Hello Team,
One of the many options to run cuOpt is on an Azure VM that runs Nvidia Enterprise AI. ref: Microsoft Azure Marketplace
The cost of running Nvidia AI enterprise with a 4 GPU plan in Southeast Asia is around $7500 per month. Is this the only way to access and run the cuOpt services?
What if a customer of ours wants to use just the cuOpt service out of the whole lot and wants the same service (the cuOpt solver and the other microservices that can be run as containers) to be run on a serverless cloud service such as Azure functions or AWS lambda? The reason for asking this question is based on the following scenario:
Invoking cuOpt to calculate the optimized routes in response to a change in the supply chain (e.g. a route becoming unavailable due to a natural disaster or labor shortages). In these cases, the cuOpt pipeline can be invoked in response to these events and the serverless FaaS can be triggered only when needed.
Any information on this topic would be appreciated.
Thanks,
Sriram