In essence, the data centre consists of two realms. The first is the IT realm, governing the systems that address the information processing aspects, such as servers, storage arrays, and network switches. The second is that of the physical infrastructure and controls that allow IT to perform its function, and includes all physical infrastructure that support both IT, and the data centre facility itself, including power, cooling, and security systems.
And although the two realms are interconnected, the subsystems within each are procured, managed and operated by different groups of users. Typically, the facilities and engineering teams own and operate the physical infrastructure, and IT department operates the IT realm. Sometimes, the two teams share a common backbone, and as the data centre evolves, they become increasingly intertwined as do the management systems that support them.
Data centres are at the heart of the digital world that controls our business and personal lives. Our society is growing increasingly dependent the internet, and without a data centre, video streaming, remote control of energy consumption, transformation and digitisation across the board, would not be possible. We would be unable to be driven by an autonomous car, to take distance learning courses at university. Anyone who has tried to separate a teenager from their phone will understand the generational change, not only in wanting to be connected but needing to be. The digital economy and new uses of augmented reality, virtual reality and artificial intelligence are generating a wealth of data that is stored, processed and secured 24/7 in data centres.
By 2020, analysts predict there will be 50-billion connected objects, and 4-billion people connected to the internet. Over the next five years, 20 times more equipment than people will be connected to the internet, representing 10 zettabytes of annual IP traffic.
Ten years ago, analysts thought that all internet uses would go through the cloud and predicted the “death” of on-premise data centres. Today, we are witnessing a step back. It is clear that the internet of things (IoT) will soon produce more data than can be transported to large, centralised data centres. One-hundred percent cloud architecture is not enough. Data flow requirements, burgeoning amounts of data, and the need for security, have given rise to the hybrid data centre ecosystem that is made up of huge, centralised cloud data centres, and their smaller, enterprise-class counterparts. In addition, we have colocation data centres (regional edge) and micro-data centres (local edge), which prevent networks from being inundated, by processing the data as close as possible to where it is produced.
The cloud-centralised network architecture was not designed to optimise latency. Firstly, because users require quick access to data and applications, and cloud networks are too far away, and require too many steps to reach the user. Secondly, due to bandwidth limitations, and thirdly, due to the regulatory environment.
To aggregate and process these large amounts of data, there is a need to deploy localised data centre solutions closer to users, data, and objects, to avoid latency issues and effectively manage the real-time capabilities required by the IoT. This is called edge computing, which sees the type of applications and trades determine where the compute resources will be physically located.
These trends are also leading us to a hybrid and complex data centre ecosystem where cloud-centric data centres coexist with regional data centres, and local micro data centres are found on the edge. At Schneider Electric, we solve these challenges head on through energy infrastructure solutions tailored to each type of data centre, from the largest to the smallest.
In this complex environment with more data centres, interconnected with a larger number of sites, and an exploding volume of data, the management and maintenance of this ecosystem become the next challenge. Today, the datacentre or IT manager doesn’t only have to deal with the centralised data centre, but also with myriad complexities in a distributed, and hybrid IT environment. “On-prem” data centre infrastructure management (DCIM) alone does not cut it anymore, and that’s where DCIM-as-a-service steps in. This gives businesses the ideal way to leverage the potential IoT and big data with the addition of over the top services such as those that handle massive data crunching and carry out analytics to provide insights to the end user.
EcoStruxure is our answer. It is a platform of technologies composed of three levels: the connected products which collect data and send them to an edge; control software which ensures the monitoring, alarms; and if applicable, the control of equipment. In addition to this, there are cloud hosted applications / data analytics and digital services that will bring visibility, simplification, automation, predictive analytics for a faster and more proactive response
This integrates the management of the three main architecture blocks – the clean room, the cooling of the building and, of course, power. But to meet the challenges facing our industry, including complexity, dependence and the energy paradox, we believe that a management system hosted in the cloud will be the only way to solve these challenges.
Send your comments to engineerit@ee.co.za