Watch our 2-minute video on AutonomousHPC.

Every supercomputer is unique. It may share similar compute, storage, network, software, and middleware components, but there’s no machine like it. Every subtle difference sets it apart from all other machines. Its function, its workload ― every feature defines that machine.

Nobody can feel the pulse of your machine like you do. AutonomousHPC is the tool that help you comprehend the next level of detail about your machine.

AutonomousHPC Solves Today’s Performance Problems

See how we rise to the challenge of processing 8.64 billion data points per day.



Re-Store ( has focused exclusively on technical computing systems for 12 years (Our Products & Services). During that time, Re-Store has developed a broad base of real-life administrative and support experience across hundreds of engagements. AutonomousHPC™ represents the codification of that experience and the intersection of modern analytic tools from Elastic Search.


Elastic Search is a JSON-based distributed, restful search and analytics engine that centrally stores your data, so you can discover the expected and uncover the unexpected.

Under the hood of AutonomousHPC beats a heart of Elasticsearch, a powerful open-source tool to store, search, analyze and visualize data. With each node of your cluster offering up to 1000 datapoints per second, this new flood of billions of datapoints is overwhelming. AutonomousHPC’s machine-learning algorithms make sense of this have been trained by Re-Store engineers with rich real-life experience to deliver a product that reduces the human requirement for a more efficient, more reliable, more predictable business or research workflow.


The automobile has been relying on sensor-driven on-board diagnostics for 30 years. AutonomousHPC applies similar sensor-driven on-board diagnostics that drive autonomous features to understand, automate, and optimize your HPC cluster. AutonomousHPC is able to collect those thousands of data points from every member of your cluster each second. With modern machine learning techniques, AutonomousHPC is able to use this massive amount of data to:

  • Understand how the cluster members behave
  • Understand the relationships between memory, processor, network, storage, and software
  • Automate routine tasks, that administrators would want to do, if only there was enough time
  • Use machine learning to forecast future behaviors
  • Interpret variation of observed values from predicted values
  • Select from three levels of technical assistance and support
  • Administrator Assist – detect, alert and escalate to local administration
  • Automated software response
  • Call home to Re-Store help desk through SupportLink
  • Delivered as an on-premise clustered system or as a secure cloud subscription


  • IT organizations are relying on human administrators for tasks that can easily be automated.
  • Human administrators cannot respond in real-time to the billions of data points available to manage a medium-sized cluster each day.
  • Traditional monitoring and management tools are device/vendor-centric, which are not aligned with business IT systems.
  • Organizations rely on vendor support resources that have no telemetry of the installed systems.


Today’s technology headlines are filled with news of the self-driving car, but consider the significant amount of automation already available in current cars that allow the driver to focus on the safe, efficient, reliable operation of the car: electronic ignition, electronic fuel injection, automatic transmission, anti-lock braking, traction control, automatic headlights, engine diagnostics, etc.

Each feature is discrete; it relies on the data from specific sensors and applies a heuristic, which generates an action. In each case, the amount of information and the need to make continuous real-time decisions would distract and overwhelm the driver.

Data Warehouse & Analytics

Collect, warehouse, correlate and perform powerful analytics on 13 channels of telemetry with 1300+ data streams. AutonomousHPC can collect efficiently billions of data points of health, performance, capacity, system events and configuration. This data can be used for trending, machine learning, root-cause analysis, time series analytics, and predictive analytics.

Real-Time Event Handling

The scale of HPC precludes the use of standard monitoring tools. A human administrator cannot watch, correlate and action the deluge of telemetry that’s produced by a modern cluster. AutonomousHPC uses machine learning to raise awareness of statistical changes in the behavior of the machine, forward log and system errors, and spotlight trending concerns.

Automation & Optimization

Like the autonomous car, we’re not ready for our machines to start contradicting our decisions. We want positive control of our machines, but we understand the need for automation. Consider anti-lock brakes. It’s a simple result of a 16 hertz data stream of tire rotation speed. When the car detects a significant difference, it cycles the brakes. A human being simply cannot collect and respond to that data in real time.

You Determine How Autonomous You Want It to Run

Just like the latest generation of autonomous cars, which can drive for you but allow you to operate with help (lane notification, emergency braking, etc.), AutonomousHPC can either adjust for you or notify you of which changes are needed.

Alerts sent via email, SMS, and dashboard. You have the data necessary to manage your system the way you want to manage it.