Who keeps an eye on the equipment while the IT manager is on holiday?

You have probably all seen the NASA control center in a film. A room with a large screen and some smaller monitors, filled with people intently watching the screen and knowing the status and vital parameters of every piece of equipment on board the spacecraft. They will be the first to know if it reached orbit successfully or if the Mars landing was a success. Of course, they will also be the first to know if anything goes wrong.

Network operations center at Greenergy Data Centers. Photo: Tõnu Tunnel.

A similar control center, with a huge screen and a dedicated staff, can also be found in larger data centers. They will not be going into space, but much like the people working at NASA, they have a critical role to play in the success of the mission. Their mission is ensuring the operation of the vital systems of organisations active in the digital world.

Why is this necessary?

Everything we see on the internet resides on physical IT devices, which, in turn, are located in data centers. Put simply, if something were to go wrong, a piece of our digital world would stop working. To avoid this, we need a control center or a network operations center (NOC) as it is called in data centers.

‘Just like astronauts can enjoy a peace of mind thanks to the back-up forces on Earth keeping an eye on everything, the peace of mind of IT managers is ensured by specialists in the NOC,’ Toomas Kell, head of operations at Greenergy Data Centers, draws an analogy.

According to him, customer feedback shows that this is especially important now, during the holiday period. ‘IT managers, too, want to disconnect once in a while and take some time off. Keeping their equipment in the data center allows them to do this without worrying. If there are any problems, they will let us know – but in the meantime, they can rest easy,’ says Kell.

Keeping constant watch

To justify a large-scale NOC, it needs to be manned around the clock. This ensures the best balance between automation and human decisions. In the largest data center in the Baltics, information from more than 3,000 nodes, devices, or sensors reaches a central point, which the software processes into more easily understandable signals and infographics. For example, coloured lines and infographics on the screens show, among other things, the temperature and humidity in the server rooms, the status of the cooling node and the speed of the cooling fans, as well as the indicators of security and fire-fighting systems. An experienced eye can glean a lot from these graphs and indicators and this information, and therefore anticipate problems or react to them if necessary. Each node can also be isolated from the big picture and the situation can be examined in more detail, down to the level of individual devices and sensors.

‘In fact, from one proper NOC, you can coordinate the work of several buildings in different locations. For example, the NOC might be in Estonia and the building being controlled in Poland. We are also planning to monitor and manage the next data centers across the border from here,’ says Kell, who recently exchanged experiences with a data center in Barcelona, where the NOC monitored and managed the daily operations of three centers.

The NOC can also be a good source of information for the customer. Through dedicated one-way channels, a constant flow of information on the vital environmental parameters of the equipment rooms is ensured, and the IT manager can keep an eye on the humming equipment and the environment of the room from anywhere in the world. It is not really a holiday activity, but it does provide an extra layer of security.

‘The NOC can also be an extension of the customer. For example, if you want to reboot a piece of equipment or change the connections between machines, you will not have to come in. The specialists in the NOC can organise the work. This saves a lot of time,’ explains Kell.

What information does the NOC collect and analyse?

First and foremost, the NOC is a collaboration between people and technology to keep an eye on the key parameters of different equipment, systems, and equipment rooms, such as their statuses, pressures, temperatures, airflows, humidity levels, operating speeds, etc. ‘The importing part is comparing the information with how one or another piece of equipment, system, or equipment room should be performing. For example, the amount of water circulating in a cooling system, the position of valves or taps, the air quality in server rooms, whether there are compounds that should alert the fire suppression systems, whether oxygen levels are within normal limits in rooms with reduced oxygen,’ Kell gives examples. In addition, information from security systems is also gathered there, and the staff of the NOC has a complete overview of what is happening on the premises of the complex. If someone tries to climb over the fence of the data center, it will be registered by cameras as well as vibration and movement sensors, meaning that alarms will come to the NOC from several directions.

‘There is also an unexpected side to this. As our data center is located near Tallinn and surrounded by open countryside, the lives of various animals are also caught on our cameras. For example, we have seen a family of wild boars, hares, and the hunting expeditions of foxes behind our main gate,’ says the person in charge of the day-to-day operations of Greenergy Data Centers, giving examples of the versatility of the data collected.

Is artificial intelligence the future?

After all, a data center is, by its very nature, a vast array of equipment and automation. Could its management be handed over to a clever AI? The role of AI in the day-to-day running of data centers is undoubtedly a growing trend. Systems are becoming more sophisticated and accurate, and new technologies and software solutions are emerging, creating unprecedented opportunities. The research firm Gartner predicts that by 2025, half of all data centers will use artificial intelligence and machine learning in one way or another. Just a few years ago, the risks seemed too great, but the world is changing fast.

‘For example, the data center of GDC is one of the first in the world to use the self-learning software from Siemens to manage and optimise the cooling system, which can use data collected and analysed in the past to decide, for example, where and how much cool air needs to be channelled in the server rooms. In addition, optimising the operation of the cooling system helps to prolong the lifetime of equipment and systems and also reduces running costs. Google and other tech giants have also successfully applied artificial intelligence to optimise specific parts of the workflow,’ says Kell, illustrating the ongoing technological advances. However, AI will not be taking over the comprehensive management of a data center with hundreds of clients any time soon, because too much is at stake. Humans are able to perceive context better and ultimately make better decisions. At least for now. Actually, there is currently a global shortage of highly skilled potential employees for data centers. Inspired by this, attempts are being made to replace them, at least in part, with robots.

Returning, however, to the example of the control center of a rocket given at the beginning of the article, it is difficult to imagine the command and control center of NASA being run by robots or by artificial intelligence. The same goes for the NOC of a data center. Both the astronaut and the client of the data center will feel more comfortable knowing that, in addition to the machines, their well-trained humans are monitoring their well-being.

In short, the NOC is the brain of the data center. It brings together information from thousands of sources, and the people who work there make sure that everything works properly. The better the NOC is organised, the better the IT manager can rest.