Sign up for our newsletter
Technology / AI and automation

Our future is at the edge

Take smart cars as an example. They produce gigabytes of data on everything from passenger comfort, to music preferences & safety records, let alone all the GPS requirements and vital systems information to make them fully autonomous. Scale that up to millions of cars and the potential multi zettabytes of data being produced is too much for our current infrastructure to handle. Factor in the ever-growing Internet of Things (IoT) and the 5G revolution and the technologies of tomorrow quickly become today’s serious problems.

The datacentre forms the fundamental backbone of our technological success and advancement. Every key stroke, swipe or even step or snore that we take is increasingly recorded and tracked into a data centre. Whether it’s primary centres for the big technology companies or colocation sites for smaller businesses, everything we do feeds back to a data centre. Many businesses rely on having easy, instant access to their data centres but as we look to the increasing data that is being produced, it will no longer be possible to have fingertip access to these facilities.

There are only so many data centres that can be developed in Tier 1 locations to deliver the lowest latencies going. Cities like London, New York and Frankfurt have to accommodate their vital economic infrastructures with the necessary technology infrastructures to support them. In cities around the world, space is at a premium and less and less of this space is being made available to data centres that are equipped to handle the needs of tomorrow.

With that in mind, Tier 1 sites are not fit for purpose when we look at how they can handle the zettabytes of data produced by the growth in connected devices. So how are data centre providers looking to make good on their ability to power us into the technological revolution?

White papers from our partners

The edge

Increasingly, edge data centres are taking prominence in the way they can provide support and infrastructure for even the most basic of businesses requirements. There is no need to have back office functions for high end trading houses located in the same facility as the trading computers. The latency required to power financial trading is not the same latency required to access personnel records or historical data, so this is where edge data centres can support.

As newer, more data hungry technologies get powered up and adopted by the mainstream populations, Tier 1 data centres will not be able to cope with the amount of data being fed into them from sources, hundreds or even thousands of miles away. So, what’s the solution? Smaller, almost modular data centres that are easily able to be set up in smaller towns, or even remote locations, that can capture the data from passing smart connected devices, whether that be cars, watches, medical devices etc. Being able to deliver distributed compute, power and storage capabilities, not only to hotspots but to local areas as well, will set to be the new future for the data centre. Think of it as a local petrol station. One for every town but some are bigger and more well equipped than others, depending on the traffic that it sees.

But once we have our distributed, modular data centres equipped and running on the edge, how do we have enough compute, power and storage to process the requirements of passing connected devices? This is where the software in the data centre has to change and how AI is best placed to do that.

 

AI at the edge

The data produced by today’s technology centric businesses and tomorrow’s technology-dependent infrastructure needs to be accurately and securely processed and filtered in a smart way. The way we look at data, particularly at the edge, will change in how we manage it, process it and prioritise it. New data management frameworks are essential. By deploying artificial intelligence into data centres, both Tier 1 sites and those at the edge, we are equipping the facilities with the ability to intelligently understand the data that it is processing. The deployment of AI within the data centre also helps to reduce the need for more sophisticated hardware to be capable of processing all the data capture. AI vs Machine Learning

Typically, data centres that have vast amounts of compute, power and cooling capabilities are high-performance computing (HPC) centres that look after government and research computing projects. These sites are usually very expensive, solely owned by one corporation or government and focus on number crunching for one application at a time. With the use of AI in the data centre however, the compute power could be significantly less as it will have already filtered out the non-essential information before it reaches the facility. With the compute requirements lowered, so are the power and cooling requirements, helping to make the facility much more efficient and effective.

 

Are we future ready?

As the edge infrastructure becomes increasingly more valuable and secondary facilities take a great prominence in data processing, the need for greater fibreoptic communications networks between organisations and Tier 1 and Tier 2 or 3 facilities will have to become stronger. This underlying communications infrastructure is vital for the success of edge computing facilities and indeed the growth of modular colocation facilities.

Artificial intelligence will play an increasing role within the sites and communication lines between, helping to filter out the information that isn’t necessary to secondary or even tertiary sites. Future technologies need to move hand in hand with the infrastructure developments in order to be able to support these new developments in the connected world. The integration of these technologies and infrastructure is a key component to a successful AI revolution.  Greater exploration in ways to enhance man-machine collaboration and related trust levels is needed for AI to be a successful partner in business development.

Businesses behind the technology revolution need to be prepared to invest their capital in artificial intelligence in order to allow them to sort through the reams of data that their products produce and store. To become a more efficient technology company and maintain a competitive edge in the market, streamlining data processes and looking to the edge for support are the best ways to plan for the future.
This article is from the CBROnline archive: some formatting and images may not be present.