Latest News

Will latency drive more business adoption of real-time and on-demand services?

Written by Brian Doricko, Vice President Of Strategic Sales at CyrusOne

Advances in computing power have accelerated the pace of digital transformation at the enterprise level. One of the key areas of innovation is the increasing growth of on-demand services and real-time applications, particularly with media streaming services and AI applications. The question is, are there other definite beneficiaries of this lower latency in today’s network systems?

The massive increase in data farms, data science, compute and storage interaction has driven increasing in technology to bring more services to businesses, individuals and governments worldwide. The largest advantages in our world, that benefits wide areas of populations formerly without service, come or more usable applications, served locally and with very efficient economics.

Most infrastructure deployment in the last few years have had to first consider, what is the application going to be delivering? From there, optimal decisions can be made. Applications used to execute time-sensitive services are particularly reliant on low latency because they have a short window of time to leverage usable data and insight. For instance, for applications such as high-frequency financial trading and digital advertising (such as ad services on Facebook and Google), the speed of delivery is critical to their performance. But this is currently not the case for every business.

For the high frequency trading applications and ad serving, these are full on Olympic style sprints, fastest wins. If you miss on a trade and your competitors executes first, you’ve lost that revenue opportunity forever.  These are perishable opportunities and really do benefit by as fast a system and lowest latency possible.

But consider this: would it matter to users whether or not their weekly grocery sites load up a fraction of a millisecond faster? The answer is generally no because the change would be almost imperceptible and would not impact the end-user experience. Another example to consider is whether a work from home employee benefit from a network upgrade to her home, if she is working on email most of the time, occasionally transferring a spreadsheet to a co-worker? Probably not.

All network performance is measured per second, while latency is measured in milliseconds, and indicates the quality of your connection within your network. So, money can be unnecessarily spent trying to improve performance to no avail, because network performance is not determined by speed but in capacity.

Although low latency technology providers are under increasing pressure to create more competitive technologies at lower costs, that doesn’t mean demand will follow in kind. Cost-conscious organisations will always need to weigh up the cost of replacing legacy network infrastructure against the benefits. This cost restraint will limit the general adoption of operating a low latency network system across many small and medium-sized businesses.

A 2021 report found that although investment in edge computing solutions – the most exciting applicable solution to improving latency – has gone up in terms of priority, latency is still outranked by bandwidth and infrastructure cost, and data security’s impact on operations and apps. This demonstrates that network speed is not a leading motivator for investing in the edge. We know that the largest areas for improvement at the enterprise level is as a result of more powerful computers and increasing intelligent artificial intelligence, machine learning and data science-enabled applications to make the most of available data to build usable insights.

Is low latency important?

The answer is definitely yes. Latency is always important at the level that the output is not improved by any further increase in latency performance.

However, the paradox with edge solutions is that data scattered across the network is often difficult to gather and analyse as a whole. That is why many large scale organisations adopting edge solutions rely on a central data centre hub to hold all the data from their network for analysis and insights.

Real-time data allows companies to assess an asset throughout its lifecycle and environmental monitoring at a granular level, allowing companies to act immediately when a problem arises. Instant data on temperature, pressure and humidity will reduce potential environmental risks and allow smart decisions to be made and implemented quickly.

IoT growth will continue to push data centres into the edge space, however, not at the expense of hyperscale or colocation data centre facilities. Exciting new, emerging technologies with the capability to change our world for the better, such as AI, automation, and augmented reality, to name but a few, need low latency networks to realise their potential but this does not apply to the majority of applications.

So, what does all this mean?  Perhaps it means, the real killer application is not in fact an application at all.  It’s the zettabytes of data we are creating. This data needs to be distilled to make more informed decisions, improve people’s lives, lower cost of services, make businesses more competitive.