Software-defined data centres (SDDC) are often considered to be the next step in the evolution of virtualisation and cloud computing. Able to provide solutions to both legacy enterprise applications and new cloud computing services, the SDCC approach has had a positive impact in many data centres.
According to Edward van Leent, Chairman and CEO of EPI, SDCC has helped to reduce physical and power utilisation footprint as well as the number of staff required to run operating instances. However, although deployment is continuing to increase, it will not result in a total phase-out of legacy or on-premise data centres.
“First of all, companies have heavily invested in hardware, software and personnel and to think that they will throw that all away in a few months in favour of SDDC or cloud environment is just not realistic. Besides that, organisations have come to learn that cloud is not the perfect fit for all, and as such, the market will see a mix of on-prem with cloud or the so-called hybrid environments. Whether the weight is more on cloud vs on-prem will very much depend on each organisation themselves.”
In order to remain competitive in the industry, Edward, who has over three decades of experience in working with global high tech, high-availability and mission-critical data centres, pointed out that data centre providers need to understand that customer requirement is of top priority. While service levels and costing are important factors as well, he said that the fight for price per rack has evolved in other factors such as cost for storage, CPU utilisation and most of all, cost of connectivity.
“Organisations have come to realise that moving data into the cloud is often cheap as chips, but moving data out of cloud comes at a high cost. The competition will see those costs eventually reduce in order to win over customers. Outsourcing to cloud is often a long-term game, so making sure you win a customer is important as losing them could mean a loss of many years.”
As an active member of various data centre professional groups, Edward mentioned that there are several standards and guidelines in the market on how to design and build a data centre. However, as there never was a standard on how data centre operators or owners could effectively and efficiently run data centre operations, Data Centre Operations Standard (DCOS) was developed to fill that gap.
Edward explained that DCOS covers all 11 disciplines to run a data centre and is written in a very practical way. This helps the customer use the DCOS immediately compared to the typical ISO standards where they need to hire consultants to understand what has to be done.
For large players to decide where to set up their data centre, Edward said there are a number of factors influencing their decision, but the key one is connectivity. “One would [also] look at skilled labour, ease of labour laws, cost of power and connectivity, cost of operations including taxes and stability and laws of the country.”
Speaking about Singapore’s emergence as a data centre hub in the ASEAN region, he added, “Singapore is very business friendly with a government that clearly understands what investors are looking for and the willingness to move fast to ensure new business requirements are met as soon as possible. That, together with its great stability, has made it an almost de facto choice. For other countries to mimic this is not an easy task, and as such, I believe Singapore will continue to be the place of choice for many more years to come.”
Going back to SDCC, new technologies are becoming highly in demand for data centres. One such technology is Artificial Intelligence. AI is highly sought after when it comes to enabling automation in many areas of IT today. Interestingly, Edward has a different view of this.
“AI is a word which is often abused in the industry. It’s almost a fashionable word to be used to give others the impression that the organisation is leading. Fact is that most of the AI implemented as we know it today is merely slicing and dicing of data upon which a decision is made. It is, therefore, much more data-based decision than true artificial intelligence. Having said that, no matter at what stage various so-called AI solutions are, all of them are typically aimed at automation to reduce the cost of operations and to drive (energy) efficiency. Whether it is by automatically shifting workloads or influencing air condition behaviour based on actual requirements.”