Library & FAQs

datacenter cooling is one of the fundamental aspects in managing and operating a datacenter, especially today when it hardware density is getting higher and higher. in general, the term of "datacente cooling" refers to the all the equipment, tools, techniques and processes necessary to ensure an ideal operating temperature within the datacenter facility
cooling a datacenters typically takes a lot of energy (>50% depending on location and design) and a lot of planning. traditionally, there are three different ways to cool a datacenter: air-based cooling, liquid-based cooling and a hybrid of the two, and within these categories are a couple different methods:
- the cold aisle/hot aisle method
- the cold or hot air containment method
- the in-rack heat extraction method
- water-cooled racks or rear-door chillers
- direct liquid cooling
- immersion cooling

a datacenter is a facility (it can be a single building or a group of buildings) housing a company's computer systems, servers, network and storage system and centralizing all the the it operations of the company.

colocation (also known as “colo” or “co-location”) refers to a fundamental set of technology infrastructure services provided by third-party “operators” to clients from one or more purpose-built buildings known as datacenters. these services most commonly include:

• secure, conditioned floor space suitable for 24/7 long-term operation of critical informationtechnology (it) equipment (servers, storage arrays, telecommunications gear, etc.)
• uninterrupted sources of electrical power designed to withstand unexpected outages
• local telecommunication connections to facilitate data network services to remote locations

open bath immersion cooling is a datacenter cooling technique that implies fully submerging servers in dielectric coolant fluid. it's called "open bath" because servers are placed side-by-side in large tanks that assimilate bath tubs. these baths operate at atmospheric pressure and allow the coolant fluid to be pumped through the hardware components or servers submerged in it.

a heatsink is an object that absorbs and dissipates the heat produced by any electronic or/and mechanical devices (in case of immersion cooling, to a surrounding fluid). the goal of a heatsink is to avoid any overheating of the device generating heat. usually, heatsinks are made of metal and disperse the heat throughout a combination of convection and conduction.

the compound annual growth rate (cagr) is a measure of growth. in particular, it measures the growth rate of a certain asset or investment (or anything that can fluctuate in value), over time (at least 3 to 5 years) while taking into account the overall compound growth. cagr is used to evaluate and compare past performance of investments, but also to project how much their return will be in future.

datacenter infrastructure efficiency (dcie) is a metric used to evaluate the power or energy efficiency of a datacenter. dcie represents the ratio of the total amount of energy consumed by all it equipment and resources to the entire energy consumption of a datacenter.

the concept of net-zero emissions refers to the process of achieving an overall balance between, on one side, all man-made greenhouse gas emissions, and on the other, the greenhouse gas emissions that are taken out of the atmosphere. this process is realized thanks to a series of reduction measures.

an evb (electric vehicle battery) is a rechargeable battery used to power the motor of a bev (battery electric vehicle) or of an hev (hybrid electric vehicle).

a microgrid is an energy system that operates self-sufficiently. this means that the microgrid can work connected to a bigger grid (macrogrid), but at the same time, it can be disconnected from the main grid and operate autonomously and independently. microgrids are used to provide energy in a smarter way to determined geographic areas, meeting specific energetic, economic or operational needs of customers.

the open compute project (ocp) was launched in 2011 by facebook, along with intel and rackspace, goldman sachs and andy bechtolsheim. the idea of the five funding members was to replicate in the hardware space, the basic principles of collaboration, creativity and openness already seen in action inthe open source software space.

dry coolers are heat-transfer devices that use air to remove excess heat into the atmosphere. usually, a dry cooler presents a heat exchanger (it can be either microchannel coils or a finned tube) and fans. the fans are responsible for directing the air flow through the heat exchanger. here is where the fluid (it can be water or another solution) is cooled and then circulated back into the system.

clearly, for a dry cooler to be efficient, the air temperature needs to be cooler than the water (or any other glycol solution) in the system. this explains why, for example, dry coolers work better in cold climate regions (while wet cooling towers are more efficient in hot and dry climates). dry coolers tend to not be affected by air humidity, on the contrary of wet cooling towers.

dry coolers have a limited cooling capacity, compared to wet cooling towers: the first one only brings the water temperature down to the ambient temperature (additional cooling can be achieved in other ways), while the latter can bring the water temperature up to 5 degrees down the ambient temperature.

immersion cooling, also known as liquid submersion cooling, it is the practice of submerging computer components (or full servers) in a thermally, but not electrically, conductive liquid (dielectric coolant). liquid submersion is a routine method of cooling large power distribution components such as transformers. still rarely used for the cooling of it hardware, this method is slowly becoming popular with innovative datacenters the world over. it hardware or servers cooled in this manner don't require fans and the heat exchange between the warm coolant and cool water circuit usually occurs through a heat exchanger (i.e. heater core or radiator).

direct liquid cooling (dlc) mostly refers to water or other liquids that are delivered directly to a hot plate which is only on cpus and gpus, dealing with 70% of the heat. with rear door cooling and direct liquid cooling, it is normally advised to have a secondary form of cooling such as a traditional crac unit to handle the excess heat in the room.

the central processing unit (cpu) represents the brains of a pc. from an architectural point of view, a cpu is usually composed of few cores (on the contrary of a gpu), with lots of cache memory. in terms of capacity, a cpu is less powerful than a gpu, being able to handle only a few software threads at a time. cpus are good for serial processing.

overclocking is the action of running a computer component (tipically it is a cpu or a gpu,but it can be another component) and run it a higher speed than the manufacturer designed it for.

a graphics processing unit (gpu) is a specialized computer chip. gpu's can deliver a better performance than a cpu (central processing unit). it is an electronic circuit designed to rapidly manipulate and alter memory to accelerate the creation of images in a frame buffer intended for output to a display device. gpus are used in embedded systems, mobile phones, personal computers, workstations, and game consoles. a graphics processing unit (gpu) is a computer chip that performs rapid mathematical calculations, primarily for the purpose of rendering images. in the early days of computing, the central processing unit (cpu) performed these calculations. as more graphics-intensive applications such as autocad were developed, however, their demands put a strain on the cpu and degraded performance. gpus came about as a way to offload those tasks from cpus and free up processing power.

an original equipment manufacturer (oem) is a company that makes components or equipments that are then assembled and used in the products of another company (it can be a manufacturer or a reseller). the client company - who can also configure those components and equipments - then sells the finished item to users.

blockchain is a record-keeping, distributed technology that allows data storage globally on thousand of servers.the bitcoin network for example, is based on the blockchain technoloy. blockchain, as the name itself suggests, is a chain of blocks, where "chain" stands for "a public database" and "block" for "digital information". each of the blocks contain digital information such as:

• date, time and dollar amount of the most recent purchase of the user
• the participants in the transaction
• the specific identity of the block itself, that helps distinguish it from other blocks

return on investment (roi) represents a ratio between net profit and the cost of investment. the roi is used as a metric to determine the efficiency of an investment or multiple investments.

operating expenditure (opex) corresponds to those expenses that a company incurs on a daily basis to run a business or system. the index depends on the type of industy. for datacenters, two key factors determining the opex are maintenance and electricity costs.

capital expenditures (capex), are funds that a company uses for acquiring, upgrading or maitaining certain physical assets (e.g. buildings, equipment, warehouse, technology, etc.) in the datacenter industry, some of the elements that concur to determine the capex are: design, real-estate, construction, it hardware, cables, cooling, piping etc.

the water usage effectiveness (wue) is the ratio between the annual site water usage in liters by the it equipment energy usage in kilowatt hours (kwh). the sustainability metric was created by the green grid in 2011 to attempt to measure the amount of water used by datacenters to cool their it assets. wue takes into account water used for cooling, regulating humidity and producing electricity on-site.

power usage effectiveness (pue) is a metric representing the ratio of total amount of energy used by a computer data center facility to the energy delivered to computing equipment. it is used to determine the energy efficiency of a data center.

total cost of ownership (tco) refers to the direct and indirect costs of a product or system. the tco is calculated taking into consideration the capex (capital expenditures) and the opex (operational expenditures) indexes.

a kilowatt (kw) is a unit of measurement of energy transported over time. 1 kilowatt corresponds to one thousand watts. watt and kilowatt are used to express the output power of engines and the power of electric motors, tools, machines, and heaters.