Life is full of paradoxes such as:
- How am I supposed to gain experience [to be hired for a job] if I’m constantly turned down for not having any?” – Joseph Heller, Catch-22.
- People ask questions, but they don’t want answers. Why? Because most people just want to hear:
- Confirmation of what they already believe
- Something that makes them feel good
- Validation of their opinions
- The only thing we learn from history is that we learn nothing from history.” – Georg Wilhelm Friedrich Hegel
Well it turns out that there are several Data Center Paradoxes as well.
All Workloads Are Considered Equal
The IT team will tell you that this is NOT the case, but the mission-critical facility/power team assumes that it is. Why don’t these two groups have more data points to sort this out? What’s even worse, is that the application software has extremely high levels of resiliency built in (as do virtualization, containers, Big Data and distributed applications) that was not there ten or more years ago.
With these higher levels of resiliency built in, the software expects the hardware to fail and the power to go out – AND STILL KEEP RUNNING! But since IT and mission critical facility managers don’t speak the same language, the power infrastructure gets built the exact same way for a “batch workload” or non-production workload as it does for the production, order processing, revenue generating, mission critical workload. Yikes – talk about throwing away money! Well, here’s another paradox.
The Cloud Paradox: Less Energy Used, More Energy Wasted, Known as Jevon’s Paradox
This concept demonstrates that, as energy efficiency increases, so does the consumption of resources due to the demand spurred by lower costs.
A 2017 Greenpeace report illustrated this paradox with the following example: “With the marginal cost of memory, CPU and bandwidth are tending toward zero, with the resulting proliferation of so many ‘free’ and ‘unlimited’ online services like YouTube and Netflix, the impressive efficiency gains achieved in delivering cloud-based computing services would appear to already be serving to significantly increase overall data center power consumption.” The report points out that, by 2020, video streaming traffic will represent over 80 percent of total consumer internet traffic.
While this may not be an issue in and of itself, if somehow energy efficiencies gained in the cloud can make up for electricity used by consumers, the rate of the cloud’s growth could potentially be problematic. If the cloud grows faster than sustainable energy systems that replace fossil fuels, the cloud may receive more of its electricity from dirty sources.
The study highlights the fact that there are parts of the world still heavily reliant on coal and other fossil fuels, but cloud and colocation companies continue to build significant data center projects in these locations: “While important progress has been made in driving renewable energy investment in several markets, the dramatic increase in the number of data centers in markets such as Virginia, dominated by utilities that have little to no renewable energy, is driving a similarly dramatic increase in the consumption of coal and natural gas.”
Enter Software Defined Power – Optimize Power Infrastructure by Workload
The big cloud providers, Amazon Web Services, Microsoft Azure, and Google are growing at an extraordinary pace (40% to 80% year over year) because they are all about “pay for what you use/need” by workload. Yes, that is right, they have optimized every IT asset, CPU, memory, storage, networking/latency, GPUs, apps, operating systems all by a “pay by the drink” model.
Guess what – enterprises and colocation providers can now do the same thing with Virtual Power Systems’ Intelligent Control of Energy (ICE) Software Defined Power (SDP) platform.
SDP is a software-based power control platform that uses machine learning and big data analytics to provide awareness and recommendations, including how, when and where power can be optimized by data center, row, rack, workload, or tenant. Furthermore, when SDP is combined with hardware, the union automates data center operations and power management by workload. The result increases utilization of installed capacity by removing unnecessary buffers and automating service level agreements (SLAs) while mitigating risks and improving workload uptime. SDP can also increase data center resiliency with our industry first Power Aware Workload Orchestration.
For example, with SDP, enterprise greenfield data centers that originally were designed at 20 MW now can be designed for 14 MW. With capex at $10 million per MW, that’s $70 million in savings just on the original design! For enterprise brownfield facilities with 50% to 80% of their power stranded and unusable, SDP enables reuse of 30% to 60% of the power. See our ROI for an Enterprise Client to understand how SDP can recapture 352 kW to reuse out of 1 MW provisioned.
For colocation firms, SDP offers many advantages including the ability to increase utilization of power distribution infrastructure without accessing the racks as well as monitor and automate SLAs and manage risks. See our ROI for a Colo – this is a real game-changer for this industry segment.
We live in a transformative time and yet the axiom “if it ain’t broke, don’t fix it” too often prevails in the data center environment – especially as it relates to maximizing uptime. However, taking a different approach in better managing power utilization can help to achieve the Holy-Grail of increasing uptime while advancing power efficiencies, while substantially lowering costs and reducing a facilities’ carbon footprint. No paradox at all.