This is a part 2 of a two part series. Read more about the future of streaming analytics in part 1.
While the data lakes landscape has matured over the last few years, the streaming analytics market is continuing to develop as more enterprises embrace the use of streaming analytics to provide real-time insights for their business systems. Early returns suggest that embracing streaming analytics in industry can lead to significant efficiency gains and decreased operational costs. For example, a study by McKinsey & Co. found that “using real-time data to predict and prevent breakdowns can reduce [manufacturing] downtime by 50 percent.” With real-world cost savings at stake, the momentum toward real-time has never been stronger.
Forrester Research recently published their “The Forrester Wave: Streaming Analytics, Q3 2017” report, which compares streaming analytics vendors and draws conclusions about the state of the real-time data industry. In part 1 of this series, we defined streaming analytics and discussed the benefits of performing streaming analytics at the network edge. In this post, we will discuss how to evaluate streaming analytics solutions.
How to Evaluate Streaming Analytics Solutions
One of the key aspects when evaluating potential streaming analytics solutions is scalability and support for multiple devices or sensors. “Enterprises have dozens, hundreds, and often thousands of applications that employees and customers use,” the author explains. “All these applications generate data that has the potential to be valuable for real-time analysis, especially when combined from multiple sources.” Streaming analytics that cannot scale to support and utilize data from multiple devices are doomed to stall out in the proof-of-concept phase. Forrester concludes that “streaming analytics solutions provide connectors to ingest and sense numerous application and IoT data sources in real time. This is a key value proposition of streaming analytics — to pull all relevant sources together in memory in real time.”
Another key criteria for evaluation is the ability to intelligently filter raw data, based on real-world context, in real-time. “Streaming data moves in real time from multiple data sources, but not all of the data may be relevant,” the report explains. Streaming analytics solutions are the first opportunity for applications to reduce raw data into a more consumable, information-rich format. This functionality is especially relevant for Machine Learning applications, where feeding refined data can exponentially increase the efficiency and training speed of Machine Learning models.
Taking Action on Edge Data
Forrester notes that the “insights generated using streaming solutions are immediate but not valuable unless they are used to take action.” This is, perhaps, the most important aspect of evaluating a streaming analytics solution. Acting on real-time information can be as simple as human intervention, occur via “APIs, or take any other arbitrary action that can be defined in code.” While how this occurs will vary by application, Forrester describes how this might look with a real-world example where a “predictive maintenance application may monitor temperature and vibration data streamed from a conveyer belt. The streaming analytics solution could detect a spike in either temperature or vibration to indicate a looming shutdown.”
Now aware of an impending failure, “the solution could then push an alert to an operator or trigger an automatic shutdown of the machine. In addition, if the cadence of the streaming data is interrupted, that may also indicate a problem with the sensors on the machine.” These types of insights can trigger maintenance alerts to operations professionals or order replacement parts before a machine fails, which can lower maintenance costs and reduce downtime of industrial systems.
Like the Forrester report says, “we live in a real-time world, and so too should enterprise applications.”
This is a part 2 of a 2 part series. Read more about the future of streaming analytics in part 1.
Learn how Swim ESP™ powers Machine Learning and intelligent decision-making at the edge with a streaming analytics engine for industrial and other real-time applications.
Gualtieri, Mike, et al. “The Forrester Wave: streaming analytics, Q3 2017.” Reprints.forrester.com, Forrester Research, Inc., 7 Sept. 2017, reprints.forrester.com/#/assets/2/308/'RES136545'/reports.