Blogs

Extracting insights with Internet of Things data analytics platforms

Cloud Architect, IBM

A variety of value-adding, insight-generating and decision-enabling data emanating from a variety of geographically distributed sources—which, incidentally, are increasingly common—are in need of comprehensive, yet cognitive, analytics. Thanks to a sharp increase in the number of data sources, and also as a result of a wide array of transitions and disruptions in the IT domain, we are seeing massive volumes of data being carefully collected, cleansed and stocked in large-scale storage appliances and networks. Such network infrastructures, in synchronization with wide area networking optimization technologies, are designed to efficiently transfer data originating from multiple locations.

Other associated aspects, such as data formats (representation, persistence and exchange), data transmission protocols, data virtualization and data visualization platforms, are systematically accelerating the process by which we extract sense from data heaps. Moreover, standards-compliant data analytics platforms are contributing immensely by helping automate the transformation of data into information—and, indeed, into knowledge—for dissemination to people and machines, thereby enhancing decisions.

The strategic significance of data analytics

However, the extremely large data volume involved poses the real challenge—and data variety and velocity also inhibit knowledge discovery and dissemination. The open-source community and product vendors across the globe have long pondered the strategic significance of methodical data analytics for sustainably enhancing business efficiency and value. As might be expected, they offer some solid advancements in their analytical solutions and services that intelligently streamline next-generation data analytics to arrive at diagnostic, predictive, prognostic and prescriptive insights.

Data can be described as big data, fast data, streaming data and Internet of Things (IoT) data. Data processing can be batch, interactive, iterative or real-time. And big, real-time, streaming and IoT data analytics platforms enable the extraction of insights from all kinds of data. What’s more, IT infrastructures are optimized (using cloud technologies and any of many available automation tools) to be lean, green, dynamic, elastic and workload-aware. In short, software-defined cloud centers are being proclaimed as the one-stop IT solution for activating and accomplishing sophisticated—yet affordable—analytics. But here I’d like to highlight the need for end-to-end, converged and cloud-based IoT data analytics platforms that can be used to build and sustain pioneering and path-breaking applications and services tuned for people empowerment.http://www.ibmbigdatahub.com/sites/default/files/blog_analyticsplatform_blog.jpg

Making sense of IoT data

IoT data analytics helps make sense out of IoT data, empowering all sorts of connected, embedded (physical, mechanical, electrical and electronic) devices, appliances, instruments, equipment, wares, utensils and machines to be intelligent in their operations and outputs. But IoT data analytics also provides great benefits to users and knowledge workers in their day-to-day activities.

Despite the availability of IoT application enablement platforms (AEPs) aplenty, building and sustaining highly competent IoT data analytics platforms is a must if we are to leverage IoT data efficiently and effectively. The whole purpose of such platforms is to connect, extract, pre-process, analyze and mine data from a variety of sources, including sensors; actuators; transactional and operational systems; technical, scientific, engineered and social systems; and the like.

Capabilities of next-generation IoT data analytics platforms

Both operational and IT systems are transitioning from systems of records to systems of engagements, not least because different and distributed systems are engaging in purpose-specific collaboration, correlation and corroboration while staying loose and lightly coupled. Accordingly, IoT data analytics platforms must integrate seamlessly with different reporting tools, dashboards, consoles, portals and other visualization platforms to disseminate the resulting information in preferred formats.

Thus an IoT data analytics platform that synchronizes intelligently with IoT AEPs and IoT gateways must be able to help do the following:

  • Uncover timely and actionable insights.
  • Enable smart objects, devices, networks and environments.
  • Lead to pioneering and people-centric applications and services.
  • Produce precise predictions and prescriptions.
  • Facilitate process excellence.
  • Guarantee preventive infrastructure maintenance.
  • Ensure optimized use of distributed assets through monitoring, measurement and management.
  • Safeguard and secure people and properties.
  • Monitor complex environments to ensure business performance, productivity and resilience.

IoT data analytics capability can significantly enhance people’s choices, care, collaboration, convenience and comfort. This being so, the ready availability of versatile technologies accelerates and augments the arduous journey from data to information to knowledge, bringing about a smart planet. Indeed, the timely and actionable insights that result go a long way toward producing both smart machines and smart people.

Essential modules of IoT data analytics platforms

Considering the growing complexity of next-generation analytics, the quality and quantity of contributing modules for any IoT data analytics platform should be consistently on the climb. Here’s a brief description of essential modules and their unique features:

  • Data virtualization, integration and ingestion modules: Millions of connected devices and sensors are pumping out data about themselves and their environments—data just waiting to be captured, transformed and ingested into IoT databases.
  • Data analytics platforms: Big, real-time, streaming and IoT data analytics platforms are widely available from both the open-source software community and commercial-grade product vendors.
  • Data visualization platforms: Dashboards, portals, consoles, reports, maps, graphs, charts and the like play a primary role in visualizing obtained knowledge. Accordingly, seamless integration with visualization tools to display knowledge in user-preferred formats is a must for IoT analytics platforms.
  • Application enablement platforms: Typically AEPs, which come with a variety of libraries, connectors, drivers, adapters and the like, can offer efficient applications for all kinds of industry verticals through configuring rather than coding.

Switching to non-functional aspects, IoT data analytics platforms must also be, among many other things, high-performing, highly scalable, available, secure, easily configurable and customizable.

Reference architecture for IoT data analytics platforms

A look at the macro-level architecture of IoT data analytics platforms vividly illustrates how the principal components help identify hidden patterns, beneficial associations and fresh opportunities. In doing so, IoT data analytics platforms can drive business efficiency, create business value, shrink total cost of ownership, grow return on investment, shorten time to market and offer real-time alerts—and much, much more.

 

A typical use case—connected cars

A specialized IoT data analytics platform can bring much-needed transformation to automotive enterprises through next-generation capabilities. These capabilities may offer such things as real-time data about vehicles and their parts or pragmatic insights for car users and producers—all arrived at through advanced diagnostic, prescriptive and prognostics analytics. Moreover, automotive service providers are increasingly collaborating with analytics service providers to help automobile manufacturers gain valuable insights by systematically analyzing real-time data being streamed from connected cars. The information gained makes possible a bevy of insights-driven capabilities.

Big and real-time data analytics platforms can process sensor data from connected cars to aid knowledge discovery. The insights so garnered include real-time alerts about driver behavior, traffic levels, road conditions and maintenance state. Any end-to-end platform ensures a single—yet a comprehensive—view of vehicle data and any associated insights, informing automotive engineers about everything from safety records and performance to predictive maintenance. Actionable insights can help bring about smart parking assistance and traffic management systems, self-driving cars and so much more. And not only car drivers and owners, but also mechanics, manufacturers and passengers stand to gain immense benefits from IoT data analytics solutions and services.

Because the amount of IoT data generated, collected and subjected to deep analysis is already measured in exabytes, with no end in sight, IoT technologies and strategies must be able to deliver actionable intelligence from IoT data. So it should come as no surprise that software-defined clouds, which are emerging as a one-stop IT infrastructure solution—along with cloud-hosted, end-to-end and converged analytical solutions and services—are increasingly seen as mandatory in the knowledge era that has ensued.