Converged Infrastructure Helps Make Sense of the Internet of Things
When cloud service provider Virdata needed to develop a highly scalable platform to collect information from millions of devices and offer Big Data analytics to its customers, it chose a converged infrastructure.
The solution uses FlexPod, manufactured by Cisco Systems and NetApp, which integrates computing, storage and networking resources within a pre-tested bundle, and runs OpenStack’s open-source computing platform.
Virdata executives say the scalable, high-speed infrastructure enables customers to merge large volumes of historical and real-time data, and then quickly analyze the information for insights that lead to improved decision-making.
Virdata isn’t alone in its commitment to converged infrastructures. Enterprises are expected to adopt the technology for Big Data, the Internet of Things (IoT) and other workloads at a 24 percent annual growth rate over the next four years, totaling sales of close to $34 billion in 2019, according to the market research firm MarketsandMarkets.
“IoT and Big Data applications may collect, store and analyze video from security systems or sensor data from manufacturing equipment. That type of data doesn’t play well in traditional data warehouses,” says Charles King, principal analyst at Pund-IT. “But converged infrastructures have the necessary computing, storage and network capabilities to effectively support those applications, and as a result they can offer very compelling value in these use cases.”
Pre-Tested Flexibility
Converged infrastructures come in two main implementation models. The first is a complete package of IT components assembled and certified to work together by the vendors. The second option is a reference architecture, which provides a blueprint for IT staff members or system integrators to assemble their own converged systems.
Either way, converged infrastructures reduce integration and testing burdens for IT departments, which helps them more quickly deliver resources to support shifting business needs. “IoT and Big Data represent important types of applications for businesses, but most IT departments spend more time and money managing their complex IT infrastructures than on new applications,” says Bob Wambach, vice president of portfolio marketing at VCE, a provider of converged solutions. “Converged infrastructures take complexity out of data centers by creating units of IT infrastructure rather than individual compute, storage and networking resources.”
IT complexity does more than drain budgets. “Line of business managers become frustrated when it takes too long for IT to deliver new capabilities,” says Manoj Suvarna, director of Big Data solutions for Converged Data Center Infrastructure at Hewlett-Packard Enterprise (HPE). “When that happens, business units may decide to undertake shadow IT projects rather than wait for IT’s help.”
A Range of Options
Demands for sophisticated technology deployments have fueled growth in recent years for converged infrastructures to support IT operations such as virtual desktop infrastructure and private cloud implementations. Now, vendors are creating converged designs optimized for Big Data and IoT, which have unique requirements for computing and information management, based on the volume and data formats they must handle.
The percentage of senior IT professionals who planned to increase spending for converged infrastructure during the fourth quarter of 2015
SOURCE: 451 Research, “Voice of the Enterprise: Converged Infrastructure,” October 2015
For example, HPE offers converged infrastructure appliances and reference architectures based on its Vertica SQL analytics engine and technology from Cloudera, Idol Analytics, Oracle, SAP and others.
Similarly, NetApp and Cisco sell designs that use FlexPod and Hadoop Hortonworks, as well as Cloudera’s distribution of Apache Hadoop, says Abhinav Joshi, a senior product manager with NetApp.
Increased Resilience and Resources
Vendors say converged infrastructures that are optimized for IoT and Big Data remove guesswork over what models and configurations of IT resources best serve these use cases. Also, IT managers can quickly plug in additional units when demand for services grows, such as when a Big Data application starts out supporting a terabyte of data but quickly expands to multiple terabytes as business managers increase their reliance on the solution. “With converged systems optimized for IoT and Big Data, IT can increase compute, storage and networking while handling both structured and unstructured data,” says Survarna.
Fast access to sophisticated technology is also fueling adoptions. “Converged infrastructures offer the best-of-breed components for services, network and storage, with the fastest — Flash — and highest-bandwidth — Fibre Channel, SAS or Infiniband — storage,” says Pamela Kerman, NetApp senior manager for strategic marketing.
She adds that workloads handled by converged solutions can recover from disk failures with negligible effect on performance and much more quickly than with internal server storage. “Meanwhile, external data protection reduces both the storage footprint and data-replication overhead,” Kerman says.
Another plus: If problems occur within a converged infrastructure, IT managers turn to a single vendor for service and support. This saves time and effort compared to traditional approaches where IT departments must sort out issues among multiple suppliers.
The cloud-in-a-box characteristics of converged infrastructures also mesh well with overall cloud strategies and the regulatory requirements of individual industries. For example, healthcare and financial-services companies may use converged infrastructures to manage and analyze sensitive customer information that must be kept within the confines of local data centers. Enterprises in other sectors have the flexibility to use the infrastructures as part of native cloud or hybrid-cloud environments.
A Bridge to the Future
In the future, converged infrastructures may become building blocks for software-defined data centers that create highly fluid pools of IT resources. To support this vision, HPE is creating what it calls workload templates, Survarna says.
“The templates will specify X number of computing cores, memory reserves and disk storage, along with input-output levels to meet performance requirements for various workloads,” he explains. “So for eight hours a day, a converged infrastructure may run an enterprise resource planning platform. Then, at non-peak hours, an IT manager may use a different template on the same infrastructure for a Hadoop-based application that requires a different mix of resources. When that happens, IT becomes more flexible and agile than ever.”
To learn more about how converged infrastructure can simplify the deployment of advanced technologies, read the CDW white paper “The Ready-Made Data Center.”
For more coverage of data center technology, check out, The New Tools Needed to Defend Next-Generation Data Centers From Cyberattacks, and Automation and Analytics Emerge as Key Internet of Things Use Cases.
