We are at the beginning of a long and exciting journey, where robotics can complement humans to improve data centre operations, says Giuseppe Leto, Global Data Centre Portfolio Manager at Siemens
For everyone involved in data centres, the evolution of robotics is of great interest. It is a topic that is getting relevant visibility in all kinds of industrial sectors and with that will affect data centres as well. It is well known that humans are still responsible for most of the errors in data centres: robots can help to reduce those and make centres more efficient.
Technological precursors
The development and adoption of deep learning algorithms gave a strong impulse to robotics. With the help of artificial intelligence, researchers at Siemens have developed a two-armed robot that can manufacture products without having to be programmed. As a sneak-peak of the future of automated production: the robot’s arms autonomously divide tasks and work together as one.
Similar solutions and approaches can be found within other market players as well, who look at those ML algorithms as key to developing increasingly intelligent applications that are able to improve our daily work.
In the data centre market, operators have been adopting infrastructure management systems – like DCIM (Data Center Infrastructure Management), or DDIM (Distributed Datacenter Infrastructure Management), as recently coined by Gartner – in order to bridge the gap between facility and IT departments, manage assets and capacity, and ultimately improve data centre operations.
Some DCIM vendors provide the software with ML algorithms and AI engines that can learn from the vast amounts of data generated by the assets. Uncommon behaviour is identified way faster by those algorithms and engines than humans could ever do.
In the future robots can extend – whether to act as an autonomous entity or to be driven by an Infrastructure Management software – to bring operations to the highest level of efficiency.
Identifying the benefits
Robotics will help data centres achieve a higher degree of autonomy by automating their operations. By having the ability to execute complex tasks without errors, robots will complement humans to better perform MAC (move, add, change) requests. This will ultimately reduce risks and maintain the SLAs in place with tenants, regardless of whether they are internal or external.
Time is one of the main benefits robotics brings to the data centre. Any change request can be executed at any time of the day – or night – 7 days a week, without any errors or delay. Delivery between shifts and inevitable reprioritisation of tasks is avoided and execution times can be reduced and become easy to predict.
The data centre operator will, with the help of robotics, experience a reduction of operating costs (e.g. OPEX) and be able to allocate resources to improve the processes of educating robots. Figures will have to be defined to make this transformation process a real success.
According to the Uptime Institute, 80 percent of those who had suffered an outage believe their biggest/most recent outage was “preventable”. Artificial intelligence can predict when an asset might break, with an accuracy that depends on the depth of the machine learning algorithm.
The fact that information exists but no corrective actions have been taken, suggests that human inertia increases the risk factor. Artificial intelligence can, with the help of a fully automated process, reduce MTTR (Mean Time To Repair).
Think about the benefit you could get if after the prediction of a failure the system is able to replace the defective piece without any human intervention. In circumstances where decision time is a critical factor to keep SLA safe, machines offer better risk prevention than humans.
The customer will perceive better control and respect of the agreed SLAs and the operator will be less exposed to penalties – even if today they are usually phrased so that the operator pays only a nominal penalty.
Achieving wide adoption
Despite the great promise, we must not get ahead of ourselves about the challenges in achieving wide adoption. Robots are able to perform elementary and repetitive tasks, but need specific training to operate in a productive environment such as a data centre.
Unlike assembly lines, for example in the automotive sector, where robots are installed along the production line – in data centres the robot has to be intervening in the rack. It has to be mobile and be able to replicate the same operations that today humans carry out. This requires specific training, unique for each data centre.
Robotics also require substantial investment in order to develop the knowledge and “educate” them to do what they will be used for – investments that at the moment are not universally viable. But as has been shown with automation, as time passes robotics will be increasingly viable even for the non-hyperscalers.
There is also a cultural shift that needs to occur before robotics are widely embraced. Due to modern art, like films or literature, humans associate robotics often with a future dominated by a conflict between robots and humans. This attitude has been around since the first industrial revolution when the adoption of steam engines aroused suspicion and repudiation because – it was said – thousands of jobs would be lost.
But, history has proven that technological innovation leads to the creation of additional jobs in different sectors. Each revolution involves an inevitable transformation that – as history has shown – has led to further growth in the market and well-being.
How robotics is shaping and shifting the next generation of data centre
Fri 30 Nov 2018 | Giuseppe Leto
We are at the beginning of a long and exciting journey, where robotics can complement humans to improve data centre operations, says Giuseppe Leto, Global Data Centre Portfolio Manager at Siemens
For everyone involved in data centres, the evolution of robotics is of great interest. It is a topic that is getting relevant visibility in all kinds of industrial sectors and with that will affect data centres as well. It is well known that humans are still responsible for most of the errors in data centres: robots can help to reduce those and make centres more efficient.
Technological precursors
The development and adoption of deep learning algorithms gave a strong impulse to robotics. With the help of artificial intelligence, researchers at Siemens have developed a two-armed robot that can manufacture products without having to be programmed. As a sneak-peak of the future of automated production: the robot’s arms autonomously divide tasks and work together as one.
Similar solutions and approaches can be found within other market players as well, who look at those ML algorithms as key to developing increasingly intelligent applications that are able to improve our daily work.
In the data centre market, operators have been adopting infrastructure management systems – like DCIM (Data Center Infrastructure Management), or DDIM (Distributed Datacenter Infrastructure Management), as recently coined by Gartner – in order to bridge the gap between facility and IT departments, manage assets and capacity, and ultimately improve data centre operations.
Some DCIM vendors provide the software with ML algorithms and AI engines that can learn from the vast amounts of data generated by the assets. Uncommon behaviour is identified way faster by those algorithms and engines than humans could ever do.
In the future robots can extend – whether to act as an autonomous entity or to be driven by an Infrastructure Management software – to bring operations to the highest level of efficiency.
Identifying the benefits
Robotics will help data centres achieve a higher degree of autonomy by automating their operations. By having the ability to execute complex tasks without errors, robots will complement humans to better perform MAC (move, add, change) requests. This will ultimately reduce risks and maintain the SLAs in place with tenants, regardless of whether they are internal or external.
Time is one of the main benefits robotics brings to the data centre. Any change request can be executed at any time of the day – or night – 7 days a week, without any errors or delay. Delivery between shifts and inevitable reprioritisation of tasks is avoided and execution times can be reduced and become easy to predict.
The data centre operator will, with the help of robotics, experience a reduction of operating costs (e.g. OPEX) and be able to allocate resources to improve the processes of educating robots. Figures will have to be defined to make this transformation process a real success.
According to the Uptime Institute, 80 percent of those who had suffered an outage believe their biggest/most recent outage was “preventable”. Artificial intelligence can predict when an asset might break, with an accuracy that depends on the depth of the machine learning algorithm.
The fact that information exists but no corrective actions have been taken, suggests that human inertia increases the risk factor. Artificial intelligence can, with the help of a fully automated process, reduce MTTR (Mean Time To Repair).
Think about the benefit you could get if after the prediction of a failure the system is able to replace the defective piece without any human intervention. In circumstances where decision time is a critical factor to keep SLA safe, machines offer better risk prevention than humans.
The customer will perceive better control and respect of the agreed SLAs and the operator will be less exposed to penalties – even if today they are usually phrased so that the operator pays only a nominal penalty.
Achieving wide adoption
Despite the great promise, we must not get ahead of ourselves about the challenges in achieving wide adoption. Robots are able to perform elementary and repetitive tasks, but need specific training to operate in a productive environment such as a data centre.
Unlike assembly lines, for example in the automotive sector, where robots are installed along the production line – in data centres the robot has to be intervening in the rack. It has to be mobile and be able to replicate the same operations that today humans carry out. This requires specific training, unique for each data centre.
Robotics also require substantial investment in order to develop the knowledge and “educate” them to do what they will be used for – investments that at the moment are not universally viable. But as has been shown with automation, as time passes robotics will be increasingly viable even for the non-hyperscalers.
There is also a cultural shift that needs to occur before robotics are widely embraced. Due to modern art, like films or literature, humans associate robotics often with a future dominated by a conflict between robots and humans. This attitude has been around since the first industrial revolution when the adoption of steam engines aroused suspicion and repudiation because – it was said – thousands of jobs would be lost.
But, history has proven that technological innovation leads to the creation of additional jobs in different sectors. Each revolution involves an inevitable transformation that – as history has shown – has led to further growth in the market and well-being.
Experts featured:
Tags:
AI automation Data Centre machine learning robotics robotsLatest Features
Google Analytics 4 (GA4) vs Universal Analytics: Preparing for the Shift
The impact of R&D tax relief reforms on UK tech businesses
Brucha: Panel Solutions for Data Centre Applications
How Whole System Thinking can deliver resilient energy for data centres
Selecting a building management system (BMS) for a data centre based on the d...