by Troy Cogburn, Senior Research Engineer, Trace3
The data center is dead…or so claimed experts as recently as two years ago, crowning the cloud champion and king of the modern IT environment. However, these prognostications haven’t proven completely true. While cloud providers have certainly enjoyed great success, data centers also remain key to many environments. Just not in the forms previously accepted. As organizations continue to back source their applications from the cloud and push towards hybrid cloud models, and technologies such as Artificial Intelligence (AI) and Internet of Things (IoT) altering data processing, the traditional datacenter has been forced to adapt. Here are some of the trends we see shaping that future.
Artificial Intelligence Impact on Data Centers
As companies incorporate technologies such as machine learning, deep learning, and computer vision, the underlying infrastructure must change. Most applications can be optimized by running on GPU arrays and specialized chips, and while this presents great opportunity to condense and optimize the data center, the majority of their racks are not ready to handle the required power consumption and cooling necessity. As a result, the idea of having liquid cooled systems has seen a resurgence, since liquid is a much better conductor of heat than air. In 2016, Google used their deep mind project to optimize data center power efficiency and saw a utility cost reduction of 30%. Although there was significant impact, Google stated AI isn’t ready to handle a data center’s cooling and heating all on its own. As such, data centers will have to upgrade to handle the demand for AI and high-performance compute.
Cisco estimates by the end of 2019, IoT devices will generate more than 500 zettabytes of data per year and most enterprises aren’t equipped to handle that amount of bandwidth. Not only will it become difficult to backhaul this data to a cloud or datacenter, many applications will also require lower latency for processing. This creates the need for not only processing at the edge but also smaller purpose-built data centers closer to the edge to assist further. Several companies, such as VaporIO and Edgemicro, have emerged to help by creating edge data centers that sit at the base of radio towers.
Hybrid Cloud and On-Prem AWS and Azure
While the public cloud has brought benefits like convenience, low upfront costs, and minimal maintenance, many companies have discovered not every application is best suited on a public cloud. Often, these applications are not even architected to be cloud-ready. As a result, customers look to stay in-house, combining the benefits of the cloud with the data locality and security of an on-premise solution. The big public cloud providers have taken note and released supporting technologies, such as Azure Stack and AWS Outpost, to enable cloud services in corporate controlled environments.
Data Center Automation (Software Defined Data Center)
The Software Defined Data Center (SDDC) is the sweet spot for businesses looking to capitalize on the agility, elasticity, and scalability of the cloud, maintaining full data control, while also avoiding the public/hybrid cloud and vendor lock-in. However, adoption of this solution is a complex process. Most companies are familiar with virtualizing compute but to achieve an SDDC, all resource pools must be abstracted, including virtualizing network and storage. Although these technologies are software defined, a lot of underlying infrastructure must change to accommodate the application of the virtualized components. Further, new skillsets must be developed to set up and maintain this technology. Another overlooked aspect of SDDC is the approach to securing the environment. Traditional physical security techniques will no longer be applicable, with focus shifted to access distribution.
Climate Change and Sustainability Impact
As appetites grow for digital services, data center energy consumption increases in turn. With this, however, customer concern over a data center’s environmental footprint becomes paramount. Many organizations have reduced their footprints by focusing on the climate surrounding their data centers. Locations with cooler temperatures and proximity to clean energy sources, such as hydro or wind energy, are becoming hotbeds for datacenter development. The Nordic region recently became a data center home to likes of Facebook, Amazon, and Google. Other companies such as Microsoft have even explored the possibility of building underwater, capitalizing on the proximity to wind, solar, and ocean currents. Further, as the effects of climate change continue to be seen and felt in higher average temperatures and increased natural disaster frequency, disaster prevention has taken an even higher precedence. In addition, automation and remote monitoring are being built into many new data centers to allow for continued system function in times of disaster.
The data center will remain a big part of enterprise strategy for years to come, even if cloud providers continue to see increased adoption. With companies becoming more agile, efficient, and secure, and applications supporting Machine learning and Internet of Things, the data center will evolve to keep pace, providing customers a breadth of options to envision the modern IT environment.
To solve the challenges of tomorrow, business leaders need access to critical insights today. Trace3 Research provides just that, identifying emerging technologies, market trends, and the collective forces behind it all. Through this analysis, and our unique relationship with the venture capital and startup community, we enable our clients to best align their business objectives and technology strategy to generate breakthrough results.
Learn more at https://trace3.com/research/