Hi Uwe, I'm providing a cloud based service that lets users run some domain specific jobs on the cloud. Some of them are batch, some are interactive and some runs for long period of time. I need to be able to provide them a minimal guaranteed QoS as per the plans they purchased. To *reduce my overall operational cost *and also to provide the lowest possible plans to my customers, I need to manage my computing resources very diligently. I also need the ability to manage the workload and jobs from within my own application [so APIs for a workload manager are required].
Do you think SLURM is appropriate for my use case ? Thank you On Thu, May 14, 2015 at 9:06 AM, Uwe Sauter <[email protected]> wrote: > > Before being able to answer your questions, you'd probably should tell us > what you want to achieve with a workload manager such as > Slurm. > > > > Am 14.05.2015 um 17:58 schrieb Pradeep Bisht: > > [Resending as I don't see my message in the archives] > > Show message history > > > > I'm looking at SLURM and it seems to have all the features that my > application needs. I have a very basic question. I see that > > SLURM is very popular for supercomputer clusters. However, I couldn't > find any use scenarios where it is being used on commodity > > Linux machines like the one being in cloud data centers. I'm a cloud > service provider with EC2 instances in Amazon cloud. Do you > > think SLURM is suitable for this kind of infrastructure ? I'm just > concerned that the overhead might be too high for commodity > > machines and may be SLURM is designed for supercomputers only ? > > > > Can somebody please also cite some companies name who are using SLURM > for cloud based services ? > > > > > > Thank you >
