the cutting-edge IT technologies you see here and many more. And you never have to worry about a new technology coming along that can’t be covered, because our flexible plug-ins allow us to build integrations to any technology with connectivity in a matter of hours or days.
Your business’s services are constantly changing and incorporating new innovative technologies, so the idea of Centerity is to be ready for those inevitable shifts to make sure you don’t ever have gaps in coverage.
To recap, Centerity allows you to scan, discover, and add to the system any devices or applications on the network.
You can then visualize that environment from end-to-end, correlate performance data from every technology layer, and integrate the platform with any associated service to feed data (e.g. service desk, CRM, etc.).
And finally, you can organize the environment and analyze its performance according to the actual services your business provides to its customers.
To that end, let’s now turn to the business services and how this all wraps up to give you a window into the business first and foremost, not simply siloed technology data.
What has traditionally stood in the way of service-centric performance monitoring was the inability to extend coverage to every technology that powers a given business service,
or the reliance on too many tools to derive data from those devices. Either way, the result was silos or gaps in data and no way to accurately perform cross-domain impact analysis in real-time.
Because the Centerity platform can extend to every technology in the Industrial IoT stack,
Now to put this in the context of Predix, you can see how every point in the framework is visible and unified as a business process within Centerity.