Distribute algorithms and applications with ease
Unlock the full capacity of business datacenters, HPC, private clouds and public clouds to accelerate algorithms and applications. With an interface accessible to business line users, enable them to leverage the full compute capacity and accelerate innovation.
Business line users and scientists, are more and more empowered by IT in accessing the compute capacity. This requires training on technologies that should be used as commodity. With Activeeon, they can use the compute capacity easily with no mention of the underlying resource. It feels like a full serverless solution. On the other hand, IT connects the relevant multi-cloud resources onto the platform to provide what will be required.
Designed with a multi-cloud approach, Activeeon ProActive includes a Resource Manager that enables smart control over the resource.
With an infinite amount of resources available, the solution in charge of distributing the load needs to scale. Activeeon ProActive has been tested to run over 20,000 CPU cores.
With a clear understanding of resource utilization of each task, Activeeon ProActive can be configured to optimize the resources usage.
At Legal and General, on the same resource pool, a distributed VaR algorithm has gone from 18h end to end execution to 16h and only 5h for priority reports before the end of the whole batch.
At large scale, an advanced error management system is required to ensure any error is contained and related data are saved. With Activeeon ProActive, multiple policies to manage errors are available to ensure automated recovery or alerting.
Designed to be self explanatory, Activeeon ProActive reduces the learning curve to access big compute capacity. Business users can focus on their needs and IT department concentrate on providing the appropriate resource at any time. For instance, for data scientists and machine learning specialists, Activeeon designed a complete solution to perform distributed processing of machine learning pipelines.