New technologies can shake up the data center world – but it will happen gradually
Businesses tend to innovate by applying a new idea or technology to generate value, so innovation is therefore linked to growth. Data centers innovate to deliver business operations reliably and resiliently while expanding to handle growing volumes of increasingly complex data.
Bandwidth-hungry internet traffic such as HD and 3D video, more sophisticated connected devices, and trends such as Big Data and the Internet of Things are driving the innovation agenda, and facility and IT administers are working with scarce, finite and expensive resources such as space, power and skills. Despite all this, data center innovation is surprisingly gradual, incremental and piecemeal.
Source: Thinkstock / MauMyHaT
Barriers to innovation
Why is this? Limited operational and capital expenditure (opex and capex) budgets are acting as constraints and barriers to innovation, causing managers to defer investments that would otherwise enhance data center innovation. Other barriers to innovation include institutional inertia, created by cultural or organizational factors within the enterprise itself. Many of the barriers to innovation are reflected in the external factors that data center operators predict will impact on their operations (page 20).
Traditional data centers are designed for a lifecycle of 10 to 15 years. This creates practical challenges, such as the expense and difficulty of expanding or upgrading a data center. It also has cultural implications, as traditional skillsets and legacy ways of thinking may dominate decision-making.
Given all this, it is perhaps not surprising that, according to DCD’s Global Census survey, the pace of industry change continues to be one of gradual, incremental progress. For example, server virtualization and outsourcing are major forms of innovation that have been available for several years, pursued by the industry for some time now. Yet growth in these areas remains slow but steady.
In 2015, 17 percent of total global data center racks were located in-house but were virtualized or managed as part of a private cloud. This figure was up from 15.6 percent in 2014, reflecting an increase of just 1.4 percent. Meanwhile, 10.4 percent of total global racks were outsourced to cloud-based service providers. This figure was up from 9.6 percent in 2014, reflecting an increase of just 0.8 percent.
“Everyone” is doing it?
Media exposure of new technologies can sometimes give the impression that everyone is doing X or Y, but in fact the DCD Census reveals that whatever new technology you choose to look at, there will be no more than a quarter to a third of data center operators investing in that technology in a given year. Once again, this points to the cyclical nature of data center investment and the need to stagger investment in costly technologies over longer periods.
Innovation is slow and steady, while legacy equipment and processes linger alongside new technologies. But despite this, most data center operators are convinced about the benefits of innovation, acknowledge that it can reduce their total cost of ownership as well as increase agility and flexibility, while satisfying customers better.
The DCD Census reveals that whatever new technology you choose to look at, there will be no more than a quarter to a third of data center operators investing in that technology in a given year.
Data center innovation can address a spectrum of things, including facility design and construction, the use of virtualization, automation and monitoring in IT, the adoption of business models such as outsourcing, and the application of specialist human skillsets (see box).
These innovations will be adopted in different ways, and at different speeds, in various geographical markets and market sectors. For example, data center sectors in so-called emerging markets have picked up new technology differently from how it has been adopted in long-established markets, sometimes picking up new ideas at an earlier stage, “leapfrogging” over the traditional market. For instance, emerging markets have adopted modular technologies and outsourced service models in the early stages of data center growth, instead of evolving through the traditional in-house model that dominated in more established markets.
Despite geographical variations, the data center industry has been converging, and the gap between “established” and emerging markets has been getting narrower. This is partly due to the international transfer of technologies and strategies – a development that is contributing to the steady globalization of data center innovation.
tortoise color thinkstock left
Source: Thinkstock / MauMyHaT
This is especially true of energy-efficiency initiatives within IT and data centers, which are no longer the preserve of Western data centers and governments, but have instead moved anywhere where there is a critical mass of data centers.
The traditional structures and culture of the data center industry mean that innovation will most likely remain evolutionary, sporadic and incremental. Most new technologies and business models will be adopted steadily across a range of infrastructure and business areas.
New technologies will need to sit alongside legacy infrastructures and processes, producing a hybrid culture with its own set of challenges. Owners and operators will have their hands full, making sure their staffs’ skills and tools are up to date and effective.
But while this happens, many will question whether data center owners and operators are responding fast enough to external pressures, including those associated with Big Data and the Internet of Things. Some organizations will be more successful than others at setting the right pace of innovation.
This article appeared in the November 2015 issue of DatacenterDynamics magazine
Sources of innovation
- Modular data centers Modular and prefabricated data centers look like an answer to a thorny problem: lengthy data center lifecycles conflict with the need for ongoing upgrades. Prefabrication promises faster project times, improved scalability, portability and standardization. Standardized data centers could be a boon for both commissioning and staff training. Modular data center solutions are also being adopted by enterprises looking to boost their green IT credentials.
- Power and cooling Energy-efficient strategies include the use of low-energy servers, or micro servers, which rely on low-power central processor units (CPUs). Others include the use of natural air cooling, which uses outside air to cool data center facilities instead of power-hungry mechanical refrigeration or air-conditioning units. Meanwhile, computational fluid dynamics (CFD) promises a scientific approach, using numerical analysis to model fluid flows.
- Virtualization By allowing consolidation, virtualization is transforming data centers. Virtualization has been applied to servers and applications, allowing faster development, better backups and speedier deployments. New virtualization technologies could improve data centers further, including desktop virtualization, network functions virtualization (NFV) and the virtual storage area network (VSAN) – a software-defined storage solution that will help to increase flexibility and automation.
- New architectures Mega data centers are being driven by cloud and colocation providers needing economies of scale. Cloud services can be delivered from multiple data centers, a trend that will increase as services and applications move to the cloud. Meanwhile, large enterprises will look for better ways to connect multiple data center facilities, trying out options such as Cisco’s overlay transport virtualization (OTV).
- Automation Data center operators are investing in automation, a benefit that has long been promised by technologies such as data center infrastructure management (DCIM). Meanwhile, software-defined networking (SDN) promises to help automate networks. At a higher level, between data centers, open trading exchanges could automate the way in which resources are bought and sold, providing compute, cloud and data center capacity more flexibly.
- Machine Learning Artificial intelligence (AI) and “machine learning” is increasingly being applied to improve data center operations. Google now uses machine learning in proprietary systems that gather data to manage and optimize its IT systems and regulate power and cooling. Commercial implementations include Vigilent’s dynamic cooling management system, which responds intelligently to temperature data collected from wireless sensors.