Intelligent Data Centres Issue 30 | Page 31

IT ’ S IMPORTANT TO HIGHLIGHT THAT CLOUD AND COLOCATION ARE NO LONGER COMPETING , ESPECIALLY WHERE HYBRID CLOUD NATIVE HPC IS CONCERNED . www . intelligentdatacentres . com
EDITOR ’ S QUESTION
SPENCER LAMB , VP , KAO DATA
or today ’ s businesses , cloud computing offers a

F number of benefits and has , in many respects , become a fundamental necessity . However , in the worlds of High Performance Computing ( HPC ) and Artificial Intelligence ( AI ), the hyperscale cloud no longer competes with colocation . One , in fact , complements the other .

In recent years , HPC has become cloud native , but the intricacies of parallel processing and other forms of intensive computational analysis means such processing has to be undertaken locally . This often requires specialised data centre expertise and a facility which has been designed to support industrial-scale computing . However , it ’ s important to note that the data coming into this process will be collected via the cloud and the findings , thereafter , will be dispersed back via the cloud – creating what is essentially hybrid HPC .
One might even consider that colocation data centres working within HPC and AI are becoming the facilitators between data gathering , processing and dispersal . Yet some datasets , especially those generated via specialised use cases such as within FinTech and healthcare , must always be located safely in a secure , on-premises or colocation facility . Here , examples might include NHS patient data , security sensitive information and large-scale data libraries such as those used in genomics or bioinformatics . For financial services organisations , the speed of processing quantitative data is also paramount , so data becomes timesensitive , which is another deterrent for hosting applications in the cloud .
For users requiring access to such data , accessibility is crucial and at Kao Data we have specifically partnered with Megaport to help fluid transit between cloud and colocation environments .
Our position as the closest hyperscale onramp to Cambridge enables businesses such as AI startups , pharmaceutical and life science organisations located in London or the UK Innovation Corridor , to gain the best of both worlds .
Of course , for enterprise organisations , cloud virtualisation offers several benefits including large economies of scale and the ability to host numerous IT workloads efficiently .
Yet cloud data centres still require access to land , low latency connectivity and resilient , renewable power . Hence the trend of hyperscale providers contracting multiple technology suites , or entire buildings with colocation operators .
To further support the cloud community , Kao Data has created a dedicated
‘ Built-to-suit ’ service , providing operators with access to a planning-approved campus on which to build bespoke and sustainable data centres .
Such services remove many of the construction and financial pain points , offering cloud service providers the ability to quickly build new availability zones and meet growing end-user demands .
As I say , it ’ s important to highlight that cloud and colocation are no longer competing , especially where hybrid cloud native HPC is concerned . Together , they form part of a global digital ecosystem where one will undoubtedly continue to require the other .

IT ’ S IMPORTANT TO HIGHLIGHT THAT CLOUD AND COLOCATION ARE NO LONGER COMPETING , ESPECIALLY WHERE HYBRID CLOUD NATIVE HPC IS CONCERNED . www . intelligentdatacentres . com

31