With the rise of AI and machine learning, energy demands are greater than ever. These and other technologies now hitting the market draw vast amounts of power, generating immense amounts of heat. Hence the need for more advanced cooling technology.
In response to increased and more intense workloads, data centers have packed more computational power in each of their racks. While high-performance, processor-heavy computational loads have already maxed out traditional cooling capacity, even average data centers now have to start making adjustments. The key takeaway here: Old, dated cooling technology just cannot handle modern cooling demands.
However, if you’re looking to transition to a solution that’s better, simpler, quieter, plus a lot more cost-effective than other cooling methods, saving on both CAPEX and OPEX costs, there are cooling technology basics that every chief technology officer (CTO) should know.
Single-Phase Immersion Cooling to Meet Increasing Demands
Single-phase immersion cooling technology represents the leading technique to address the challenge of rising data center temperatures. High-performance computing trends will only increase in the coming years, requiring innovative techniques to handle the heat. Liquid immersion cooling technology is battle-tested and head and shoulders above the alternatives on several fronts – greater upfront affordability, easier operations, less maintenance.
With single-phase immersion cooling, servers are installed vertically in a coolant bath of a hydrocarbon-based dielectric fluid that’s similar to mineral oil. Like its two-phase counterpart, heat is transferred to the coolant through direct contact with server components. But, unlike two-phase immersion cooling, the coolant does not boil off. Instead, it remains in the liquid phase (hence the name) and is cooled via a heat exchanger in a cooling distribution unit (CDU).
Single-phase immersion has proven to be cost-effective and highly reliable leveraged by GRC for a decade now. Our ICEraQ and ICEtank solutions are now helping some of the world’s largest cloud, enterprise, government, education, and telecom organizations grow while future-proofing their data centers.
A GRC Cooling Solution That Slashed an Energy Bill in Half
Geosciences firm CGG had a 10,000 square foot, legacy data center that had been mothballed because of things such as limitations of the available power envelope and an extremely inefficient Power Usage Effectiveness (PUE).
By partnering with GRC and leveraging our patented liquid immersion cooling technology, ICEraQ, CGG was able to re-energize its aging data center with immersion cooling that literally doubled its capacity for hundreds of servers and slashed its energy bill in half! The transition took place in only a few weeks without any downtime, no expensive air processing equipment, no building modifications.
Liquid-Based Cooling Beats Air Cooling
The underlying principle underpinning both single – and two-phase immersion cooling technology is that liquid is a much better heat conductor than air. What makes both these immersion-based systems superior is the proximity of the liquid to the heat source. Furthermore, both options require no chillers, CRACs, CRAHs, or raised floors. Bear in mind though, that there are significant differences between the two technologies.
Air cooling requires several extremely large, complex, and expensive pieces of machinery necessitating the need for more power, driving up costs further. And then there are maintenance costs. Liquid cooling systems, like GRC’s, perform better than traditional air cooling systems. And not just slightly better… by a wide margin.
Liquid immersion cooling is a simple, yet extremely effective technique that costs less and works in nearly any location. That same simplicity also makes liquid immersion cooling fundamentally more efficient. Capital expenditure goes down, as is the case with operational expenses when compared with air cooling systems.
Liquid can pull heat away from servers over a thousand times better than air. Dated air cooling systems eat away much of a data center’s budget pushing air through the servers. Immersion cooling hums away at a small fraction of the cost, cutting out 90% of your electricity bill.
These immersion cooling tanks also protect expensive IT equipment against dust and other air particles with the liquid acting as a buffer. Air cooling systems, on the other hand, bring with them more contaminants.
The biggest advantage of immersion cooling over air cooling, however, is in raw performance. Air cooling starts to lose its grip above around 15 kW per rack. Modern hardware for artificial intelligence and other applications can easily exceed that. Super-efficient immersion cooling can manage 100-200 kW per rack. It’s thus not hard to see why more and more computationally intensive data centers are moving to liquid immersion cooling.
More Powerful GPUs Heats Things Up
In addition to central processing units (CPUs) becoming more and more powerful, there’s a trend towards massive graphics processing units (GPUs) that devour energy and generate intense heat. In fact, manufacturers like Intel, NVIDIA, and AMD are producing chips that chew through half a kilowatt each! That’s several times more than what chips consumed just a few years ago.
People are increasingly using GPUs and other powerful processors for new uses like artificial intelligence (AI). Throw several GPUs onto a server, then several servers onto a rack, and you have tens or hundreds of thousands of cores to cool. High-performance computing is forcing data centers to pivot to new and more effective cooling technologies, simply because old air cooling methods just cannot handle the volumes of heat generated heat.
Put plainly, more powerful GPUs spell the end of air-cooled data centers. Now is the time to prepare for the future and liquid immersion cooling is that future.
Liquid Immersion Cooling is Future-Proofing
Financial, medical, scientific, and other applications will only put greater demands on data centers and their cooling systems. Immersion cooling offers the solution, effortlessly dissipating heat of over a hundred kilowatts per rack, more than enough even for modern GPU workloads.
Outdated air cooling technology can hardly keep up with today’s needs, let alone tomorrow’s. And given the direction in which the IT industry is rapidly heading, liquid immersion cooling (with the power and flexibility to handle present and future computational demands) is becoming a critical piece of the infrastructure puzzle.
Switch to Liquid Immersion Cooling With GRC
High-density server rooms are increasingly common as businesses increasingly rely on cloud-based solutions, requiring greater computing power and storage capacity. Data center challenges are mounting fast, fueled both by the rise of next-gen applications, super-hot, largely GPU-driven applications like AI. It’s thus critically important to leverage cost-effective cooling technology that not only affects performance, but generates savings, environmental benefits, and far larger capacity.
With processor-heavy computational loads, there’s an urgent need for a viable alternative. That solution is liquid immersion cooling. The tremendous efficiency of liquid immersion cooling in comparison to previous approaches to cooling makes it an easy call and you’ll immediately see much larger cooling and computing capacity, easier maintenance, lower CAPEX, and OPEX.
As temperatures rise inside data centers, operators too need to rise to the challenge and find workable, more power-efficient alternatives to traditional air-cooling. Tech giants, including Intel, Facebook, and Google, have already thrown their weight behind liquid immersion cooling technology, while industry and government are working to make this standard. If all of this has you thinking: this sounds like the future of data centers, then you are on the right track. The next step is to get in touch with GRC, we are, after all, the immersion cooling authority!
