A guide to AI, machine learning and new workflow technologies at HIMSS17, Part 3: DevOps and workflow

Charles Webster, MD, MSIE, MSIS (a HIMSS17 Social Media Ambassador and @wareFLO on Twitter) researched which vendor is showing what, and where on the show floor. The final part of his series focuses on container and microservices orchestration technology.
By Charles Webster, MD
07:21 AM
AI machine learning workflow HIMSS17

I recall continually reinstalling operating systems and software to reproduce and fix EHR software bugs. Then came long virtual machines. All I had to do was spin up a fresh image, and proceed. The problem was that running an operating system, say Windows OS, on a Macbook was slow. Each virtual machine used so many resources I had trouble getting other work done. Then containers came along, which I thought of as just the subset of an operating system necessary to run my applications. They were fast and lightweight and I could run a zillion on the same laptop. Increasingly, applications are not just developed and debugged, but also deployed in containers to the cloud.

Microservices predate containers. They remind me of remote procedure calls and method invocations during the 90s. I could call a procedure or a method out there on the network somewhere and return the result directly into my program. That approach faded with the advent of the Internet and RESTful webservices. Microservices are like RPCs and RMIs, but not they are not tied to specific platforms. Furthermore, the microservice approach emphasizes breaking apart monolithic programs and representing individual business goals and tasks as individual software services. There is natural fit between microservices and containers. Deploy microservices to containers, then containers to the cloud. Lots of problems solved. 

[Also: A guide to AI, machine learning and new workflow technologies at HIMSS17, Part 2: chatbots and workflow]

There is just one thing. Something has to orchestrate the containers and the microservices. DevOps (for development-operations) is increasingly process-aware, relying on models of the workflow of the staging of containers to the cloud. There is also a great fit between microservices and workflow technology, since workflow models represent interactions among business goals and tasks, and microservices implement these goals and tasks. It’s a short step to workflow engines orchestrating microservices.

We are enduring a phase during which APIs are being added to monolithic applications, particularly EHRs. FHIR (Fast Healthcare Interoperability Resources) is one of these APIs. Mobile apps and other services and devices, such as chatbots and wearables, will access and generate EHR and health IT data. At the same time, monolithic applications are being taken apart and modularized, or eventually replaced by more modular, and therefore more maintainable and scalable systems.

At some point, these new bits-and-pieces must be knitted together into coherent, usable workflows. This is where workflow technology comes it. Something has to trigger and track the development, deployment, execution, and interaction of all this software. Deployment and management of containers, and interactions among microservices, will be more-and-more automated and process-aware. Models of process and workflow will drive both application behavior (dev-) and infrastructure operations (-ops)

It is difficult to know which HIMSS17 exhibitors are beginning to experiment with containers and microservices. One does not usually publish backroom R&D to HIMSS exhibitor profiles or public facing websites. However, the big container & microservices providers are easy to see. They include Amazon Web Services (booth 6969), Google (6975), Microsoft (2509), VMware (3661), and IBM (1809).

Welcome to the new workflow technologies of the future!
Machine learning, chatbots, and microservices … what do they have in common? Workflow, dataflow, process, and orchestration — of a complicated flow of complicated interactions among complicated software services. We will see this pattern play out throughout IT and health IT. For example, many users do not like their EHR because they do not like their EHR “screenflow.” What if the flow of EHR screens could be represented in a model of the user’s workflow? The workflow model could be executed, or at least automatically consulted, to more intelligently offer the right screen to the right person at the right step in a workflow. If a user is unhappy with their screenflow, well, just pop open a workflow editor and fix it! 

[Also: A guide to AI, machine learning and new workflow technologies at HIMSS17 Part 1: Machine learning and workflow]

Workflow technology is appearing everywhere in IT and health IT. Healthcare marketing automation relies on it. Healthcare interface engines rely on it. Prototypes combining blockchain and workflow technology are appearing. Heard of “SecOps? It’s basically workflow technology applied to cybersecurity, a perennial health IT concern. Even 3D printing in healthcare industry settings increasingly relies on workflow technology. As Wil van der Aalst, noted BPM researcher has written, “WFM/BPM [Workflow Management/Business Process Management] systems are often the ‘spider in the web’ connecting different technologies.” The real news here is the opportunity to emphasize data and workflow equally and strategically.

Process-aware workflow is the necessary foundation to glue all these, and other, important emerging technologies together into usable and scalable workflows serving patients and healthcare professionals. 

HIMSS17 runs from Feb. 19-23, 2017 at the Orange County Convention Center.

This article is part of our ongoing coverage of HIMSS17. Visit Destination HIMSS17 for previews, reporting live from the show floor and after the conference.

Like Healthcare IT News on Facebook and LinkedIn