Home < Blog < A Reality Check for 5G Real-Time

A Reality Check for 5G Real-Time

4 min read

Real time for realz

While we applaud other vendors reaffirming what we here at Volt Active Data have been professing and delivering on, for several years, simply adding a few buzzwords doesn’t cut it. Without specific details, it merely adds to the confusion in the market and makes the job of a technology decision maker much more difficult.

For Volt Active Data, the most prominent example is the use and misuse of ‘real-time’ in the context of 5G. Volt Active Data has been operating in a single-digit millisecond world for the best part of a decade, even before 5G became a reality. Not surprisingly, vendors around the world are looking at 5G and seeing a bonanza. The web is flooded with articles professing the next break-through in ‘real-time decisioning’ and the like.

Two recent examples, while not alone, highlight the challenge:

In the first example, the author rightly explains the importance of marrying “real-time information with historical information in the cheapest, easiest-to-deploy flexible architecture”, but then goes on to describe how the product is now being used for COVID tracing in Thailand. Reporting on COVID results and aggregating is a worthy activity, but it is not real-time and it is not decision making.

A second (paid) example, which is actually a republish of an earlier article, discusses “…two technological tectonic plates moving inexorably toward one another. The first: the continued improvement and accessibility of public clouds. The second: the impending rollout of 5G.” The author claims that “These two capabilities are about to unlock latency levels that were previously only available to Fortune 500 companies.”

We’re really not sure of the connection here. The article also goes to great lengths to discuss the 100 Millisecond Rule. Again, this is completely missing the point. While it’s correct that 100ms marks a delay that humans begin to notice, traffic growth in the future will come from devices and computers – the ‘things’ in IoT – and not from people.

At Volt Active Data we know this. We have customers already operating in the World of single-digit milliseconds.

The 5G world is about ‘decisions’

To put some ‘reality’ context around this, the first thing these other vendors are missing, is that the 5G world is about ‘decisions’. Decisions are made in Transactions. A Transaction is not storing the result of a COVID test. It’s where you do something that can have multiple outcomes, and until you do it you won’t know which one will happen.

Scalable Transactions is something that legacy RDBMS struggled with, and NoSQL has never really addressed. While you can scale ‘Reads’ by having multiple copies of the data, Transactions and Writes are vastly different.

The real volumes for future transactional workloads will be found in the IIoT space, which is rapidly becoming a point of emphasis for 5G. We know from our extensive experience in Telco that a single phone call or mobile data interaction that involves dozens or even hundreds of transactions behind the scenes, which results in systems that handle more than 500K TPS per second. Given that 5G is assuming a device density of 1 million per square km, we can assume extraordinarily higher volumes will be needed for automated decisions in the future.

A more important issue is that 5G’s latency expectations, force you to deploy closer to the edge, even under ideal circumstances. Meeting these latency expectations involves architectural changes that reduce the number of trips to back end servers. The actual time constraint is not the 100ms that the US Dept of Defense wants for human / computer interactions – it’s the single digit milliseconds mandated in the 5G specs.

Summary

Other vendors are talking about what our customers are already doing, and we get it. This is an incredibly exciting time to be in the fast-data industry. But it’s also a time of change. The future is mass scale automated transactions in single-digit millisecond timescales. We see a whole series of niches emerging that other database vendors will be able to find homes in, but when it comes to mass scale automated millisecond decision-making, we don’t see many genuine competitors – just a growing collection of buzzword-filled marketing.

David Flower