New data sources from edge devices such as security cameras, drones, mobile apps, and the Internet of Things have existing storage networks bursting at the seams. Most, if not all, of the data created from these devices are “streaming.” “Streaming data” refers to a continuous data flow with no clearly defined beginning or end.
In today’s economy, a business’s ability to grow is directly related to its ability to store, manage and utilize data. Those who can harness data at speed and scale will win markets, minds, and more. Information is at the core of the new digital ecosystem and hence an essential enabler for both digital modernization and disruption – making it more important than ever for organizations to be on the forefront of this digital transformation.
The recently published e-book, Modern Enterprise Data Pipelines, discusses the modern streaming storage engine Pravega. Pravega, developed by Dell Technologies, enables endless insights and process optimization and modernization, all with a significant reduction in operational costs. In the e-book, you can read much more about the technical nuances of Pravega, but to truly understand its power, let’s look at a few examples showing how it can be used:
We’ll look at preventative maintenance in a specific industry – roller coasters – but this example can be applied to any industry. Using Pravega to ingest real-time streaming data from thousands of sensors along a roller coaster, the data can be used to identify key points as simple as how many cars are present on the roller coaster at any time, or as complex as how many vibrations per second a sensor experiences while the car passes it at a certain point. If a normal vibration reading is 3,000 vibrations per second, a threshold can be set to alert a maintenance technician when that vibration reading is too high – alerting the technician that a particular ride needs maintenance. The same data can be accessed later using the exact same tools and compared across different rides or over a longer period – to generate trends – which can be helpful in predicting failures or determining the need for maintenance.
While this is an example of a roller coaster, the same basis can be used in general automotive use cases, when thresholds can be set for oil temperatures, tire pressure gauges, and so on, in order to alert users to the need for preventative maintenance before the problem leads to much bigger problems, like a flat tire or poor engine health.
In a manufacturing environment, anomaly detection can be extremely important in saving time, resources, and money. By placing IoT sensors and cameras along the manufacturing line, Pravega can ingest images and data such as belt speed and temperature. Camera images can automatically discover parts, or products, that are out of specifications and then create an alert that something is not right. By utilizing the data from the sensors, the user easily finds out that the ambient temperature of a machine was too high and/or the speed of the belt was incorrect. Instead of just fixing the problem, a machine learning model may be created – teaching those sensors that if the temperate reaches a certain threshold or if the belt reaches a certain speed, to automatically enable a fan or reduce the speed before anomalies in the product are ever created.
Regardless of the product, this example can work in any manufacturing process. Further, anomaly detection can be used in many different industries outside of manufacturing. It can be used in the financial sector, to find anomalies in mobile check deposits, in coffee shops to monitor preventative maintenance on their machines and provide automatic reordering of supplies, or in hydropower facilities, to find shortfalls in portions of a powerplant. The practical applications are unlimited.
Project alignment, object detection, and allocation
In a construction environment, drones have been used to stream real-time video and telemetry into Pravega. By providing a real-time glimpse of the progress of a construction project, analysis can be completed to compare the digital rendering of the project to the actual progress, delivering a progress report that is always up-to-date and can ensure the accuracy of the construction and appropriate time frames for future planning.
At the same time, by attaching sensors to construction equipment on the ground, Pravega can enable object detection. Equipment, people and materials can be tracked for inventory purposes, or for allocation reasons – ensuring that they are making the best use of each portion of their resources across multiple projects.
And while this particular use case is in construction, similarly to all the other examples, it can be used as easily in other industries. For instance, nearly the same use case was completed in mining, including drone feed for progress reporting and sensors for heavy machinery allocation. The drones monitored the rendering down into the ground instead of building above it.
In a “smart kitchen,” Pravega has been used alongside digital thermometers in large storage coolers to ingest streaming temperature readings in real-time and create an alert when the temperature is out of range. This can prevent entire coolers from spoilage if a door is left ajar. It also helps maintain food costs, since keeping the food at an optimal temperature allows for the longest possible shelf life. This is only the beginning of a “smart kitchen” project, which will soon have multiple stages of sensors and alerts to maintain many different aspects of the process.
As you can see, streaming storage engines such as Pravega provide a backbone to ecosystems of data that put processes in place to leverage the value of streaming data. With those processes in place, organizations such as the ones described above are more prepared to derive value from the data that full of intricate complexities but is also naturally ripe with opportunity. Edge computing and the internet of things brings the promise of new possibilities to those organizations brave enough to work towards unlocking them.