Clunky, slow, and costly: Data center management is falling behind. As facilities race to keep up with increasingly power-hungry, high-performance chips, operators are struggling with fragmented management tools and limited visibility.

Margarita Groisman is the co-founder and CEO of Aravolta, a next-generation data center infrastructure management (DCIM) provider tackling one of the industry’s most underestimated challenges.

Alongside co-founder Jack Sutton, Groisman drew on her experience in Microsoft’s data center division, where she gained firsthand insight into the operational demands of deploying cutting-edge hardware.

“At any given time, something was breaking. We were constantly in firefighting mode,” she recalls.

Recognizing these inefficiencies, Groisman and Sutton set out to build a solution. The result was Aravolta: A single pane of glass view of all assets across building, electrical, and gray space. Features include branch circuit monitoring and cross-connect management to automate manual workloads.

Grappling for AI and GPU capacity

The data center industry is all too familiar with the challenges posed by the high-density workloads required for AI and GPU applications. These workloads demand significantly more power, cooling, and specialized infrastructure to maintain performance and reliability compared to traditional computing tasks.

Cooling typically tops the list of concerns. The heat generated during GPU processing can quickly lead to server overheating, triggering a cascade of problems that may disrupt the entire system, as Groisman illustrates:

“Overheating issues are especially common in highly utilized GPUs. So now, operators are forced to reduce the utilization of this very expensive hardware to prevent problems, or end up with shorter device lifespans and other issues over time.”

These thermal and utilization challenges directly impact the speed, efficiency, and stability of GPU training and inference tasks, making proactive management essential to maintaining optimal performance and hardware longevity.

The limitations of tradition

According to Groisman and Sutton’s conversations with a variety of data center operators, legacy and manually processed DCIM tools often have implementation timelines exceeding 18 months – and even then, their capabilities remain limited.

With a manual system, operators spend countless hours entering every single asset and its associated data, requiring them to constantly align what’s actually happening on the data center floor with what’s recorded in the system.

This can seriously undermine data quality and data accuracy, while consuming time and labor that could be better used elsewhere – especially critical in an industry facing ongoing shortages of skilled technical workers.

Redefining DCIM

As many in the industry can relate to, the landscape of data center infrastructure and management is constantly evolving. In fact, when Groisman and Sutton began developing Aravolta, they weren’t initially sure it even fit within the boundaries of the traditional “DCIM” label. As Groisman puts it:

“We thought it was something completely different, but we realized it serves the same purpose. It’s a tool for data center operators to do their job more effectively, more efficiently, and with clearer visibility into what’s going on on the data center floor.”

So, what exactly is Aravolta?

Groisman describes it as a modernized data center infrastructure platform that streamlines asset management, failure detection, and optimization – all through a single-pane-of-glass interface.

The system collects granular data from each device, from the chassis down to the GPU level, and then maps that data into correlational graphs and heat maps. This allows operators to detect subtle anomalies early on. According to Dr. Martin Rapos, EVP at Flexnode Datacenters:

"This was by far the easiest setup and the most impressive monitoring system I've encountered. While many companies promise seamless, 'plug and play' integration with a wide range of data center IT and OT systems, these guys actually deliver.”

Combining BMS, EPMS, and DCIM

Traditional DCIM systems typically focus on the gray space – servers, racks, PDUs, and switches. In contrast, a modernized tool like Aravolta offers a unified approach by combining data from all building and electrical assets.

By integrating information from both building management systems (BMS) and electrical power management systems (EPMS), Aravolta provides a single, centralized view that spans from the building level down to electrical components.

Data centers contain numerous power-related assets but often face challenges tracing how upstream power issues impact downstream systems. Aravolta addresses this by visualizing power flow throughout the facility, enabling operators to pinpoint the exact downstream effects of any upstream problem. They also detect lost power as it moves across systems.

As an automated software system, Aravolta also discovers assets automatically, ensuring a 100 percent accurate source of truth. The utility node within the data center detects and adds new devices to the platform without requiring manual onboarding by operators.

Adding further value, Aravolta is vendor-agnostic. Its custom-built middleware layers enable the system to quickly onboard a wide variety of devices from different vendors and protocols, streamlining all data into a single, unified format.

“Across all of your different devices and data collection methods, everything resides in a time series database in a single format so that you can utilize the data in the same way,” says Groisman.

AI-driven optimization

Of course, no modern software tool is built without AI capabilities in mind. Aravolta follows this approach, continuously collecting extensive asset data to build a model tailored to each data center. This model identifies optimal operating strategies and delivers data-driven recommendations to improve networking, cooling, and power performance based on the facility’s historical data.

“The AI journey starts with collecting highly detailed, millisecond-level data across the entire data center – humidity, temperature, devices, etc. This data is used to build a customized model for each unique environment,” says Groisman, adding:

“Currently, the model provides recommendations to operators to decrease energy costs and prevent downtime.”

Work smarter, not harder

The ultimate goal of a unified DCIM is to help data center teams improve their overall efficiency by automating manual tasks and uncovering smarter ways to optimize operations. According to Groisman, through real-world deployment and analysis of onboarded customer results, the platform has produced a promising set of results demonstrating its impact.

By tracking previous deployments, Groisman reports that customers have seen a 20 percent reduction in downtime – directly tied to reduced manual effort and more proactive system management.

On the optimization side, Aravolta’s analytics reveal significant energy-saving opportunities through its customized recommendations, such as lowering fan speed settings on switches.

“From those optimization recommendations, customers were able to get two to five percent in energy savings per year, which adds up really quickly,” says Groisman.

Unified in sustainability and security

In an era of tightening SLAs and growing sustainability mandates, Groisman advocates that Aravolta is designed to deliver resilience, lifecycle optimization, and built-in compliance as core differentiators.

On the sustainability front, the platform goes beyond traditional metrics like PUE to track detailed, asset-level performance using advanced measures such as data processing power. This level of granularity supports more effective lifecycle management by identifying inefficiencies early, reducing unnecessary wear, and extending the operational lifespan of critical equipment.

Equally important is compliance. Aravolta automates the tracking and reporting of required sustainability metrics like PUE, energy reuse factor, and other regional or national requirements (such as the Data Center Optimization Initiative, or DCOI, in the US and the EU Directive 2023/1791). By removing the need for manual monthly reporting, the platform ensures operators can maintain consistent regulatory compliance while reducing overhead and human error.

“It really all comes down to the data model, but just as important is working very closely with our customers – listening more than talking, truly understanding what's happening across the industry,” explains Groisman.

“Our goal is to build a platform people love to use, while aligning with best practices and guidelines from organizations like OCP, the Green Grid, and others.”

Security is at the heart of operational resilience. Aravolta meets standards like SOC 2 Type 2 and provides both cloud-based and fully air-gapped on-premises options. The on-premises setup is completely cut off from external networks, with the onsite utility node blocking unauthorized access and protecting against threats like hacking and bots. With multiple layers of defense, the platform keeps customer data and systems safe and secure.

With a rapidly growing customer base spanning a diverse range of data centers, Aravolta is positioning itself as the modern DCIM solution built for a world of stricter SLAs, sustainability goals, and demand for truly unified infrastructure management.

Go to Aravolta.com to book a demo today.