A pulse is missing. Is it the patient? Or is the pulse monitor not functioning?
Life-and-death IoT systems literally have no margin for error.
How are readings from health sensors merged and analyzed immediately?
How can single point of failures be eliminated?
An IoT backbone to connect sensors, apps and analytics into a responsive system is needed. It has to be secure, flexible and scalable.
The challenge of combining data from multiple sources
Medical errors are the third leading cause of death.
Most medical devices operate independently. It’s hard to combine information from multiple devices to understand a patient’s condition. Devices suffer many false alarms. Fatigued healthcare staff members silence the alarms, misconfigurations go unnoticed, dangerous conditions go unaddressed—and people die.
Consider a patient undergoing treatment. Pulse oximeters track blood oxygen saturation and set off an alarm when it’s low. This reading is more likely to be a concern when the patient’s breathing rate is also low. Readings from both devices should be considered before generating an alarm. Medical staff can then focus on patients who most need their care, rather than chasing false alarms.
But medical devices and sensors generate data in different formats and at different rates. How can these diverse data streams be merged and analyzed quickly? Cloud-based analysis of sensor data introduces latency that hampers timely decision making. A different approach is needed.
A growing need for real-time integration of sensor data
Many industries need real-time integration of sensor data, including naval systems, avionics, power, medical devices, consumer electronics and industrial control. Three examples include:
- Automotive: Sensors in a car detect an approaching vehicle. What corrective action should be taken?
- Manufacturing: Sensors on the production line indicate a slowdown. A motor is overheating. Loud sounds are detected. What should be done?
- Transport: Mechanical and electrical system sensors on a ship indicate a problem. Should the power be turned off?
Skilled Analysts The data integration challenge
Merging real-time data streams from two sensors to determine a patient’s condition is hard. It’s even harder when inputs from other devices are included.
Of course, the problem isn’t new. Enterprise applications are integrated with enterprise service buses (ESBs). ESBs translate application-written information into a common format that can be shared with other applications. Applications read/write information into a common format that can be shared with other applications. This avoids having to write customized interfaces between all of the possible combinations of applications in an enterprise that need to exchange information.
ESBs or distributed databus?
ESBs aren’t viable for most industrial IoT (IIoT) applications. ESBs connect large-grained enterprise systems that execute a few transactions per second. An ESB also represents a single point of failure. IIoT systems need faster, smaller-grained and more reliable services.
Tech from different providers must work well together
IIoT solutions require that technologies from different providers work well together. The Industrial Internet Consortium (IIC) was founded in 2014 by AT&T, Cisco, GE, IBM and Intel. It is a nonprofit group that promotes collaboration through open interoperability standards.
“Industrial systems need to identify, describe, find and communicate a lot of data with demands unseen in other contexts. Many applications need delivery in microseconds or the ability to scale to thousands or even millions of data values and nodes,” explains Stan Schneider, CEO of RTI and member of the IIC steering committee. “The IIC targets technology to enable new mission-critical IIoT applications.”
RTI The IIC’sIndustrial Internet Reference Architecture(IIRA) has a connectivity framework that connects parts intelligently so the system can perform, scale, evolve and function optimally. The framework provides for data discovery, exchange patterns and quality of service (QoS). Here’s how it applies in our patient care example:
- Delivery (Reliability and re-delivery):How can a patient’s pulse rate be delivered dependably for analysis ?
- Timeliness (Prioritize and inform when information is “late”):How can we guarantee that information from healthcare sensors arrives in time to make a decision?
- Ordering (Delivery in the order produced and received):How do applications know which event happened and in which order? How can we understand the distributed sequence of events?
- Durability (Support later joiners, survive failures):What happens when we add or restart new algorithms? How do they get updates from all of the sensors attached to the patient?
- Lifespan (Expire stale information):How do applications know the data they’re using is fresh? How can we keep older sensor readings that aren’t relevant from impacting the care being delivered?
- Fault tolerance (Redundancy and failover):How can we ensure that communication or subsystem failures don’t disrupt care?
- Secure (Ensure confidentiality, integrity, authenticity and nonrepudiation):How can we make sure a patient’s healthcare information is kept confidential and secure? How can we protect the system from hacks?
Data distribution service (DDS)
The Object Management Group (OMG) and DDS standard is a connectivity framework that addresses the needs of many industries:
- Power systems (huge hydropower dams, wind farms, microgrids)
- Medicine (imaging, patient monitoring, emergency medical systems)
- Transportation (air traffic control, vehicle control, automotive testing)
- Industrial control (SCADA, mining systems, PLC communications)
- Defense (ships, avionics, autonomous vehicles)
DDS provides automated discovery of new IoT devices connected to the system and their QoS settings.
Industrial Internet Reference Architecture (IIRA)
The IIRA defines an architecture for IIoT systems. It specifies a core connectivity standard that implements reliable, high-speed, secure transport and QoS described above. Some endpoints connect to the core standard directly, while others connect through “gateways.” This allows sensors to work on different protocols and need only one bridge to connect to the core. This eliminates the complexity of creating bridges between all possible pairs of endpoints.
Real-Time Innovations (RTI)
RTI’s Connext DDS connectivity databus provides low-latency, real-time QoS and high availability.
“GE Healthcare is connecting medical devices, cloud-based analytics, and mobile and wearable instruments. The future communication fabric of its monitoring technology is based on RTI’s data-centric Connext DDS platform,” saysMatt Grubis, chief engineer of GE Healthcare’s Life Care Solutions.
Acute care innovation
The Zuckerberg San Francisco General Hospitalis a safety net hospital and the only Level I Trauma Center for 1.5 million San Francisco residents. Dr. Christopher R. Peabody, MD, MPH is a practicing emergency physician and directs the UCSF Acute Care Innovation Center.
“Alarm fatigue is a huge issue in patient safety. Our new Innovation Center promotes collaboration between technology and healthcare providers to resolve it,” he says. “IoT done right, can dramatically improve acute care delivery!”
IoT has changed healthcare. Maximizing its impact requires that sensors, analytics and apps are tightly integrated. RTI’s Connext is the DDS integration platform of choice for IIoT systems.
This article is published as part of the IDG Contributor Network. Want to Join?