Servers: The Right Compute
cancel
Showing results for 
Search instead for 
Did you mean: 

The Race to the Internet of Things - Computing at the Edge

RubyNicholson

Kelly Pracht.pngGuest blog by Kelly Pracht, Sr. Manager, HPE Edgeline IoT Systems, Product Management

When Sam Bird straps into the cockpit of the DS Virgin Racing Team Formula E electric car, it’s an exercise in big data analytics. Hundreds of sensors and cameras capture video and structured data to monitor battery temperature and energy consumption, transmission performance, brake wear and temperature, chassis vibration, location, and a host of other variables. It’s the Internet of Things (IoT) on wheels, moving at up to 140 mph.

Racing electric-powered cars in the Formula E series is a balancing act, requiring teams to manage power consumption vs. battery temperature and make continual adjustments. Sensor data is analyzed real-time, on location. The DS Virgin Racing Team incorporates the computing power that you would usually associate with a data center directly on pit row. Streaming data is analyzed at the point of collection, providing real-time insight that allows Bird and his mechanics to make real-time adjustments to maximize the systems that control their car, and hopefully win the race. After the race, aggregate data is analyzed for deeper insights.

The idea of trackside analytics might sound intuitive to a racing organization, but the concept of pushing Big Data analytic computing capability out to the edge of the network – performing analytics as close as possible to the data source – should have a much broader application in the enterprise.

There can be many good reasons for not moving data back to a massive, centralized data repository and processing engine (commonly referred to as a “data lake”) before engaging in analytics, but for the moment let’s focus on two:

1. Accelerate time-to-insight

Think about your own business. You collect data out at the edge of your network, through connected devices, and you want to get insight from that data quickly. Moving the data back to a central data lake takes time; it adds latency. Why is that an issue? Well, how fast do you want to know if your connected asset is about to fail? Probably faster than the time that it takes to transfer the data and analyze it in a central data center. What if your measurements indicate that a device is about to cause unplanned downtime for a customer? Again, you need that insight as soon as you can get it. Consider also that network connections are spotty or not always “up” – it changes the time you have to act on your data to indefinite. Moving compute to the edge on the actionable data can improve reliability as well as reduce latency.

2. Conserve bandwidth

It costs money to move data, and much of what is collected could be processed from the datacenter in the data lake in aggregate form rather than as raw data, reducing the load on the network and on central resources. Even if you can afford the bandwidth, do you really want to clog your infrastructure with raw analog data from your connected devices?

The Internet of Things (IoT), by its nature, presents a Big Data problem – terabytes of data streaming in real-time and often demanding instantaneous analysis and response. A question organizations need to answer is: “Where does it make the most sense for Big Data analytic horsepower to reside?”

In an Internet of Things environment, computation needs to get done as quickly as the associated system of things requires. Data processing near the source, at “the edge of the network,” can minimize time-to-business-insight and value. A racing team can optimize power consumption on the fly, promoting battery lifetime and perhaps preventing thermal issues mid-race that can degrade battery performance, resulting in a safety risk. In contrast, collecting and analyzing “centralized” data is necessary when integrating multiple devices and data sources. For example, a car manufacturer might aggregate data from sensors on thousands of cars to understand how to build a better transmission.

How to choose the right platform Pracht blog 2-11.jpg

To optimize IoT performance and cost, companies must examine their infrastructure – including connectivity, security, data analytics, services, IoT ecosystem, and compute – and define a strategy that balances data processing at the edge and at the data center. Read the Enterprise Strategy Group white paper on how to choose an IT platform to empower your Internet of Things.

Computing at the edge can provide time-sensitive insights, allowing you to provide better service to your customers. For your business, it can be the difference that gets you to the coveted checkered flag before the competition.

Visit our Internet of Things solutions page to learn more about how HPE can help your business create value from the IoT. And I hope you can join us at Mobile World Congress 2016, coming up 22-25 February at the Fira Gran Via, Barcelona. You won’t want to miss our Internet of Things demo there – we’ll be in Hall 3, Booth A20.

  • Trending Now
0 Kudos
About the Author

RubyNicholson

I am a Senior Manager managing external content and social media for HPE Servers Awareness. Stay tuned for topics on Mission Critical Solutions, Core Enterprise and SMB Solutions, Next Gen Workload Solutions, Big Data and HPC, Cloudline and HPS Options! Follow me @RubyD_Nich

Events
See posts for dates
ONline
Connect Worldwide Events - 2017
Connect Worldwide is Hewlett Packard Enterprise’s largest independent technology user community, and has more than 70,000 global members. Visit this p...
Read more
Each Month in 2017
Online
Software Expert Days - 2017
Join us online to talk directly with our Software experts during online Expert Days. Find information here about past, current, and upcoming Expert Da...
Read more
View all