How IMP Vienna Automated HPC Cost Allocation and Billing
The Research Institute of Molecular Pathology (IMP) in Vienna meters the IT consumption of its High-Performance Computing (HPC) clusters with Exivity — cutting the wait for usage and billing insight from three months to one day for hundreds of users.
We reduced the waiting time to gain insights into the IT resource consumption from every 3 months to 1 day.
IMP Vienna
Biomedical research center, IMP Vienna
Customer story
The Research Institute of Molecular Pathology (IMP) is a biomedical research institute located at the Vienna BioCenter in Vienna, Austria. Its research spans areas including molecular and cellular biology, immunology and cancer, structural biology, and biochemistry.
Together with other Vienna BioCenter institutes—including the Institute of Molecular Biotechnology, the Gregor Mendel Institute of Molecular Plant Biology, and the Vienna BioCenter Core Facilities—IMP uses a shared IT infrastructure supporting approximately 750 users across the campus.
In collaboration with the Austrian Academy of Sciences (ÖAW) and other universities, IMP also operates a High-Performance Computing (HPC) cluster serving around 300 active users across Austria.
As demand for scientific computing grew, IMP needed a more efficient way to measure HPC resource consumption, allocate costs, and provide users with timely insight into what they were consuming.
The challenge: managing the cost of data-intensive research
Modern scientific research generates enormous volumes of data.
IMP works with advanced technologies such as cryo-electron microscopes, which generate high-resolution image and video data at a rate of up to 50 images per second. With individual images around 8 MB in size, the microscopes alone can generate approximately 15 TB of data per day.
Additional datasets are generated through technologies such as next-generation sequencing.
Because of these demanding workloads and the volume of scientific data involved, much of IMP's infrastructure is operated on-site.
Its HPC environment effectively provides scientific computing capacity as a shared service. Research groups can access the computing resources they need and are charged according to the resources they consume.
This created an important requirement: users needed a clear and accurate view of both their HPC consumption and the associated costs.
From manual HPC reporting to automated cost allocation
The HPC environment had become essential to numerous research groups, including teams within the Vienna BioCenter itself.
The infrastructure included more than 200 servers, 7,700 CPU cores, and 120 GPU cards, creating a substantial amount of consumption data that needed to be measured and attributed to individual users and customers.
Before implementing Exivity, IMP prepared HPC usage and cost information manually using tools such as Excel.
The process required considerable manual effort and lacked the automation needed to efficiently process growing volumes of consumption data. Manual preparation also increased the possibility of errors, making it harder to provide users with timely and reliable information.
As a result, users could wait as long as three months to receive an overview of their HPC service consumption.
IMP needed to replace this process with a scalable system capable of transforming infrastructure usage into accurate cost and consumption information much faster.
What IMP needed
IMP identified several key requirements for its HPC environment:
Detailed insight into HPC resource consumption
Cost allocation on a per-customer basis
Automated IT billing
Faster access to usage and cost information
Reduced dependence on manual data processing
Greater accuracy across billing workflows
Exivity provided the metering and billing capabilities required to connect HPC resource consumption with the appropriate users and associated costs.
The results
Usage visibility reduced from three months to one day
One of the most significant improvements was the speed at which HPC users could receive information about their consumption.
Previously, users could wait up to three months for a summary of their HPC resource usage.
With Exivity automating the underlying metering and billing process, that waiting period was reduced to one day.
Researchers and other HPC users could therefore gain much faster insight into the resources they were consuming and their associated costs.
Detailed HPC cost allocation
Exivity enabled IMP to create clear, detailed overviews of HPC consumption and costs on a per-customer basis.
Usage from the shared computing environment could be attributed to the appropriate consumers, creating greater transparency around how infrastructure resources were being used.
This was particularly important in an environment where hundreds of researchers and multiple organizations shared access to the same computing infrastructure.
Automated billing
Before Exivity, billing-related information was prepared manually using spreadsheets.
Automating this process significantly reduced the amount of manual work required to transform HPC usage into billing information.
As a result, fewer FTE resources needed to be involved in the process, reducing administrative overhead and allowing staff to focus on higher-value activities.
Fewer manual errors
Reducing manual data processing also reduced the opportunities for errors within the billing workflow.
Instead of repeatedly collecting and manipulating HPC usage data in spreadsheets, IMP could use an automated and repeatable process for translating consumption into cost information.
This improved the reliability of reporting while minimizing the administrative effort required to maintain it.
Making HPC consumption financially transparent
High-performance computing environments present a particular cost management challenge: expensive infrastructure is shared between many users while workloads can consume dramatically different amounts of CPU, GPU, storage, and other resources.
For IMP, providing scientific computing capacity therefore required more than monitoring infrastructure performance. Users also needed to understand how much they consumed and what that consumption cost.
By automating HPC metering, cost allocation, and billing with Exivity, IMP transformed a manual process into a scalable financial workflow.
The result was clearer per-customer HPC cost visibility, automated billing, fewer manual errors, and a reduction in reporting time from three months to just one day.