No matter what industry you are in, operations has slow and busy time periods. Using cloud computing in peak demand periods is proving to be an easy alternative to renting temporary servers or even permanent data center expansions.
One example of this is in the making of movies. In post-production environments, demand for rendering nodes can fluctuate greatly. Projects come in spurts and these studios must ramp up quickly to meet deadlines and keep artists working. When the project is over, the ability to ramp down is almost as important. They can avoid not only the cost of the nodes, but also free up space and eliminate overhead costs associated with IT systems.
In our latest Ask Avere Almost Anything video, we answer a question about the limitations of cloud computing. How many clients can you run in the cloud? How do you scale performance for these clients?
We asked one of our solutions architects to respond. In his example, he demonstrates using the Avere vFXT to burst compute workloads and explains how to maintain performance thresholds when expanding compute clients.
If you’d like to learn more about scaling cloud workloads, Eight VFX explains how they deal with peak demand periods by bursting compute nodes into Google Compute Engine with Avere vFXTs in a recent webinar.
"Is there a limit to clients running in the cloud?"
Hi my name is Aaron Wetherold. I'm a Solutions Architect with Avere Systems. Today's question is, "Is there a limit to the number of clients that we can launch into the cloud? Or does the system queue access to the FXT?"
There's really no limit as far as the compute platform. You can really expand or add additional clients to that platform as you need to. From the FXT side we've designed the system to be very easily expandable. So as you need or as you increase capacity on the compute side you can just match additional FXT nodes or add additional FXT nodes to the system to maintain the performance thresholds that you expect for your particular workload.
So we'll look at indicators on the FXT. We have onboard UI that shows you what the latency, the throughput, the ops per second are. As those exceed a certain limit then you can expand the capacity on the cluster side to again match the workload generated by the compute side.
So hopefully that answers the question. Thank you and have a great day.
Learn more about how to scale cloud workloads with one-click cloud in this webinar with Eight VFX.