In-Memory Computing and the 5G Ecosystem

In-memory computing can provide tremendous benefits for the 5G ecosystem. We’ve seen the marketing for the new fifth-generation mobile networks. The benefits of 5G for end-users are easy to understand. Speeds faster than your home broadband and latencies only a little slower promise to be game-changers for consumers, enhancing existing applications and opening open entirely new categories that we can’t currently imagine.

This shift in technology also represents an opportunity for network owners and their suppliers. As with any technology leap, there are also challenges. The dramatic increase in speed is matched only by the scale. Studies show that there are over 3.5 billion smartphones in use around the world — and that number is rising — meaning that even a relatively modest carrier can have millions of customers. Meeting these demanding latency requirements with millions of simultaneous users isn’t easy.

The Role of In-Memory Computing in the 5G Ecosystem

Where vast volumes of data and low-latency requirements collide, you’ll find in-memory technology like GridGain, the enterprise-grade version of Apache Ignite. In-memory technology is how leading 5G providers are managing this explosive combination of speed and scale.

GridGain manages to solve this problem in two ways. First, it runs on a cluster of commodity hardware, which means that it’s horizontally scalable. If you need more storage capacity or compute power, you can add new nodes. Secondly, it typically holds all the data in memory. If you need to hit millisecond-level SLAs, it’s the only reliable way. But storing data in RAM isn’t a panacea on its own. If the data is spread across random nodes in your cluster, hitting the network is almost as bad as having to access a disk. Instead, in-memory technology understands the concept of data affinity, or how the data is related, and ensures that a single machine in the cluster is used for performance-sensitive calculations.

Addressing the Challenges of 5G with In-Memory Technology

Let’s look at a couple of real-world examples of how people are addressing the challenges of 5G using in-memory technology.

Moving from 4G (LTE) to 5G means dropping the latency from around 50ms to about 10ms. But the requirements for processing requests have not changed substantially. When a phone requests some data, is it authorised? What quality of service is warranted? Has the user run out of data for the month? All these questions and more need to be answered in a split second. A large American equipment manufacturer is using GridGain for precisely that. They chose our technology not only for its performance and scalability but because it is reliable and easy to deploy.

A common use-case for in-memory technology is real-time analytics, and 5G is no exception. We have an Asian mobile operator with over thirty million subscribers using GridGain to provide users with services and alerts. An SMS would be helpful if you’re about to go over your data allowance, for example. Or maybe the provider would like to promote an offer. In the UK, the Communications Act 2003 allows consumers to set limits on the cost of their bills. In general, these rules need to run in near-real time across all customers accessing the network while also simultaneously combining data in a digital integration hub which can combine data from disparate internal systems such as billing and marketing and make the data available to multiple business applications.

While we don’t yet know all the new and unique applications that 5G will enable, we do know that legacy technology will struggle to meet the speed and scale demands its implementation brings. We’re already working with many vendors in this space who believe that in-memory software is an essential piece of solving that 5G puzzle and would love to hear from others doing the same.

Share This