How Technology Trends are Changing CI and HCI Solutions
From AI and advanced analytics to edge computing, containers and more – today's top technology trends are bringing new life to converged and hyperconverged infrastructure solutions.
February 18, 2020
There are several big technology trends whose impact is being felt throughout the data center and across many IT practice areas. These technologies are in the process of dramatically changing the IT landscape. This includes changes within the converged and hyperconverged infrastructure market, as vendors work to exploit and support these new developments. In this article we delve into how these trends are affecting CI and HCI solutions.
AI, Automation and Advanced Analytics
Technology is advancing across the board as computing systems get smart enough to see things we can’t and do things we choose not to. Automated vehicles incorporate several of these technologies, from image recognition to AI and advanced decision-making. High-volume manufacturers, like those in the disk drive industry, have been doing predictive analytics on drive failures for years. In the infrastructure space, we’ll see similar capabilities in systems that can monitor themselves and remediate some of the problems they find.
Integrated systems vendors are applying advanced analytics to the operational data they collect from their installed bases to optimize configurations and increase resource utilization and actually predict future requirements. They’re also working to simplify administration and detect anomalies that can indicate a failure scenario. This technology was introduced by HPE’s Nimble Storage a number of years ago and is being offered in one form or another by several other HCI vendors.
A related area is what I like to call the Day 2 problem. This is the situation that exists with systems that include multiple components, with software and firmware versions that must be kept in sync as they age. This is particularly applicable to converged infrastructure; Dell EMC and others are offering software that is supposed to address this problem.
Edge Computing
As a compact, self-contained environment, HCI is a potentially ideal architecture to support compute outside of the data center: computing at the edge of the network, where the data originates. These clusters can consume less space and go in more easily than traditional infrastructures while still offering a significant amount of compute power and storage capacity. Some typical edge use cases include running controls and automation on the factory floor, capturing and filtering video streams before they're sent to a processing location, or supporting mobile applications. In addition, the “edge nodes” that most HCI solutions are now offering appeal to traditional ROBO use cases, in remote offices and retail stores that need local compute power to run applications locally but still fit under the counter or on a shelf.
Containers
Kubernetes, a mainstay with developers, hasn’t replaced VMs in production, but container use in the data center is growing, and enterprise IT needs to be ready. Most HCI solutions now support one or more container platforms on premises and some can migrate containers to the public cloud. VMware and Cisco have products coming out that merge the container platform with the hypervisor, and Nutanix’s Karbon Kubernetes environment is incorporated into the Acropolis OS. KubeVirt is open source software that runs VMs in a container environment. Expect to see HCI solutions continue to add container capabilities throughout the year.
Disaggregated Architectures
The knock against HCI from the beginning has been resource efficiency. Each node typically includes storage and compute, one of the characteristics that made HCI simpler to design and deploy. But this also means you scale both resources when you add nodes. Vendors like NetApp, Dell EMC and Microsoft have changed this paradigm with HCI clusters that feature separate storage and compute nodes, referred to as disaggregated HCI. This can drive significant cost savings in hardware and software, as storage-only nodes don't need a hypervisor license and compute nodes don't need the full HCI stack.
For more information on how converged and hyperconverged infrastructures stack up and how current technology trends will affect these two integrated systems, download this report from Evaluator Group.
About the Author
You May Also Like