IoT Blog
IoT Blog

Why IoT Needs Distributed Data Orchestration

by Olivier Pauzet, Vice President & General Manager, IoT Solutions
The Internet of Things (IoT) has come a long way since the 1970s, when graduate students at Carnegie Mellon built what is commonly referenced as the first IoT application. The students installed a computer board in their office’s soda vending machine that connected to ARPANET—an early packet-switching network that was a foundation for the Internet—creating an IoT application they could use to learn if cold Cokes were available in the machine. Today, IoT does a whole lot more than ensure a cold beverage is available before we walk to the vending machine. With Smart City, Internet of Life Saving Things (IoLST), Industrial IoT and other IoT applications helping utilities increase their use of renewable energy, first responders better protect people and property, retailers provide more reliable service to their customers, shippers track assets as they move through around the world, cities make our streets safer and more. In fact, IoT is a strategic focus at the highest levels of enterprises and original equipment manufacturers (OEMs). 

Two key innovations are driving a new age for the IoT: the “Deep Edge” and Low Power Wide Area (LPWA) cellular standards. We now have the ability to integrate ubiquitous and secure wireless connectivity in very small, low-power, and inexpensive computing endpoints that are deeply embedded into the assets they are monitoring. ABI Research estimates that LPWA network connections, which enable these deep edge IoT applications, will grow from 27 million in 2016 to more than three billion by 2025. The intelligence in IoT applications can now be distributed between the cloud, the network edge and devices at the deep edge. It’s only by providing computing at the deep edge that we unlock the real potential to connect billions of things. As IoT edge device hardware continues to mature, LPWA network coverage expands globally, and new cloud, Artificial Intelligence (AI) and other technologies are available to analyze IoT data for insights, enterprises can benefit from more efficient and profitable business models and entirely new ideas. 

The IoT Complexity Challenge

There seem to be few barriers remaining that can prevent the IoT from fully realizing its original promise of transforming our world with a fully connected economy. However, those on the IoT front lines—IoT developers, enterprise digital transformation leaders, and others intimately involved in building and deploying IoT applications—know there is one significant hurdle. This hurdle is the massive complexity involved in creating, managing and updating IoT applications. These applications are incredibly complex because they require IoT infrastructure and middleware to collect data from diverse endpoints and assets, transmit this data through various networks, process this data in the cloud and then integrate it into ERP, CRM or other cloud-based systems – and then turn around and send back data to these endpoints.

Building complex infrastructure that can securely and reliability collect asset data at scale requires extensive knowledge of a wide range of technologies. Looking just at the IoT technology skills that are required, enterprises have to have deep levels of expertise in IoT hardware, embedded software, wireless connectivity, back-end software, IoT protocols and cybersecurity. So, while business cases for IoT applications are compelling, and the technology exists to create these applications, bringing technical know-how and the experience needed to actually create and launch the infrastructure needed to support these applications is hard. Really hard. This complexity is one reason why nearly three quarters of IoT initiatives fail.

Efficiently Managing Massive Amounts of Data 

As enterprises increasingly recognize the complexity involved in building the end-to-end infrastructure required for IoT applications, they are turning to partners who say they have the skills to deliver IoT platforms with the edge connectivity they need for their applications. Yet, these platforms usually still fall short in truly addressing this complexity. Some of them – often referred to as IoT cloud platforms -- simply provide generic data ingestion services that integrate IoT data in the cloud, but do not help enterprises design, manage or secure the end-to-end integration between the cloud platform, cellular networks and third-party endpoint devices. 

Other so-called platforms are off-the-shelf solutions built for specific applications or use cases that can do one thing well across both edge and cloud. However, these “stovepipe” vertical platforms lock enterprises into a single vendor and use case, and they are not able to support anything beyond basic IoT applications, forcing enterprises to manage multiple incompatible stovepipe solutions that can’t communicate with each other or be updated or improved over time.

These IoT cloud and stovepipe vertical platforms still do not address all the complexity enterprises face as they try to build IoT applications. In particular, they do not help enterprises efficiently manage and glean insights from the massive tsunami of data generated by thousands, hundreds of thousands or even millions of IoT assets. After all, IoT is about data and the insights gleaned from it. If you want to get meaningful insights, and do it efficiently, in a way will scale for real-world deployments, you will need to think very carefully about how you are handling data throughout the stack.

Nor do these platforms address a core issue: how to deliver data without exhausting the limited battery power of deep edge IoT devices. New LPWA technologies are much more energy efficient than legacy connectivity technologies, but even LPWA IoT devices consume power when transmitting data. If data is not intelligently processed and transmitted, so that they only send data that needs to be sent when it needs to be sent, the devices will use up all their battery power long before their expected lifetimes of 10 years or more. Yet, knowing how to intelligently collect data from these devices can often only be determined after they are deployed. Moreover, what data is needed, and when, will also change over the life of many IoT applications. If enterprises do not have the ability to control and update their IoT devices’ data management and transmission rules, they are likely to find these devices running out of power sooner than expected, or unable to meet new business needs or requirements.

Distributed Data Orchestration Enables a Simpler, Flexible and Dynamic IoT

Complexity hindering the growth of IoT can be addressed with a new capability: distributed data orchestration. It provides IoT application developers with control over the entire data layer stack, enabling them to orchestrate the flow of IoT data from the cloud to the edge and back, so they can make sure the right data is sent at the right time, with the right priority, to the right system of record. 

Data orchestration brings the IoT in alignment with enterprise software environments. Rather than requiring entirely different tools and skill sets, IoT applications can now be built, deployed and refined like any other type of business application. By simplifying the IoT, distributed data orchestration makes asset data much more accessible and actionable, so enterprises can use this data to generate insights quickly and easily. 

Sierra Wireless is bringing distributed data orchestration to market with its new Octave offering. This device-to-cloud (D2C) IoT solution provides all of the components of distributed IoT connectivity—LPWA edge processing, cloud processing, end-to-end security and wireless connectivity anywhere in the world—as a single, integrated offer, enabling enterprises to break through the barriers that have caused other IoT initiatives to fail. It allows enterprises to change edge device rules and configurations as easily and frequently as they would in the cloud, optimize back-end code after an IoT application is deployed, and update or enhance the application as business needs change. 

In many ways, Octave with distributed data orchestration has the potential to do for the IoT what Amazon and its cloud services did for computing. Amazon abstracted away the complexity involved in connecting, configuring, scaling, securing, updating and managing computer infrastructure, and empowered enterprises to focus on developing applications, collecting application data and analyzing this data for actionable insights. So will Octave abstract away the complexity involved in connecting, configuring, scaling, updating, securing and managing IoT infrastructure, and empower enterprises to develop IoT applications, collect IoT application data and analyze this data for actionable insights. 

In future blogs about Octave, we will look “under the hood” examining how it enable enterprises to extract, orchestrate, act on and secure IoT asset data, so that they can develop IoT applications faster, enhance these applications over time, and better use IoT data to generate insights that improve business outcomes. 

Start with Sierra for the device-to-cloud solutions you need to develop and deploy IoT applications, and to learn more about distributed data orchestration.