This article is part of a VB special issue. Read the full series here: Data centers in 2023: How to do more with less.
Cloud adoption is growing, but data centers still remain the driving force for most enterprises worldwide. The reason is obvious: You get more control and flexibility over the servers and associated components (like networking equipment and storage). But, when you have a lean IT team at the helm, keeping such infrastructure ready for complex and modern workloads can also be quite a task.
Case in point: Iowa-based Interstates, a 60-year-old provider of electrical, construction and engineering solutions, faced the challenge of frequently upgrading data center hardware to meet its business needs. We dive into its digital transformation story
Interstates’ VM-heavy workload
Interstates focuses on design-build electrical projects, plant floor automation and operational technology support for manufacturing and industrial environments. To provide these services, the company emulates its customers’ environments within virtualization stacks hosted in two data centers — one primary and the other secondary.
“On average, we run between 500 to 700 customer-specific virtual machines as we develop applications to run their environments and simulate that,” Nathan Bullock, the company’s IT operations manager, told VentureBeat. “It’s a pretty heavy workload for a virtualization stack. Plus, there is traditional Windows server architecture within manufacturing environments and traditional-type workloads.”
The AI Impact Tour
Connect with the enterprise AI community at VentureBeat’s AI Impact Tour coming to a city near you!
Bullock leads a team of about 20 IT experts who, among other things, make sure that data center infrastructure meets the demand of these business-critical applications and virtual machines. However, with core infrastructure aging, his team often struggled to quickly spin up required systems as needed. They had to update the hardware footprint now and then, which demanded major upgrades of the main chassis and took a lot of time.
“We had to overhaul the hardware from time to time,” Bullock explained. “We had to look at upgrading server midplanes and backplanes (circuit boards) to be able to sustain new blade infrastructure and maintain our growth. It wasn’t allowing us to be very agile.”
He noted that the issue was too disruptive and costly for the business and was happening too often, thus creating clear scalability and management bottlenecks for their tight-knit group.
In search for a simpler and easily scalable compute infrastructure to deliver 3D models in real-time and enable customers to review projects virtually from construction trailers, Interstates landed on Cisco’s UCS X-Series servers and Intersight operations platform.
“We looked at Dell, IBM and others, but they (Cisco) came in easily 20% less than what we were getting from some of these other vendors,” said Bullock. “That allowed us to invest in Intersight and potentially other tools.”
But, pricing is not the hero here. It is modularity.
According to Bullock, with its latest UCS platform, Cisco was able to showcase how within one chassis, one could have not only compute nodes but also need-specific memory nodes and GPU nodes — something very relevant to their line of work. The vendor has partnered with Nvidia and Intel for GPU footprints, giving them the flexibility to pick and choose based on the workload at hand.
Meanwhile, the cloud-based Intersight offering brought a single pane of glass to the table, giving cross visibility to manage not only compute, memory and GPU but also other infrastructure in the data centers, like storage and networks. This was, again, very lucrative for the company.
It purchased the first UCS-X platform for its secondary (disaster recovery) data center about 12 months ago. The company purchased the next set three months ago for its primary data center.
Data center footprint reduced by half
While Interstates is still in the process of migrating workloads from its production environment to the new platform, it already claims to have witnessed notable benefits, including a reduction of more than 50% in the server footprint of its two data centers.
“We’ve significantly reduced our rack space, we’re using less power and cooling, and we can upgrade components over time instead of full rip-and-replace overhauls,” said Bullock.
His team can now tune each box for specific workload needs, independently scaling memory, storage, CPUs and GPUs on a server-by-server basis. No more need to constantly overprovision or buy new ones.
This has helped Interstates reduce the frequency of full-fledged server infrastructure overhauls, improve systems integration, automation and scalability and simplify and accelerate data center operations.
“Every time you upgrade, there’s greater performance,” said Bullock. “We’re seeing at least two terabytes of RAM in each system so we were able to reduce our blade count or node count from five to three, and still have the growth capacity we need.”
Plus, the reduced consumption of power and cooling as well as the 10 to 15-year shelf-life of the new platform (as claimed by Cisco) is helping with cost savings, he added (without sharing the exact numbers).
Spinning up systems 75% faster
With this modernization, Interstates is also leveraging automation support in the deployment experience, spinning up physical and virtual machines 75% faster than it could with the former infrastructure.
“The installation of (Cisco UCS X server with Intersight) is extremely faster than what we’re accustomed to… it would usually take us a day to set up a basic VMware host and it was able to do that within an hour and create automated profiles, which allowed us to have our entire environment set up within a day,” said Bullock. “Traditionally, it would probably take us a couple (days), at least.”
“Because of the simplicity of the platform, we can do more on our own than ever before,” he added. “And we can do it all through a single pane of glass.”
Moving ahead, the company plans to extend its new computing environment to the cloud and align its internal cloud and data center network with the public cloud to manage it all as a single environment.
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.