How Pairing In-Memory Streaming with IoT Will Fundamentally Change Your Business

Knock-knock

Your business is already under assault, whether you know it or not. It doesn’t matter whether you’re B2B or B2C; the same technology variables are driving pressure on your IT infrastructure. The growth of the global technology ecosystem has always been organic. While new technologies often layer on top of previous enablers (e.g. mobility over the top of the internet, social media on top of both), the introduction of something new is rarely coordinated or systematic, which means sectors are always trying to catch up with or leap-frog each other. The bigger the event, the more pervasive this effect becomes and nothing so far has reached the potential scale of the Internet of Things (IoT). While mobility added billions of devices, IoT is adding trillions of input points, most of which (in addition to sheer numbers) also stream data nonstop. This is not only a whole new level of technology; its size is expected to be an order of magnitude higher than anything that preceded it. So how can a business prepare for this brave new world, and what is going to be the core enabling technology?

Who’s there?

When you’re using terms like trillions and nonstop, the baseline requirements are going to be speed and scalability. What kind of speed? Milliseconds? No, that’s your dad’s speed. We’re talking microseconds, and not very many of them at that. These microsecond frameworks also need to scale to millions of simultaneous transactions and all executed essentially in real-time. And to make it extra fun, the transactions are likely to be complex. Most consumer-facing businesses are already dealing with the surface effects of being accessed by millions of mobile devices at once (Facebook, for example, has already had multiple instances of over a billion people being on their system at the same time). A B2B company will have the same challenge – sensors and IoT devices are everywhere because they’re inexpensive, add a lot of value and drive a substantial inbound volume of information. There’s also the issue of inter-systems communications (referred to as messaging systems), which is software and hardware systems and subsystems batting information at each other at incredible speeds, while billions of consumers or IoT devices dogpile on top.

In-Memory Streaming

From an application infrastructure perspective, the need for speed is being addressed by 1) significantly faster and lower cost chips, which means 2) cost of RAM is also less expensive and can scale much higher (terabytes of RAM is now common), which means 3) applications that used to have to traverse a network to access information in a database can now store relevant information in RAM because it’s now fast enough and big enough. Once you eliminate the latency associated with traversing a network and accessing a database, your latency drops by order of magnitude; events that were previously transacted in seconds now execute in thousandths of a second. The combination of in-memory computing coupled with a high-powered streaming engine (which together form an in-memory computing platform) provides the speed and scalability that will not only help you keep up with ecosystem pressure, but can also keep you comfortably ahead of the inbound streaming data curve.

In-memory streaming who?

Oil and Gas: Energy companies have riddled their extraction and refining infrastructure with IoT devices, which means not only far more inbound data, but data processing occurring at the edge of the network (referred to as Edge Processing). Since in-memory technologies are both lightweight and cloud-enabled, they can act as remote sensors.

Healthcare Informatics: A company that manufactures mattresses has (same as Oil and Gas), riddled their product with IoT devices that can report changes in sleep conditions for patients. The sensors can track events like sleep apnea and will soon be able to detect respiratory or cardiac events. This company is already streaming billions of input points per night and is still warming up.

Media: Your set-top box and the device you use to control it are both considered IoT devices by your service provider. They use them as a streaming source to track your viewing habits, purchase history, customer service interactions, billing, etc. This is used to develop a 360° view of how you interact with their IT resources, which then powers an AI application to handle service calls optimized to your particular needs.

In-memory streaming, your new best friend

The core enablement of in-memory technology has been the in-memory data grid (IMDG). Significantly different than in-memory databases, IMDGs are ridiculously fast, easy to deploy and designed to work with your existing infrastructure. This technology has been around for nearly ten years and evolving with the addition of fully integrated stream processing engines built from the ground up and optimized for today’s high-speed digital world. When combined with 3rd generation stream processing engines such as Hazelcast Jet, the potential scope of innovation increases by an order of magnitude. To find further information on IMDGs and stream processing engines, please click here.