By Christopher Brown
From the April 2019 Issue
The data center industry continues to grow and change at a significant pace. Large build-outs are causing strain on the supply chain, we are seeing continuation of skill shortages and hiring issues, and there are innovations in power, networking and resiliency with hybrid infrastructures that also cause challenges.
The top concerns for operators are availability and security, which are compounded by increasing interdependencies and complexity across ecosystems. The key to solving these issues is the balancing of risk mitigation against cost controls.
What should be expected in 2019 and beyond, and which innovations will make a difference? Where are problems and challenges likely to emerge? Uptime Institute Intelligence recently identified 10 areas worth paying attention to when planning and operating data center facilities.
1. Big cloud builds push the ecosystem to its limits.
The accelerating demands of big cloud operators for more data center capacity is straining the ecosystem of suppliers, builders, operators, and power companies. We expect operators and suppliers to focus more on standardization, including data center and equipment designs, build approaches, and incremental power requirements. Meanwhile, the bigger suppliers are pushing for earlier and better visibility into the hyperscalers’ expected future demand.
2. Worried governments step up oversight and regulation.
Around the world, governments are becoming more concerned with the profits and power of large IT companies, and societies’ dependency on invisible infrastructure. The overriding result of this concern is additional government vigilance, oversight, regulations, and taxes. Expect that new taxes, including those levied on digitally fulfilled or ordered goods and services, will be introduced with the goal of protecting local business and sharing national infrastructure. Expect too, that large US-based tech giants and low-tax nations that currently benefit from the status will protest that their investments are hugely beneficial regardless of direct taxes, energy use, or other negative impacts.
3. The transition to distributed resiliency will not be smooth.
Disruptive and often high-profile outages will continue as operators seek to support more applications and services on a low-cost base, grappling with the complexities of deploying distributed, hybrid systems across multiple data centers and services. The IT industry is in the middle of a large and difficult transition—from single, secure, and tightly managed data centers to a network of distributed, dynamically interconnected systems deploying clouds, microservices, and software-defined networks. Evidence suggests the best way to ensure resiliency is to combine site and network level redundancy with distributed IT and resilient architectures using cloud technologies.
4. Edge data center deployment.
The promised explosion in demand for small, edge data centers is indeed coming, but issues with security, costs, business models, integration, networking, and 5G rollout will hold back large-scale deployments.
The data generated is driving demand for data center of all types:
- Those nearby (“local edge” or “edge”) for first-line processing, analysis, and routing
- Those within a local area (“near” or “regional” edge) to connect, integrate, and re-route
- Those far away—such as economical hyperscale facilities, (“core”) for further processing, analysis, and archiving
5. Connectivity is king, and operators work to build the fabric.
Demand for fast and secure network connections to both trading partners and cloud operators continues to grow. Big data center operators and suppliers of software-defined networking fabrics are working to establish themselves as essential providers in a software-driven, distributed world.
6. Skills shortage will force new strategies.
Even with automation and artificial intelligence (AI), the data center sector’s staff shortages are set to intensify. To keep pace with demand today and to avoid a shortfall tomorrow, data center operators (especially the biggest ones) will begin to focus on diversifying the talent pool, with new initiatives, hiring strategies, and workforce training.
7. Climate change forces fresh review of resiliency planning.
The risks associated with climate change may be more varied and extensive than IT planning had previously anticipated. Without reassessment resilience plans, major service failures will be more likely. Rising seas, higher and frequent floodwaters, more violent storms, and other effects caused by climate change may pose an unexpected and potentially large challenge to data center operators. According to Uptime Institute, even where a facility rides out the challenges unharmed, damage to local infrastructure can deprive a facility of staff, utilities and fuel, and access to telecommunications.
Data center operators can mitigate these consequences with frequent facility availability assessments, with emphasis on how flooding or drought, high winds, and warmer temperatures would affect the facility’s availability to cope with new extremes. Assessments should involve local government, utilities, and telecommunications companies with information about local disaster recovery plans and with specific actionable items.
8. Economics will drive acceptance of data center AI, eventually.
AI-based approaches to analyzing data center risk and efficiencies, including via new cloud services, will be proven at scale, driving mainstream acceptance and, over time, high levels of adoption.
9. Growing threats will necessitate new “zero-trust” approaches.
Security vulnerabilities to corporate IT now encompass mission-critical facilities. Organizations will increasingly realize they can take nothing on trust and will adopt more stringent policies affecting all data center equipment, services, contractors, suppliers, and staff.
10. Programmable power unlocks new efficiencies and agility.
A combination of lithium-ion batteries and other new forms of energy storage, along with configurable, intelligent management tools, means data center operators have a new set of levers to help improve data center performance.
As data centers get larger and larger, the cost of power distribution equipment and the power itself become ever greater. Reducing these costs, especially in an environment where cloud operators are trying to drive them down, is becoming a management priority. A big challenge for operators is to cut power costs without increasing the risk.
Brown is chief technology officer of Uptime Institute, the global data center advisory that works with stakeholders responsible for IT service availability through industry standards, education, peer-to-peer networking, consulting, and award programs. Brown’s 14+ year career in critical facilities includes analyzing requirements, retrofitting existing systems, performing capacity planning, and new facilities design. With more than 11 years in data center design and operation, he also has experience in engineering design, equipment selection, installation, and troubleshooting and is a member of the Uptime Institute Intelligence team.
Do you have a comment? Share your thoughts in the Comments section below or send an e-mail to the Editor at [email protected]