Data Center risks: choose a location
“Only those who consider it possible can avoid a catastrophe.”
V. Schwebel
We are increasingly dependent on the achievements of progress: we read mail in cinemas, we mark the places of our presence in foursquare. And business has become no less dependent on technical achievements. And if for us the breakdown of the phone becomes a small inconvenience, then for companies the failure of any element of the IT infrastructure turns into a huge loss. One hour of idle time of a Russian bank, included in the TOP-100, is equal to the cost of a luxury car. Now imagine the size of losses and lost profits, if the corporate data center collapsed walls or a heating main broke near it. Do services start there quickly? How long will it take to restore health if there is no backup data center?
Such a catastrophe can be avoided by initially designing the data center, paying attention to its location, the effectiveness of the solutions used in it, energy intensity, reliability, and payback.
')
There are two world standards on which to focus when designing a data center: BICSI 002 2011 Data Center Design and Implementation Best Practices (6: Site selection) and TIA-942. They contain many factors that influence the future data center:
• Natural : such as seismic activity, mobility and soil moisture, groundwater and river beds, the effect of wind, height above ground level.
• Man-made : the level of dust and vibration, noise, proximity to highways, airports and parking lots.
• Public life phenomena : the criminal situation in the area and even the presence of several tenants in the building.
The manifestation of one of these factors at least once a decade can destroy or for a long time suspend the work of the data center with the finest IT content. That is why abroad, great attention is paid to preparing for emergency situations at the design stage, guided by the principle: forewarned is forearmed.
The same reasons gave rise to the trend of building data centers from scratch, and not in existing buildings. In an effort to minimize risks, Western designers, as a rule, place data centers outside of megacities and densely populated administrative units. Sometimes these purposes are served by old industrial buildings with a good bearing capacity of floors, high ceilings and spacious squares. This greatly simplifies the technical solutions, since there is no need to adapt to the existing premises and operating conditions of the entire building.
“Every unpleasant event is unexpected, even if it was prepared for it.”
E. Severus
Do you think that your data center is not threatened by disaster? - A controversial statement. Here are some examples to consider:
• July 2013. As a result of the strongest thunderstorm, the Moscow-City complex flooded elevator mines and underground parking.
• May 2013. A strong wind in Bashkiria blew off the roofs of a dozen buildings, including a district hospital and a polyclinic.
• April 2013. Due to subsidence of soil in the Nizhny Novgorod region, three buildings went underground.
• October, 2011. A heating breakthrough flooded one of the most popular nightclubs in Yekaterinburg with hot water.
Are you still sure that it is not necessary to foresee external factors? No longer? 85% of the owners of Russian data centers will not agree with you. Their data centers are located in buildings not intended for this. As a rule, there are two reasons for this: either the desire to have a data center closer to the center of business activity (recalling the communication channels, raider seizures and visits to the organization by various law enforcement officials), or very poor opportunities when choosing a room.
And you can also make mistakes when designing. So recently, the number of requests for upgrading the data center, built in the 90s, when no one made special demands on the engineering infrastructure, has significantly increased. Built as God put on the soul. Therefore, now the most common problems of the data centers of the 80s-90s are an inefficient cooling system, lack of conditions for upgrading and increasing equipment (with an average update time every three years), low efficiency of power supply, unsystematic development of SCS with a web of patch cords of different lengths, different categories and all colors of the rainbow. To organize all this chaos, it is necessary to introduce individual engineering solutions for equipment and reinforce building structures.
In addition, many customers are not sufficiently responsible approach to the design of server rooms with a capacity of 10 kW, believing that the problems with them are less than with megawatt data center. The task often sounds like: “I need a simple, maaaaa little server. Here I have a duct passes, and here the heating pipe. But this is nothing! Yes, I forgot to say, I only have a place under the data center in the basement ... ”. Convincing the customer that cable routes and engineering networks that are not related to the data center need to be moved outside the data center, or it is extremely difficult for a designer to choose another place for the server.
As they say, to solve a problem is to find a person who will solve this problem. Therefore, first of all, you need to find a responsible and sensible contractor. Collaboration of an integrator with a customer is an integral part of the planning stage, on which key goals and objectives depend. The quality of the concept development directly affects the successful implementation of the project and the cost-effectiveness of the data center during its operation. Therefore, the experience of the integrator in this field and the portfolio of projects successfully completed by him are mandatory when allocating a credit of trust.
We do not forget to carefully ensure that no unnecessary and dangerous communications take place near the future data center. Such communications, for example, can be heating mains or conduits. If the load on them increases or they break through due to disrepair, then the data center may be in danger of flooding or overheating. Moreover, in case of large-scale industrial accidents, the consequences turn out to be even more deplorable: the walls are deformed due to temperature drops, sedimentary cracks appear from rising water levels and ground movement and the data center can cease to exist.
Another danger is insufficient bearing capacity of building structures. It is not for nothing that international standards recommend building separate buildings for the data center. This is primarily due to the fact that the equipment used in data centers has a lot of weight, far exceeding the capabilities of standard ceilings. Imagine how your neighbors from the offices below will be delighted to see your diesel generator set or UPS instead of the workplace of your boss (and with a particularly lucky set of circumstances and instead of the boss himself). It is worth paying special attention to the access roads and the ways of carrying large-sized equipment to the data center, incl. dimensions of doorways and lift capacity. It is unpleasant and moreover, it is very inconvenient to supply the UPS to the 10th floor window using a mobile crane or to memorize the sequence of actions of engineers from Europe, disassembling the super highland server for its entry into the building in parts, and most importantly then remember this sequence during assembly. But such situations, unfortunately, occur very often.
Of great importance is the return on investment. Sometimes it is more profitable to make significant investments at the beginning of the project so that they quickly pay off later. To reduce operating costs, you can use energy-efficient solutions and budget planning. For example, while working on one of the projects, we developed a scheme for using a system of subceiling air conditioners instead of precision ones for cooling the data center room. The advantages of such a solution is the principle of natural air convection, which means lower costs for the movement of air masses, isolation of corridors. In addition, the advantages include ease of maintenance and high maintainability, energy efficiency and smooth control over a wide range of external temperatures and internal loads, the ability to scale and regulate the mode of operation. According to its technical characteristics, the subcapital blocks are not inferior to floor precision air conditioners, and in some cases even surpass them.
Our experience and the experience of colleagues in the workshop shows that for the successful implementation of the project are necessary:
• Responsible approach to creating a data center both at the customer and at the integrator.
• Focus on modern technologies and the use of foreign developments.
• Use effective solutions.
• Individual approach to each new object.
• The emergence and use of innovative ideas and technologies.
We wish you success in designing the data center!
In the future, we will talk about other errors that may affect the smooth operation of your data center, and how to avoid them in the subsequent stages of creation.
V. Schwebel
We are increasingly dependent on the achievements of progress: we read mail in cinemas, we mark the places of our presence in foursquare. And business has become no less dependent on technical achievements. And if for us the breakdown of the phone becomes a small inconvenience, then for companies the failure of any element of the IT infrastructure turns into a huge loss. One hour of idle time of a Russian bank, included in the TOP-100, is equal to the cost of a luxury car. Now imagine the size of losses and lost profits, if the corporate data center collapsed walls or a heating main broke near it. Do services start there quickly? How long will it take to restore health if there is no backup data center?
Such a catastrophe can be avoided by initially designing the data center, paying attention to its location, the effectiveness of the solutions used in it, energy intensity, reliability, and payback.
')
World practice
There are two world standards on which to focus when designing a data center: BICSI 002 2011 Data Center Design and Implementation Best Practices (6: Site selection) and TIA-942. They contain many factors that influence the future data center:
• Natural : such as seismic activity, mobility and soil moisture, groundwater and river beds, the effect of wind, height above ground level.
• Man-made : the level of dust and vibration, noise, proximity to highways, airports and parking lots.
• Public life phenomena : the criminal situation in the area and even the presence of several tenants in the building.
The manifestation of one of these factors at least once a decade can destroy or for a long time suspend the work of the data center with the finest IT content. That is why abroad, great attention is paid to preparing for emergency situations at the design stage, guided by the principle: forewarned is forearmed.
The same reasons gave rise to the trend of building data centers from scratch, and not in existing buildings. In an effort to minimize risks, Western designers, as a rule, place data centers outside of megacities and densely populated administrative units. Sometimes these purposes are served by old industrial buildings with a good bearing capacity of floors, high ceilings and spacious squares. This greatly simplifies the technical solutions, since there is no need to adapt to the existing premises and operating conditions of the entire building.
How is it designed in Russia?
“Every unpleasant event is unexpected, even if it was prepared for it.”
E. Severus
Do you think that your data center is not threatened by disaster? - A controversial statement. Here are some examples to consider:
• July 2013. As a result of the strongest thunderstorm, the Moscow-City complex flooded elevator mines and underground parking.
• May 2013. A strong wind in Bashkiria blew off the roofs of a dozen buildings, including a district hospital and a polyclinic.
• April 2013. Due to subsidence of soil in the Nizhny Novgorod region, three buildings went underground.
• October, 2011. A heating breakthrough flooded one of the most popular nightclubs in Yekaterinburg with hot water.
Are you still sure that it is not necessary to foresee external factors? No longer? 85% of the owners of Russian data centers will not agree with you. Their data centers are located in buildings not intended for this. As a rule, there are two reasons for this: either the desire to have a data center closer to the center of business activity (recalling the communication channels, raider seizures and visits to the organization by various law enforcement officials), or very poor opportunities when choosing a room.
And you can also make mistakes when designing. So recently, the number of requests for upgrading the data center, built in the 90s, when no one made special demands on the engineering infrastructure, has significantly increased. Built as God put on the soul. Therefore, now the most common problems of the data centers of the 80s-90s are an inefficient cooling system, lack of conditions for upgrading and increasing equipment (with an average update time every three years), low efficiency of power supply, unsystematic development of SCS with a web of patch cords of different lengths, different categories and all colors of the rainbow. To organize all this chaos, it is necessary to introduce individual engineering solutions for equipment and reinforce building structures.
In addition, many customers are not sufficiently responsible approach to the design of server rooms with a capacity of 10 kW, believing that the problems with them are less than with megawatt data center. The task often sounds like: “I need a simple, maaaaa little server. Here I have a duct passes, and here the heating pipe. But this is nothing! Yes, I forgot to say, I only have a place under the data center in the basement ... ”. Convincing the customer that cable routes and engineering networks that are not related to the data center need to be moved outside the data center, or it is extremely difficult for a designer to choose another place for the server.
How to avoid mistakes? Share experience
As they say, to solve a problem is to find a person who will solve this problem. Therefore, first of all, you need to find a responsible and sensible contractor. Collaboration of an integrator with a customer is an integral part of the planning stage, on which key goals and objectives depend. The quality of the concept development directly affects the successful implementation of the project and the cost-effectiveness of the data center during its operation. Therefore, the experience of the integrator in this field and the portfolio of projects successfully completed by him are mandatory when allocating a credit of trust.
We do not forget to carefully ensure that no unnecessary and dangerous communications take place near the future data center. Such communications, for example, can be heating mains or conduits. If the load on them increases or they break through due to disrepair, then the data center may be in danger of flooding or overheating. Moreover, in case of large-scale industrial accidents, the consequences turn out to be even more deplorable: the walls are deformed due to temperature drops, sedimentary cracks appear from rising water levels and ground movement and the data center can cease to exist.
Another danger is insufficient bearing capacity of building structures. It is not for nothing that international standards recommend building separate buildings for the data center. This is primarily due to the fact that the equipment used in data centers has a lot of weight, far exceeding the capabilities of standard ceilings. Imagine how your neighbors from the offices below will be delighted to see your diesel generator set or UPS instead of the workplace of your boss (and with a particularly lucky set of circumstances and instead of the boss himself). It is worth paying special attention to the access roads and the ways of carrying large-sized equipment to the data center, incl. dimensions of doorways and lift capacity. It is unpleasant and moreover, it is very inconvenient to supply the UPS to the 10th floor window using a mobile crane or to memorize the sequence of actions of engineers from Europe, disassembling the super highland server for its entry into the building in parts, and most importantly then remember this sequence during assembly. But such situations, unfortunately, occur very often.
Of great importance is the return on investment. Sometimes it is more profitable to make significant investments at the beginning of the project so that they quickly pay off later. To reduce operating costs, you can use energy-efficient solutions and budget planning. For example, while working on one of the projects, we developed a scheme for using a system of subceiling air conditioners instead of precision ones for cooling the data center room. The advantages of such a solution is the principle of natural air convection, which means lower costs for the movement of air masses, isolation of corridors. In addition, the advantages include ease of maintenance and high maintainability, energy efficiency and smooth control over a wide range of external temperatures and internal loads, the ability to scale and regulate the mode of operation. According to its technical characteristics, the subcapital blocks are not inferior to floor precision air conditioners, and in some cases even surpass them.
Our experience and the experience of colleagues in the workshop shows that for the successful implementation of the project are necessary:
• Responsible approach to creating a data center both at the customer and at the integrator.
• Focus on modern technologies and the use of foreign developments.
• Use effective solutions.
• Individual approach to each new object.
• The emergence and use of innovative ideas and technologies.
We wish you success in designing the data center!
In the future, we will talk about other errors that may affect the smooth operation of your data center, and how to avoid them in the subsequent stages of creation.
Source: https://habr.com/ru/post/189448/
All Articles