Specifying the appropriate Uninterruptible Power Supply (UPS) for a data center or IT environment is a multi-dimensional task that involves decisions on efficiency, integration into existing environments, availability, flexibility, cybersecurity and beyond.
It’s a decision that must be taken based on the organisation’s overall strategy and modernisation investment.
There are some obvious considerations around the type of UPS that’s suitable – single-phase or three-phase, depending on whether the UPS is destined for computer rooms, network closets and small datacenters or a large data center.
There are also considerations around availability, reliability, form factor, ease-of-integration and environmental footprint, but these are minimum expectations. A UPS must go way beyond these to deliver smarter power management and business value.
Data Center demand is going up, and while data centers must respond accordingly with increases in capacity, the clear trends are not for organisations to open more facilities, but to respond by investing to make the existing infrastructure more efficient.
Up-to-date infrastructure is the key to running a successful, efficient and cost-effective facility and it’s met, largely, with a move towards virtualised and more energy-efficient systems.
Modernisation through replacing older equipment, refurbishing facilities, virtualising environments and automating wherever possible makes it easier to consolidate and reduce operational costs in real estate, energy and manpower.
It’s not just about core IT systems, investing in the supporting facilities infrastructure around power, cooling and racks is important too, both because of virtualisation and changes in equipment density, energy rating and increased resiliency requirements.
Outdated infrastructure and data center management techniques do not only lead to energy inefficiency or a lack of compute capacity, they also create a higher risk of failure and power-related outages.
Virtualisation is Vital
Modern, virtualised IT systems enable organisations to pack more compute power, storage capacity and more into a much smaller footprint, while cloud-based architectures drive up resource utilisation. The investment needed to modernise is usually returned through the efficiencies, reliability and uptime of a modern system, compared to living with the costs and constraints of an older environment.
When considering power management in virtualised environments, the most important aspect is the ability of the UPS to integrate with the existing virtualisation platforms. If it can’t, then the UPS is just a dumb box with batteries.
In virtual environments, more services run on less hardware, which puts more focus on making sure that hardware – and the services that run on it – are always available. VMware, Microsoft, Citrix and other server virtualisation software vendors all enable “live migration” to transfer virtual machines from one host server to another for load balancing purposes or in case of maintenance or outages.
However, none of those systems include built-in functionality for responding to power outages, and typically power protection systems have used their own separate command console, leaving IT managers and data center professionals to use one tool for virtualisation and a separate one for power management, which reduces their ability to monitor infrastructure and can increase the time it takes to respond to any failures.
In contrast, Eaton’s Intelligent Power Manager integrates with the major virtual platforms, enabling IT managers and data center professionals to view and manage their entire power system from their current dashboard. It also triggers live migration and, in the event of an extended power outage, will gracefully shutdown virtual machines and hosts in a cluster.
This is important because server virtualisation makes safeguarding data during power failures significantly more complicated. The challenge is that when a virtualised data center loses power, technicians must shut down not only their physical servers but the virtual machines running on those host servers as well.
Additionally, they must execute the many steps in that process in a specific sequence, often in the face of intense time pressure. For example, they must shut down virtual machines before physical ones, and core devices – such as domain controllers and shared storage arrays – after the servers that depend on them.
Efficient by Design
Alongside rising capacity needs, data centers are also being driven by the twin pressures of rising energy costs and increased awareness of their environmental impact. The challenge facing IT and data center professionals is to reduce energy consumption without affecting the reliability of their servers, data storage equipment or network.
Of course, a UPS should be measured by its value. What it brings to the business is business continuity, resilience and disaster recovery – values that will ensure a business can continue to operate even in the event of prolonged and unexpected power outages. Even so, when the availability that a UPS provides is a given – or at least, it should be – it is often down to the efficiencies and ability to reduce total cost of ownership that can differentiate UPSs.
A modern UPS is designed with a low total cost of ownership in mind and equipped with technologies to increase efficiency and reduce maintenance and servicing costs. It does more than simply reduce energy costs, it gives IT professionals the tools to maximise network runtime, resiliency and prolong its service life.
Eaton has invested significantly in its technologies to ensure that it can offer the most advanced, most reliable and highest efficiencies across its complete UPS portfolio. Eaton’s UPSs have some of the industry’s highest efficiency ratings of up to 99%. Alongside this, UPS power density has improved by greater than 50%, meaning that UPSs can be built to smaller footprints to reduce space requirements, while UPS generated heat losses have reduced by 40% due to lower power losses, reducing cooling requirements.
The inherent efficiencies that can be found in a modern UPS have meant that data center and infrastructure have improved to reach Power Usage Effectiveness (PUE) levels that are hard to better. For the same reasons, UPSs are green by design.
So, attention is now turning to how a UPS’s efficiencies can be turned to deliver more return on investment and drive down the data center’s environmental impact even further.
Renewable energy can be inconsistent, but a modern UPS can help grid operators to overcome some volatility, ensuring the data center contributes to the use of renewable energy and is rewarded for it.
The primary example of this is using a UPS as a reserve to contain a grid’s demand-supply balance, where data centers immediately respond to grid-level power demands to keep frequencies within allowed boundaries to help the grid operators avoid widescale power outages.
This means that data centers are supporting the grid to use renewables for power generation and are also maximising their UPS investments by receiving compensation either for not drawing power, or for offering capacity back to the grid.
Keeping critical workloads operational is a data center’s top priority during power outages, but a modern UPS is expected to deliver much more performance and functionality. The minimum expected is that it provides high availability, is green by design, integrates easily into all physical environments, is future-proofed and optimised to increase its service life. Going beyond this is what separates a smart power management system from a dumb box.
Of all the trends in data center power, virtualisation and efficiency are the most impactful in terms of increasing runtime and lowering total costs of ownership.
Companies should view advanced power management systems as an essential component of any well-designed virtualisation environment, while seeking out the features that will reduce lifetime costs and enable the UPS to provide more value back to the business.
Jaco du Plooy is Eaton’s Power Quality Product Manager for Africa.