Data centre (DC) providers looking to capitalise on the demand for high-performance compute (HPC) services, which are expected to grow over the next five years, must not underestimate the size of the financial outlay needed to support the technology, warns Greg McCulloch, CEO of Aegis Data.
Historically, end users of HPC services have been researchers, engineers, scientists, educational institutes and healthcare professionals, but the growth of data demanding technologies, including cloud, big data, and IoT now means that HPC is becoming more readily embraced across the wider business community. According to recent research the HPC market is expected to reach $33bn in value by 2022, growing at a compound annual growth rate (CAGR) of five per cent from 2016.
While many are keen to leverage the benefits of HPC, the reality is that many DC’s don’t possess the infrastructure needed to support the demands of the technology. In line with further research from Gartner predicting increased spend in DC services over the next 12 months, McCulloch suggests we could see a greater proportion of DC providers wanting to allocate more of their budget towards investment in HPC services:
“Traditionally, HPC services are seen as a niche technology reserved for academics or government entities, but as global data use has continued to rise so has the demand for HPC capabilities needed to process greater volumes of information. This is something we are very much seeing first hand – two years ago, approximately one in ten customers would come to us asking if our facility was HPC ready, today, that figure is closer to seven in ten. Our experience is not isolated and as we have seen from wider industry data the demand for HPC services is increasing rapidly.”
McCulloch goes on to warn that while many providers might be keen to invest in HPC services, they must not underestimate the size of the financial outlay needed to do this:
“While we’ve seen evidence that DC spend is expected to increase over the next 12 months, but the challenge for a lot of DC providers wanting to invest in HPC technology will be the size of that cost.
“Any investment into HPC is likely to represent a significant CAPEX expense and is typically hindered by space and infrastructure boundaries. For a lot of firms in this position often the more viable option is to build an entirely new site, but again this can only be achieved through significant investment and therefore we are seeing a lot of providers reaching a crossroads as to where to take their operations. Because organisations are now required to carry and process unprecedented data volumes it’s important now for DC providers to understand where the current and future demand for HPC capabilities is likely to stem from.
“Industry research combined with the continuing growth of technologies including cloud, IoT and big data reinforces that demand remains on an upward curve. Because of this any investment into making your facility HPC ready will certainly allow you to meet any future demand, but this can’t be at the expense of your existing customers. It is therefore imperative that any investment is still able to account for traditional mission-critical applications,” McCulloch concluded.