Containers, cloud, fast networking and IoT still top IT trends

From containers and faster networking to IoT and edge computing, these trends continue to define the role of IT pros in future data centers and the skill sets they need to hone.

ORLANDO, Fla. -- IT professionals have a tall order to fill. They must keep applications running as they evaluate new technology in preparation for whatever the future brings.

At Gartner's IT Operations Strategies and Solutions Summit here this week, Gartner analyst David Cappuccio summarized the top IT trends of the year. While not necessarily new, these 10 trends have an increasing impact on IT teams and the data center.

Data centers are disappearing. Traditional, on-premises data centers have faded, as companies provision business workloads using more efficient and cost-effective alternatives. Today, about 80% of business workloads run on premises, while the remaining 20% are in a cloud or an outsourced facility, Cappuccio said. By 2021, local workloads drop to about 20% to 25%, and the majority of workloads will run in the cloud or otherwise off premises.

Data centers will be interconnected. Connectivity is a hallmark of modern data center technology, but the growth of cloud and other off-premises facilities will spawn even tighter levels of connectivity using fiber links for high-speed, long-distance connectivity. Future interconnects will merge data centers and resources in a manner that more closely resembles peer-to-peer networks. The idea puts public cloud and software-as-a-service (SaaS) provider connectivity performance and resilience on par with on-premises private compute and storage resources, and it provides flexibility and agility on a grander scale.

Containers change applications. Containers continue to be one of the top IT trends in the enterprise. This is because they enable new forms of application development and deployment that would be impossible with even virtualized data centers. While this offers fertile ground for developers, IT professionals must implement and support containers, as well as address new management issues related to their use, such as governance, software licensing and so on. Containers can have a life span of milliseconds, which requires a solid infrastructure to create, monitor, manage and destroy containers, while organizing and tracking all of the resources used to operate them.

IT drives the business. Business leaders understand the need for IT and the role of IT services in the business, but the days of in-house IT dictating the applications, services and other technologies used by business units are over. The rapid growth of public cloud, SaaS and other competitive services has put traditional IT models on the defensive. There is no longer a need to wait for IT to assess, recommend or deploy services and workloads when effective options can be engaged faster, and with less hassle, within lines of business. Cappuccio said more than 20% of business IT spend is currently outside of IT control, which forces IT to innovate and compete to keep its seat at the table. Part of the answer is for IT to rethink its role as a service enabler and broker, rather than a sole source of enterprise IT.

Data centers will increasingly be a service. The notion of data center as a service (DCaaS) has gained traction, as businesses refocus away from infrastructure toward applications. But the choice of cloud, outsourcing, SaaS, PaaS, human resources as a service and even DCaaS is not an all-or-nothing proposition. One delivery model will probably not support all of an organization's needs, and teams will likely employ multiple models to meet different goals.

Stranded capacity must be identified and reduced. Stranded or orphaned capacity is certainly not new -- consider VM sprawl -- but the speed and agility offered by modern computing models make sprawl problems much more pronounced and potentially expensive. Twenty-eight percent of compute instances are now ghost servers -- instances that run and consume resources, but deliver no value to the business, Cappuccio said. Forty percent of racks are underprovisioned, servers are performing at just 32%, and 40% of storage provisioned is more than required. This all translates to wasted resources that cost the organization money.

IoT is still emerging. The internet of things (IoT) has quickly become one of the top IT trends, able to collect and transform tremendous amounts of data into unmatched insights for sales, process control and analysis. But the actual devices and software stack used in IoT deployments are still being defined. IoT devices can provide fast time to market, with prebuilt functionality, high scalability, vendor support and a reasonably complete vendor ecosystem. But IoT remains fraught with challenges due to its overall expense to install, product immaturity and a lack of standards overall.

Remote devices must be managed. Where IoT poses challenges for selection, deployment and storage, it also requires IT teams to manage remote devices -- all of those "things" -- which can pose a serious logistical challenge. IT staff must install and configure, operate and maintain, and then replace or retire every remote device in the environment. This might include tracking details, like battery replacement dates, IP addresses and even physical locations. Even device audits might be required periodically.

Micro and edge computing deserves attention. Instead of investing in more expensive network bandwidth, many organizations deploy small-scale compute and storage resources closer to the location of the actual work -- a strategy known as edge computing.

For example, a remote industrial site that compiles varied sensor data could deploy a small-scale computing infrastructure at or near the site to collect and even preprocess the data. It then performs some preliminary normalization and analytics before sending results to the main data center. Still, such remote or edge computing efforts pose other problems, such as resilience and maintenance. Organizations must consider the requirements and vulnerabilities of a remote computing platform.

Look ahead to new IT roles. These top IT trends, and the changing relationship between business and IT, have spawned new IT roles. While a single organization probably won't realize all of these new roles, chances are that one or more will gain acceptance over the next few years.

  • An IoT architect will select and implement the components of an IoT system to ensure proper management and integration with other data center infrastructure and software.
  • A cloud sprawl specialist will monitor resource usage in the public and private cloud to right-size resources and implement and enforce policies that minimize waste.
  • A strategy architect will review existing infrastructure and offer guidance to optimize infrastructure that best accommodates business goals, and identify the technologies needed to achieve them.
  • A capacity and resource specialist will analyze and forecast resources for the business to identify and use cost-effective alternative resources when possible.
  • A broker and provider specialist will identify business services from a wealth of potential providers to guide the business to cost-effective services that can integrate with other business platforms.
  • And, finally, a performance and end-to-end specialist will focus on application performance management to ensure workloads operate at peak effectiveness and user satisfaction.

Stephen J. Bigelow is the senior technology editor of the Data Center and Virtualization Media Group. He can be reached at [email protected] and @DatacenterTT.

Next Steps

Cloud-style pricing comes to the on-prem data center

When to pursue an IoT professional certification

More IT trends for data center teams to track in 2017

Dig Deeper on Emerging IT workload types