Seraphim Vector - Fotolia
ORLANDO -- Every IT professional faces daily operational challenges, but the opportunity to look ahead and consider the many trends taking place in the IT industry allows architects and administrators to meet those challenges most effectively.
Ten strategic, tactical and organizations trends will shape IT for the next five years, shared David Cappuccio, vice president and distinguished analyst at Gartner, speaking to a packed ballroom here at the Gartner Symposium/ITxpo.
1. Data centers are disappearing -- but not entirely
IT professionals have long known that on-premises data centers are disappearing in favor of off-premise resources. By 2020, only 20% of enterprise workloads will remain on premises, Cappuccio predicted. Those workloads aren't disappearing but instead moving to outsourced and cloud facilities that may ultimately number into the dozens.
Although this hybrid deployment model offers greater agility for IT and the enterprise, it also poses new challenges for IT professionals. "As workloads move off premise, our lives are not getting easier," Cappuccio said, noting that administrators and CIOs must still manage and monitor key performance indicators (KPIs) within this increasingly complex web of diverse providers.
Implementing and expanding mixed environments demands copious staff training. Workload migrations between on-premises and off-premises facilities -- even between facilities -- can be difficult and expensive. Idled resources, such as servers and storage, should be shut down or repurposed; this is an often overlooked part of the migration process. And service brokers have a new prominence in the IT process, ensuring continuity and consistency between providers, KPIs and service-level agreements.
2. Interconnected fabrics are expanding
A typical resilient data center uses fiber to peer discrete assets together into a multitenant fabric, and evolving interconnect fabrics are building on this idea to also peer remote sites, Cappuccio noted. The goal is familiar: ensuring the availability of applications and the improvement of customer or user experiences, an idea dubbed "service continuity" rather than the more traditional considerations of business continuity. Direct connections are also appearing between providers, including Amazon Web Services, Microsoft Azure and Google Cloud Platform, to boost workload performance and lower latency.
The idea of IT infrastructure "is not just on prem, but [involves] all services provided to customers" Cappuccio said. This must extend to centralized management, allowing IT professionals to operate, orchestrate and automate many processes. This should improve workload and traffic patterns for enterprise applications, while supporting centralized management over serves, storage, networks, data centers and providers' resources.
Ideally, "services are delivered from the right place, for the right price, from the right platform," Cappuccio said.
3. Containers will continue to grow
One of the most prominent trends in IT, containerization supports new forms of agile application development and compartmentalization into microservices that can be deployed and scaled on physical systems or VMs. Containers will become a primary tool for application developers, Cappuccio suggested, but IT must provide the infrastructure and operations support to run containers. It's an important challenge because containers represent highly scalable and transient entities. Consider that a physical server may last up to five years, and a virtual machine might run for weeks or months -- a typical container may last only milliseconds. Containers will demand detailed levels of control over workload management, such as orchestration and automation.
4. IT is increasingly business-driven
IT exists to solve business problems, but IT professionals don't always have the best picture of the business problems that need solving. Business leaders can't wait for traditional IT processes to play out, and are increasingly leveraging outside IT services without on-premises IT support: a full 29% of IT spend is made outside of the IT organization, Cappuccio noted. This trend is expected to increase as cloud use proliferates and IT costs escalate. IT cannot stop this trend -- nor should it -- but instead can provide services to support and guide business users in their use of IT for the business.
5. Think of the data center as a service
The goal of IT is to provide applications that the business needs -- not infrastructure. We've seen the changing role of IT as a service provider, and Cappuccio thinks this trend in IT will continue: "I look at data centers as a provider of services to the business," he said.
The concept of data center as a service means selecting on- or off-premises resources that can best meet the business' needs for factors including compliance, security, recovery time objectives, workload latency, service continuity, performance, data protection/recovery and even business reputation. Eventually, businesses will be able to select from any number of potential providers depending on these and other future considerations, allowing the delivery of business services, such as platform as a service as well as human resources (HR), payroll and other corporate functions as a service, to users from anywhere.
6. Identify and deal with stranded capacity
IT still has the unfortunate habit of wasting costly resources, and that waste is often traced to four principal culprits. Cappuccio pointed to statistics showing that 28% of IT is ghost servers: systems that are running properly but aren't doing any meaningful or quantifiable work. Forty percent of racks are under-provisioned, leaving unused rack space that limits consolidation and wastes physical space -- not to mention power and cooling. Part of the reason for such waste is the ownership idea -- that a rack is owned by the development team or the HR department -- and an overall lack of control.
"We must put more governance in place to understand what's running and why," Cappuccio said.
Cappuccio listed several ways that IT leaders can drive cost optimizations by reining in stranded or wasted capacity: right-size resources (provision only the resources that a workload needs); implement a comprehensive tagging system to track and monitor workload lifecycles; limit data egress to prevent critical data from being unnecessarily duplicated in multiple locations; reclaim or repurpose orphaned resources; throttle poorly utilized resources so that less active workloads get less resources such as network bandwidth; choose a pricing model that makes the most sense for the particular business; and consider free or open source tools for monitoring and management.
7. IoT is growing
The internet of things (IoT) promises sensors and information sources for countless tasks, but with 20 billion devices connected to the internet by 2020, the infrastructure needed to support IoT must be evaluated carefully.
"IT must start thinking about an infrastructure to support IoT," Cappuccio said, noting that failure to plan now may create even bigger headaches in coming years. Factors including network complexities, competing (often vendor-specific) platforms or specifications, differing protocol standards and competing ecosystems can add to the complexity of IoT implementation.
The resource demands for IoT can also be startling. Consider a police force with 100 body cameras recording video and audio for three shifts per day. How much storage is required, and how long must that data be stored? How will it be secured? These are just a few of the potential issues that a business must think about before even attempting to support IoT.
8. IoT devices must be managed
The notion and benefits of IoT can be compelling, but the proliferation of new devices presents important new management challenges for businesses. IoT devices must be installed, registered, calibrated and tested. IoT devices must also be maintained -- this involves tasks ranging from battery replacement, firmware updates, and reboots to audits and status monitoring. And finally, IoT devices must eventually be retired from service where they are removed and disposed of in a responsible manner.
Although IoT management may be embraced as part of other infrastructure initiatives, such as network bandwidth upgrades and edge computing, it's the scale that is truly challenging. Considering the potential for many billions of connected devices in just the next few years, "going forward, this can be a major headache, so we need to get ahead of it," Cappuccio said.
9. Computing is moving to the edge
The familiar centralized data center paradigm is changing to accommodate the evolution of specific computing demands. For example, workloads may be moved closer to the user base to reduce latency and improve workload performance. A single central workload location may not service every user adequately. In other situations, the proliferation of IoT devices promises to unleash torrents of data that are difficult to pass back to a central location while people wait for processing and responses.
In these cases, it may be desirable to move computing resources closer to the users or the data -- often through the deployment of edge or microcomputing sites, Cappuccio said. For IT architects and planners, this requires a better understanding of workload distribution, synchronization, and performance/user management.
10. Look for new roles in IT
Finally, the evolving trends in IT are provoking an assortment of new roles that IT professionals should evaluate in the coming years. Cappuccio highlighted six emerging roles that may appear in the years ahead.
Roles such as IoT architect provide specialization in the networking, computing and management needs of IoT devices for the enterprise. Cloud sprawl management may become a more prominent role as per-use or per-hour billing models can precipitate unexpected costs for forgotten or unused cloud resources. Strategy architect roles can help organizations realize the best approaches to meet specific business computing goals.
Capacity optimization and recovery roles, closely aligned to sprawl management roles, help match provisioned resources to requirements, reducing wasted (over-provisioned) resources. Broker roles are growing as organizations expand their stable of various service providers, allowing businesses to determine all of the available services, their performance and their cost to the business. And performance and end-to-end management roles reflect the growing importance of workload performance and user satisfaction management in the enterprise. Knowing that each aspect of an application is running well can offer early warning for potential problems, as well as insight for improvement.
The direction of data centers and IT in the next five years points to greater agility and performance with the goal of providing superior services to the business. But success will require much closer attention to a complex hybrid infrastructure, and copious management of both on-premises services and an ever-growing array of off-premises services.
The future lies in abstraction for IT resources
IT skills to pay the bills
What cloud computing means for job titles