Digital Transformation Drives Converged Infrastructure Solutions
The same IT trends that create challenges for CIOS are also driving their suppliers to come out with products and technologies that can help. Some of the latest converged infrastructure solutions include the use of automation, machine learning, microservices and APIs.
November 26, 2019
![converging flat arrows converging flat arrows](https://eu-images.contentstack.com/v3/assets/blt07f68461ccd75245/bltf57191547ba31f83/661ce8386115a5d38aeee6cf/1139052717.jpg?width=1280&auto=webp&quality=95&format=jpg&disable=upscale)
Enterprises are struggling with digital transformation: what it means, how to adopt technologies to enable it, how to manage the expectations of management and stakeholders etc. DT is a process, not a product, one that manifests internally as improvements in the way companies operate and externally in the new products they develop and how quickly they come out.
Infrastructure vendors are going through their own digital transformation journeys as well and are incorporating new technologies into their products, the ones their customers are using to implement their own digital transformations. I’m not trying to create a circular discussion here, but the same IT environmental trends that create challenges for IT managers and CIOs are also driving their suppliers to come out with products and technologies that can help. Some of these are the use of automation, artificial intelligence/machine learning, microservices and APIs.
Converged infrastructure was created a dozen or so years ago to simplify the procurement and deployment process for traditional "three-tier" solutions. Enterprises were buying storage arrays, SANs and open systems compute components (bare metal and virtualized servers) and needed a way to design and deploy these systems faster and more easily. Over the past decade or so, every storage vendor has come out with a line of converged infrastructure solutions focused on specific software applications or hardware platforms.
Hyperconverged infrastructure systems have had automated setup routines for years, which typically take an hour or less. They're also easy to expand, with most allowing nodes to be added to the hyperconverged infrastructure cluster with no or minimal disruption. But converged infrastructure solutions are much more complicated and require the coordination of multiple components instead of a few nodes that all run the same software. The deployment experience is simple, but updates are not.
Automating Hardware and Software
Issues confronted during updates or changes are what I call “Day 1 problems.” Day 0 is the installation and configuration of the systems, and all vendors do this well. The certified, tested and often pre-assembled stack goes up easily. The problem comes after, say, six or 12 months -- Day 1 -- when the storage system, servers, software or switches need firmware updates or the system needs to grow. This typically requires a manual check of certification matrices to make sure that upgrading one component won’t conflict with another.
Vendors are starting to automate the update process to eliminate the manual coordination of firmware versions and hardware models for each component in the converged infrastructure solutions stack. Controller software can identify the latest compatible versions of firmware based on the components in the system and download the new versions, alerting administrators to perform the installations. The next step is to fully automate the update process, using APIs to install firmware and then keep the stack updated (Dell EMC claims its new PowerOne has this capability). All the vendors promise some level of automation but, like we say often about technology features, the devil is in the details. There’s a lot of potential benefit here, and the technology is generally available to make it a reality. We expect the industry to keep raising the bar, but IT needs to have the confidence in their IT infrastructure vendors to turn this aspect of management over to the system.
AI and ML
As technologies such as artificial intelligence (AI) and machine learning (ML) have become more commonplace, they’re finding their way into converged and hyperconverged infrastructure systems. Storage vendors have been collecting telemetry data from their installed base for years, to monitor system uptime and troubleshoot problems. Now, they've taken this a step further and are applying AI and ML techniques on these data to derive insights about how to improve converged and hyperconverged system uptime and increase operational efficiency. We talked about this cloud-based analytics last month.
Microservices and APIs
APIs (application programming interfaces) are standard software routines and protocols that facilitate communication between components and systems. APIs enable controllers to monitor, configure and sometimes operate the various components that make up a converged infrastructure. Microservices, meanwhile, is a software development technique that breaks up large, monolithic code bases into independent "services" that can be modified easily, reducing the complexity in large software projects. In addition to improving their software development processes, a microservices architecture can incorporate the API calls that controllers need to do their work.
Data volume continues to grow and the shortage of “human capital” is becoming more acute, so automation is one of the solutions infrastructure companies are adopting. We're seeing these suppliers use microservices architectures and APIs to assume the long list of tasks associated with setup, operation, expansion and updating of complex converged infrastructure solutions. They’re even converting legacy products' software to microservices to improve the development process and to incorporate more advanced features, like automation.
About the Author
You May Also Like