Enabling the Internet of Things in the enterprise

Article

March 2016

Click image to view the Lighting up Enterprise IoT video.

The Internet of Things (IoT) is making the world around us smarter. From self-programming thermostats that detect your presence and orchestrate with other devices, to new industrial uses for sensors, IoT can increase the sophistication of practically anything using intelligence and automation.

At the heart of IoT are low-cost sensors and computing platforms that feed data to the cloud. The value of IoT is often realized by analyzing and acting upon sensor data. In the enterprise, there are opportunities to learn more about the environment that will allow you to make smart decisions, increase user productivity, and enable new experiences through IoT technology.

Piloting IoT

The Microsoft IT Innovation Group, led by Pankaj Arora, is an early adopter and is leading the way for IoT use and learning within Microsoft. The group has combined cloud services, IoT hardware, and even industrial design to light up modern experiences, increase productivity, and improve decision making as part of early IoT piloting.

For our initial pilot, we focused on an increasingly common scenario: employees working in open floor plan environments. In these spaces, the balance between collaboration and private work spaces is important. For example, if employees need to take a phone call, they may need to spend time searching for an unoccupied room. It also helps employees to be informed about the environment so they can make decisions about where to sit, because noise and other environmental conditions can be a factor in employee preferences. Furthermore, there's an opportunity for us to use IoT data to learn more about how an environment is used and to then make facilities adjustments to optimize it.

One of our objectives was also to learn more about what it means to deploy IoT in the enterprise, and apply those learnings as part of developing our internal IoT strategy for the variety of environments at Microsoft locations worldwide.

Key elements

Numerous elements came together to make our pilot successful.

Planning

To kick things off, we conducted some light-weight research to determine an initial set of scenarios for the pilot, and how to best present the information to various audiences. This research included brief user interviews and surveys. Using that information to set priorities, there were then four major workstreams: Software, Hardware, Deployment, and Operations. As part of our planning, we had to think through various factors in each category and make tradeoff decisions.

Below is a table that summarizes selected planning trade-offs followed by a few key points on each workstream.

Table 1. Planning matrix and key points

Software. From a client front-end perspective, we created a UX/UI design flow for the initial pilot scenarios based on the user research. We decided to build a modern app for Windows 10 devices, using web controls for the map to allow for easy cross-platform portability in the future. From a services perspective, we used Microsoft Azure, which provides IoT services that can handle mass amounts of data at scale while speeding up the development cycle for IoT scenarios. Then, as part of the Azure services development, we decided to experiment with machine learning for additional value, such as making predictions based on the data. Finally, we created animated heatmap reporting using PowerView for Microsoft Excel to provide insights to facilities. All these scenario and software decisions were made balancing the anticipated value with overall development effort that would be required.

Hardware. The key trade-off was between off-the-shelf, appliance-style sensor units on one end of the spectrum, versus a more customized stack with higher flexibility on the other. For the pilot, we decided on a hybrid. We used off-the-shelf hardware versus developing custom hardware. However, instead of selecting an appliance-style unit, we opted to assemble the hardware stack, write software to run on it, and do the industrial design of the packaged up hardware. For the pilot, this provided greater aesthetic, sensor, and development flexibility. At scale, appliance-style units or mass-produced, customized units may be preferred.

Deployment. Although deploying more sensors costs more, we prioritized greater coverage and even some redundancy to increase learnings—and to help right-size long-term coverage patterns. We emphasized data collection opportunities and learning, and refined deployment patterns to manage installation costs for the long term because installation can be the biggest initial cost driver in many regions. A key takeaway is to design an IoT solution with deployment effort in mind. Longer-term options such as wireless sensors are one way to reduce deployment costs, and we are experimenting with that as well.

Operations. To ease operational manageability, we invested in proactive monitoring, including device heartbeats, system state, and sensor status, to ensure the computing hardware and attached sensors were healthy. Also, we developed a support plan to address hardware issues, including the worst-case scenario of replacement. This included having a very clear inventory of where devices were deployed, and even creating logical hierarchies representing device locations such as building, floor, room, etc. This approach helps manageability and provides useful constructs that help enable the app experience, such as averaging data across nearby sensors.

The next section touches on the solution we developed, including infrastructure (software and hardware), as well as the resulting client modern app experiences and the facilities insights that were generated.

Infrastructure

At the heart of the solution are the cloud and IoT devices with sensors. We used Azure as the cloud infrastructure, which has a whole set of services for IoT including IoT and event hubs, machine learning, services to help store data, power app experiences, and managed IoT infrastructure.

We built the necessary web APIs and infrastructure using Azure capabilities. The following graphic shows a high-level diagram of the architecture, key system components, and data flows.

Figure 1. Architecture, system components, and data flow [view larger version of image in new window]

For the IoT stack, we used the Raspberry Pi platform with off-the-shelf sensors. The key sensors for this pilot were motion and sound, but we also tacked on temperature and brightness—knowing that others could be attached in the future and that there are various other IoT hardware options longer-term.

The physical environment is the canvas for IoT in this scenario, and there were important considerations both in terms of what people see and how components were installed in the environment. Being an innovative team, we performed the industrial design in-house, to determine what it means to properly install IoT in our facilities in a sustainable way, and to ensure proper aesthetics of how the sensors look in the environment. We developed and fabricated components that house the sensors and allow the Raspberry Pi devices to be mounted and installed easily and in compliance with facilities codes. We did this to learn, iterate, and have control over the design process. Throughout the process, we partnered closely with the Microsoft Real Estate and Facilities group, working at the intersection of IT and Facilities—where many IoT projects will live.

Figure 2. Sensor enclosures were designed in-house

Modern apps

All of these elements—Azure, IoT hardware, and industrial design—come together for the user in modern app experiences built on top of Windows 10. The app runs on a Surface Hub, which is centrally situated in the floor plan. The app is available for tablet and mobile devices. The app helps people find unoccupied places to work, which is particularly valuable for non-bookable spaces. It also provides real-time noise and temperature conditions, predictions on booked rooms that might actually be available, and areas to work that might stay quiet in the near future. If employees opt-in, they can be discoverable in the building using Wi-Fi triangulation. Also included are other non-IoT services, including social and campus related capabilities.

The team is also experimenting with automated triggers, notifications, and personalization, as ultimately the vision includes having technology automatically work on behalf of the user. The solution improves employee productivity by saving people time, and represents the foundation to build even more IoT scenarios upon in the future.

Figure 3. A modern app was created for Windows 10. It works on a Surface Hub in a centrally located space as well as on mobile devices.

Facilities insights

The same data can be used to provide facilities insights, such as space utilization as well as temperature, noise, and lighting patterns over time. The data can be presented as screenshots or animated heatmaps using PowerView for Microsoft Excel. Insights such as which areas of the building are being used, temperature conditions and stability, quiet and noisy zones, as well as information to reduce energy costs are all possibilities using IoT sensors. This data can then be used to make changes to the physical environment ranging from layout and furniture to HVAC improvements. A key area of interest to many organizations is space utilization, because finding ways to increase it can reduce the amount of office space needed, resulting in potentially significant cost savings.

Figure 4. IoT data presented using PowerView for Microsoft Excel

Conclusion

Microsoft IT is at the beginning of its journey into enterprise IoT. The future will be full of smart buildings that can automatically respond to your needs and preferences, guide you to your destination, and make tasks—such as starting a meeting—as easy as pressing a button or even just walking into a room. The Internet of Things is just one of the many reasons we are excited about the future.

"The Internet of Things provides tremendous opportunities to make the world around us smarter—and is key to enabling a future where technology truly works on our behalf, instead of people having to work to utilize technology."

Pankaj Arora
Leader, Modern IT Innovation Group
Microsoft Corporation

Additional resources

Microsoft Azure Cloud Resource Center offers white papers and webinars to learn the basics

Azure IOT Hub where you can connect, monitor and control millions of IoT assets

Windows 10 IoT Core support for Raspberry Pi 3

Microsoft IT Internet of Things pilot in action video

For more information

Microsoft IT Showcase

microsoft.com/ITShowcase

© 2016 Microsoft Corporation. All rights reserved. Microsoft and Windows are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. The names of actual companies and products mentioned herein may be the trademarks of their respective owners. This document is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.