We continue our review of data center trends in part two where part one concluded and explore each technology more in depth. Technological developments are occurring at a rapid pace, policies are struggling to keep up, and budgets are poised for 2018 and beyond – we look toward tech giants to identify where they are focused and where industry will ultimately follow.
AI to Hyperconvergence: Top Data Center Trends of 2018
The deep dive into the technology drivers for 2018 includes:
Blockchain – data centers are not elastic orbs of data warehouses with which to respond in near real-time to data retrieval, users, and increasing data but focus more on disaster recovery preparations and data storage – essentially, it resolves trust between parties for data transmission at the basic level. Blockchain is set to change that with a new, rapid, secure and seamless infrastructure which supports wide bandwidth and traffic spikes of data transmissions as data and users increase in numbers. Most data centers have adopted one of two types of architectures: traditional and the spine and leaf mesh.
Artificial Intelligence (AI) – the science and engineering of making intelligent decisions is the basis for AI. The intelligence part is something is still being defined but can connote the use of algorithms and processes which feed data into, process it, and produce an output for control and action. Typical applications of AI include firewalls, Intrusion Detection Systems (IDS), Intrusion Protection Systems (IPS), security information and event management (SIEM), gaming, speech recognition, computer vision, expert systems and more.
Hyperconvergence – predictions for hyperconverged infrastructure will more than double by 2019 to the tune of $5 billion according to Gartner. Hyperconvergence is the means to combine compute power with storage, networking and virtualization technology because of the ease of deployment and management – shrinking everything down to a small footprint reduces power consumption and costs. IDC forecasts the hyperconverged infrastructure (HCI) market to reach $7.15 billion within the next three years.
Machine Learning (ML) – according to the Deloitte Technology, Media, and Telecommunications Predictions (TMT) 2018 Report, IT-as-a-Service is expected to approach $550 Billion globally by next year. In addition, over 300 million smartphones will include ML, within four years autonomous braking will reduce auto accidents, by 16 percent, and biometric fingerprinting reader devices will exceed one billion. ML, as a subset of AI, will be used to spawn new tools to fill talent gaps in the data center where repeatable and sometimes mundane workloads are involved – from patching, performing predictive maintenance, control cooling systems, fire suppression and more.
New Data – data that is generated due to the proliferation of Internet of Things (IoT) including connected car and wearables which create a deluge of information which, in many cases, has a short life span of usefulness in the corporate world. A digital representation of sensors and IoT systems is the digital twins which help with efficiencies, actionable information, and may be integrated with AI and ML to produce real intelligent, holistic systems perspectives.
Software Defined Networking (SDN)/Network functions virtualization (NFV) – it’s been conservatively predicted that by 2020, SDNs and NFVs in data center could exceed 5.2 zettabytes worth of traffic. SDN and NFV promises to make a paradigm shift towards managing future networks by reducing operational costs, improved resource utilization, and managing requirements.
Storage – adoption is growing for unified storage architecture that consolidates block and file services with open system clients including physical and virtual servers. The adoption continues with all-flash storage arrays that results in reduced infrastructure costs while unifying storage and that offer 50 percent reduction in energy costs. The Massachusetts Institute of Technology (MIT) researchers have created a new way of caching using flash memory to be more competitive with DRAM called BlueCache – although not new, it relies on pipelining subsequent queries to cache before the result of the initial query is received.
Green Energy – electrical power is often an afterthought as to the type and availability until either a man-made or natural disaster strikes. Identifying the most cost-efficient and green energy source typically offers tax credits as well as improvements to the corporate image and environment. Innovative alternative energy sources are maturing and becoming more capable in providing adequate amounts of energy – types of energy technologies include fuel cells, solar, wind, hydroelectric, hydrogen, and others but just as important is energy storage.
Edge Computing – with the explosion of data driven by advancements of wearables, sensors and connected cars - there is a need for localized compute power with very low latencies at the edge. More importantly, the timeliness of intelligent systems requires that round-trip times be minimized for time-sensitivity. Other industries have already begun to implement solutions including the Industrial Internet of Things (IIoT) which is expected to reach $7.5 trillion within the next seven years.
Platforms – specifically, conversational platforms are mostly associated with chat bots and virtual assistant systems that utilize text or natural language processing speech to interact with and accomplish tasks. According to Research and Markets, within the next five years, the intelligent virtual assistant market will grow at a CAGR of 38.82 percent.
Stick around as the data center landscape continues to grow and shift throughout 2018.