3 Next-Generation Data Center Solutions to Watch
Flash storage, liquid cooling and AI-enabled management give data center operators powerful tools to maximize uptime, improve performance and enhance efficiency.
- by Tim Kridel
- Independent Analyst and Freelance Writer | June 13, 2019
In the era of digital transformation, data is the lifeblood of business. This puts the data center at the heart of it all. A trio of technologies — flash storage, liquid cooling and artificial intelligence (AI) — are set to help data center managers keep that heart pumping strong.
Flash Maximizes Performance and Simplifies Management
The price of flash storage solutions has decreased significantly in recent years, while demand has increased for the high performance, energy efficiency and uptime that solid-state storage provides. Flash also offers peace of mind by enabling frequent, fast backups, which can improve disaster recovery and minimize the potential impact of ransomware.
“We are seeing a rapid shift to all-flash storage,” says Kevin Wittmer, senior director of product management for Dell EMC Storage. “A best practice is to choose an all-flash array that provides easy-to-use, native features for importing applications off of old technology seamlessly with minimal impact to the applications.”
Simplified management is another hallmark of all-flash storage. The performance of networking and computing solutions has historically exceeded disk performance, so many data center operators pooled dozens or hundreds of disk drives to keep up with workloads, creating management challenges.
“This required storage admins to often move workloads around during off hours or required complex, Tetris-like layout schemes to meet performance requirements,” says Chris Gibes, a CDW data center solutions technology practice manager. “A single SSD has the same performance as a rack’s worth of spinning disks. By adopting an-all flash array, the manual and time-consuming job of managing storage for performance is mostly eliminated.”
Evolving Workloads Increase the Need for Powerful Cooling
Advanced IT capabilities are placing new demands on data centers. Among the challenges this creates for data center operators is a greater need for cooling as high-density systems run hotter than older solutions.
“AI and machine learning workloads running on graphics processing units [GPUs] are becoming big concerns because of the level of disruption they present to traditional data centers,” says Jennifer Cooke, research director for IDC's Cloud to Edge Datacenter Trends and Strategies team.
A data center that can’t stand the heat can undermine an organization’s bottom line and lead to outages that result in lost business and frustrated customers.
“Today, one in five enterprise data centers experiences downtime due to HVAC outages or hotspots, and 22 percent experience delays in IT deployments due to power or space constraints,” Cooke says.
Traditional cooling systems often struggle to keep up.
This has increased interest in liquid cooling: Water is more effective than air for moving large amounts of heat.
Data center operators have multiple options for implementing liquid cooling. To support racks with increased thermal densities, rear door or in-row heat exchangers can be used to apply liquid cooling to the rack, while inside the rack, the systems are still cooled by air.
“But to cool more efficiently and gain power savings, direct-contact liquid cooling can be used to directly cool processors, GPUs, etc. with liquid,” says Ed Turkel, a senior strategist with Dell EMC High Performance Computing. “This can eliminate the power consumption of the fans moving air around the servers. The power savings can be significant.”
Data Center Services Can Simplify Complicated Challenges
As data centers become more complex, IT teams are challenged to maintain a variety of infrastructure components. For many organizations, a trusted service provider can deliver valuable assistance and enable in-house staff to focus on more business-oriented tasks. Among the areas where a service provider can help:
Storage: Organizations strive to unify their storage environments to improve visibility, enhance performance, boost reliability and scale to meet new demands. Numerous services can help IT teams achieve these goals. A storage review can identify the best solution to meet an organization’s needs, while a data migration assessment supports the development of a plan to move data to a new storage system. Managed storage services ensure that monitoring, upgrades, maintenance, reporting and incident management of storage solutions are taken care of without placing demands on IT staff. Configuration services ensure that storage solutions are custom-configured to an organization’s needs and specifications.
Power and cooling: Data center operators need solutions that protect equipment, increase uptime, maximize efficiency and save money. A power and cooling assessment helps IT leaders to see where power may be wasted as well as how to best address current needs. Other important services include planning, design, installation, assembly and configuration.
AI Optimizes Performance, Uptime and Efficiency
Applications that use artificial intelligence are placing greater demands on data centers, but AI can also enable automation that helps organizations better manage and optimize their data centers, including by reducing energy consumption.
“We see a great interest in creating what we call ‘smarter data centers’ that use sensors and connected equipment (such as intelligent power distribution units and AI-driven cooling) to optimize the environment,” Cooke says. “Data centers have been gathering massive amounts of data for a very long time, but until recently haven’t really turned it into actionable data to optimize the use of energy and cooling.”
AI also can be used to maximize uptime. For example, AI can ferret out problems faster for quicker resolution, and conduct trend analyses to identify upgrades that can help organizations stay ahead of workload demands. AI also can optimize configuration and enable better workload execution with dynamic settings and adaptive capabilities.
But achieving these benefits requires expertise.
“Many of our customers are starting their initial deployments of AI technology by targeting relatively simple workloads, to gain familiarity with the tools by their internal staff,” Turkel says. “Others are hiring AI consultants, or full-time staff, to help, especially with the data science part of implementing AI solutions.”