Research Company
They had software from more than 20 vendors, running on over 30 license servers and approximately 150 users.
Research Company Uses OpenLM to Manage their GPU Licenses
Any organization that makes extensive use of artificial intelligence for predictive analytics and machine learning requires high-performance computing that exceeds the capacity of even the more powerful CPUs. This has led to a trend where graphics processing units or GPUs are being used in preference to CPUs.
The parallel processing capabilities of graphics cards can also be used for applications as diverse as finite element analysis (FEA) and cryptocurrency mining. Nvidia is the leader in the design of GPUs and their use is managed via a license agreement.
At OpenLM, we regard our product as primarily a license management application of engineering and scientific software, so when one of the customers of our partner Secoptena approached us in 2017 to ask if we could manage Nvidia licenses, we had to apply our minds to this new possibility. Our ability to manage software such as AutoCAD and Ansys is based on recognising features within the application. This functionality allows us to also manage and monitor custom software or other products that are not normally included within our scope, but we had never tried to monitor firmware before. After consultation with the customer, we gave the challenge to our development specialists to come up with a solution, and they did. We know that many of our existing customers make use of GPUs, as well as VDI (virtual desktop infrastructure), and that they will be interested in this new development. We will be adding further enhancements in the future, based on the requirements of our customer.
About our Customer
They are a leading German research and development organization, who use mathematics to derive new and optimized processes in fields ranging from production to communications. Simulations of the real-world problems are explored, replacing the traditional trial-and-error approach to improving processes. These virtual solutions are requested by and utilized in diverse industries, from the automobile industry to banking. The company has an impressive headcount of over 25,000 people, predominantly scientists and engineers, in over 70 institutes and research units. Customers from small and medium enterprises to large corporations such as Daimler AG make use of the organization’s services to innovate and optimise their products.
After successfully using OpenLM for almost 4 years, an integration of the Nvidia licenses into OpenLM looked perfect.
Manually tracking the usage of Nvidia’s licenses to stay compliant was not an option.
Original Purchase of OpenLM
Their internal customers had specific requests that needed to be fulfilled.
● Whether specific software features were being used, and to what extent
● Mean and peak usage over defined time periods
● The prevalence of denials and their causes
● The financial managers required an annual breakdown of usage by cost centre, together with peaks and troughs in demand for those centres.
● End users wanted a single view of installed and licensed software, rather than searching through each vendor’s proprietary license management tool.
They also acquired some additional features, namely:-
● OpenLM Advanced Alerts
● OpenLM Role Based Security
● OpenLM External DB Support
● OpenLM LDAP Groups Synchronization
● and OpenLM Group Billing
which made tasks such as cost centre chargeouts simple and automated.
Integrating Nvidia Licenses into OpenLM
This was a relatively easy journey, and the new enhancement provides for understanding the capacity of the GPUs, which will help when deciding whether more GPUs need to be purchased. Currently the Nvidia servers are not being used to any great extent, but this will change following some hardware upgrades that are to be done shortly, as well as rolling out various clients.
There are a few additional requirements our customer would like to see in the future. Because GPUs are performing processing in the background sometimes for several hours or even a day or more, a feature that permits the user to upload and check progress and possibly adjust and restart the job was requested, as well as the capability to roll back and track changes for a particular license. Finally, the ability to simulate and test scenarios would help to estimate whether new licenses are needed to avoid denials.
Our customer is satisfied with the new capability, and is sure that any of our other customers who make use of GPUs or have converted to a VDI environment will also be glad to hear that they can monitor and manage their Nvidia licenses via OpenLM.
