The Evolution of Data Centers Has Changed the Face of Power Protection
New technologies and the resulting challenges are transforming UPS and power management requirements
In recent years, the IoT, ever-escalating data requirements, and ongoing cloud adoption have contributed to a shift away from traditional, enterprise data center facilities. Instead, many organizations are adopting new approaches — including hybrid IT, edge computing, distributed IT, and micro data centers — all of which afford numerous benefits but, at the same time, create critical challenges.
As data centers continue to evolve, power remains an integral yet often overlooked part of the progression. With so many alterations to the data center landscape, it’s not surprising that there has also been a change in power protection requirements, driven largely by the complexities these new methods have created.
“The biggest issue is that these new technologies are a mix of IT and OT [operational technologies], where people do not speak the same language and do not look at the same things with the same lens,” said Hervé Tardy, vice president and general manager of distributed power infrastructure at Eaton.
Going forward, organizations will need the ability to manage power anywhere — whether in remote locations, at a collocation site, in the cloud, or within an enterprise data center — as well as move workloads dynamically from one location to the other based on power events.
“It will all be about managing power in a holistic way wherever data is being processed,” Tardy said. “This goes way beyond the four walls of the legacy data center.”
Goodbye, traditional data center
The struggle for organizations to maintain explosive amounts of data, the escalating cost and security concerns over the cloud, and the expectation for IT departments to respond quickly to rapidly changing business requirements have all contributed to the formation of new data center models. Meanwhile, popular trends like machine learning and AI have driven the need to keep data as close as possible to the consumer for both latency and sovereignty purposes.
As a result, these changes have propelled many organizations away from enterprise data centers and fueled the adoption of a fresh batch of tactics. One of these is hybrid IT, which allows some IT workloads to be managed on-premises, while others rely on cloud-based services.
Similarly, the market has seen a continued migration toward distributed IT, which enables organizations to leverage colocation facilities, network closets, and nearby remote sites to place applications in closer proximity to users.
Companies are also living on the edge, deploying solutions that generate, collect, and analyze data on the edge of the network rather than in centralized servers and systems.
The appeal of edge computing , which offers the ability to increase network performance by reducing latency, is underscored by the findings of a recent Vertiv study. In it, nearly 500 qualified participants who currently support edge environments predict their number of sites will increase 226% by 2025. Edge applications have become especially attractive to retail, education, banking, health care, and industrial sectors as well as government organizations.
Another option that has gained momentum is the micro data center, a smaller or containerized architecture designed with no more than four servers in a single 19-inch rack.
While the progression of IT systems and interconnected technologies bring a number of benefits, there are also potential risks that must be addressed when relying on these rapidly changing infrastructures. One of the most significant issues is how to approach power and backup to ensure critical data remains safeguarded.
challenges created by new IT trends
It is said that everything comes with a price, and modern technology is no exception. Potentially devastating downtime threats, intensifying security concerns, and essential remote management requirements are forcing organizations to rethink how they will implement a strategic, end-to-end approach to power management in these new environments. Among the chief concerns are:
Maintaining continuous uptime — As downtime threats continue to magnify, organizations that lack adequate power protection leave themselves vulnerable to serious ramifications. IT downtime can wreak havoc on an organization’s bottom line, ongoing business operations, and its reputation. In addition to equipment damage and data loss, in some instances, an unplanned outage can even cause a company to go out of business. Just a momentary blip in power has the propensity to trigger a hefty toll, with Gartner estimating the average cost of IT downtime at $5,600 per minute. It is important to note that costs vary dramatically depending on the organization, extending as high as $9,000 per minute.
Ensuring security amid escalating threats — The proliferation of smart, connected devices has created an unprecedented opportunity for hackers and cyber criminals. Considering that forecasts predict there will be some 75.4 billion connected devices by 2025, cyberattacks represent a huge concern for IT staffs as they migrate networks off-site. And the trepidation is merited — cybercrimes against businesses nearly doubled in five years, according to the World Economic Forum 2018 Global Risks Report. Cyber fears have also dramatically slowed the deployment of the IoT and prompted legislative action. The state of California passed SB-327, a law that requires manufacturers of connected devices to equip them with a “reasonable security feature or features” that protect devices and their information from “unauthorized access, destruction, use, modification or disclosure.” UL also weighed in, issuing UL 2900-2-2 — Outline of Investigation for Software Cybersecurity for Network-Connectable Products, Part 2-2: Particular Requirements for Industrial Control Systems.
Lack of on-site expertise — While organizations may recognize the need to secure power in edge and distributed IT environments, just like an enterprise data center, they face an additional challenge: In all likelihood, these locations will not have an IT pro on-site.
“Nobody will know what to do with the system,” Tardy said.
However, he went on to explain that there will be an expectation for the equipment to remain up and running, not for the typical three-year IT system lifetime but more likely for five to 10 years. As a result, companies have begun to seek products that afford a “set it and forget it” mentality, where solutions can operate with little to no human intervention. The approach has been especially advantageous in the retail and health care sectors.
Power protection strategies evolve too
Regardless of the type of data center strategy an organization chooses, the need for continuous, clean power remains paramount. With so much IT infrastructure now residing locally, premium power protection has never been more critical to preserve both equipment and data.
UPS systems are being tasked with protecting network servers, nodes, workstations, peripherals, and hubs/routers. They are also frequently attached to Voice over Internet Protocol (VoIP), safety/security, and point-of-sale (POS) systems.
Among the changes the power protection market is experiencing are:
An upsurge in single-phase UPS solutions — Edge computing has sparked an increased demand for highly reliable single-phase UPS systems, as organizations require small power systems in close proximity to network switches and other devices that facilitate a connection to cloud.
However, Tardy said the next generation of UPS systems will likely be different than the models we see today, largely due to their need to operate for a much longer time with no trained personnel on-site. In addition, he predicts the next big innovation will be a UPS display in your pocket, where IT professionals can easily and comprehensively control and monitor equipment from a smartphone.
“I can’t believe that we are still putting LCD or LED displays on UPSs,” Tardy said. “Just think, the UPS is the only device inside a rack that still sports an LCD display. No server or router has one.”
Adoption of lithium-ion batteries — Already the preferred choice in many three-phase UPS applications, lithium-ion batteries are poised to become mainstream in the single-phase UPS space as well. Uniquely qualified to handle edge computing load requirements, these batteries offer an eight- to 10-year life cycle compared to traditional valve-regulated lead acid (VRLA) batteries, which generally need to be replaced every three to four years. Not only do they ease maintenance and eliminate replacement requirements, lithium-ion UPS systems offer the desirable “set it and forget it” value proposition especially conducive to edge computing environments. For example, in retail locations where UPS systems are tasked with protecting critical network operations, the vast majority of store employees will not be IT proficient, making them unlikely to recognize the gravity of a blinking LED warning of impending battery failure. Yet another benefit of the extended battery life provided by lithium-ion is that organizations can align their UPS refresh cycles with the rest of IT stack, saving time and money spent on labor and replacement batteries. In fact, a recent Schneider Electric report concluded that some single-phase lithium-ion battery solutions lowered total cost of ownership (TCO) by 53% compared to traditional solutions. Coupled with additional benefits, such as smaller footprint and increasingly competitive pricing, some industry experts predict lithium-ion could ultimately displace VRLA in UPS batteries.
The need for cybersecurity solutions — As IoT expands, server-gateway connections continue to multiply, and the risks associated with cyberattacks grow even more dangerous, organizations must ensure their connected technology is secure and resilient — power is no exception.
With the introduction of SB-327 and UL 2900-2-2, power protection manufacturers have begun to introduce UPS systems, network cards, software, and power distribution units (PDUs) to address these concerns.
“You can’t assume that you are safe if your device is behind the enterprise firewall,” Tardy said. “Every device connected to the network must have its own level of cybersecurity.”
An increased demand for remote capabilities — In network closets and remote locations at the edge — where there is rarely an individual trained in IT and technology — there is a growing need for systems to run seamlessly. An unskilled operator may need to perform simple plug-and-play deployments, deliver secure edge application updates, or debug in the event of a problem.
This is where remote management has become an essential component to an effective power management strategy, eliminating the need for skilled personnel to be on-site to deploy and manage the solution on a daily basis.
“The days of a standalone, dumb UPS are over,” Tardy said. “It is now a connected device that must be remotely managed — just like anything else in a data center or in a remote location.”
The rise in PDU popularity — Going forward, PDUs are expected to be among the smart devices of choice, as they give administrators direct control over power flows to IT hardware. Monitoring with smart PDUs also improves efficiency, uptime, and growth.
“The PDU is actually the smartest power device of all,” Tardy said. “The level of intelligence on the output of each outlet is similar to what you have at the output of a UPS, and you have typically 42 outlets.”
More software-defined power — The rise of software-defined data center technology has led to infrastructure operators demanding solutions that can virtualize, configure, and adjust power services dynamically. The proliferation of sensors, beacons, and other IoT technologies has fueled demand for IoT-enabled power protection products that provide office-grade plug-and-play simplicity for use in remote, largely unsupported environments. IT pros will rely on smart devices that enable personnel to access the power distribution network and provide control functions; capture and analyze a range of data and performance information; and embed intelligence inside other kinds of equipment, allowing for autonomous control.
As the modern data center continues to evolve and the demand for data and faster services continues to magnify, so does the necessity to effectively protect these systems. Continuous, clean power cannot be an afterthought. Yet in order to ensure this requirement, power solutions are simultaneously evolving to meet the needs of today’s IT professionals and systems.