Entering edit mode
4.4 years ago
acamukhin
•
0
I have question about infrastructure part of bioinformatics organisations that develop web-service!
- If you use Kubernetes, why not Nomad, Swarm, k3s, etc.? Do you really use autoscalling in your process?
- How do you explain orchestration in same products?
I need a product which can migrate docker apps to a another node if a old node is down. I can understand business cases, but how does bioinformatics use it?
Thanks
Most of us don't provide application services that need an absolute 100% uptime so an immediate answer may not be forthcoming for your question.
Since google compute/AWS support kubernetes that may be a reason why those use that environment use it. I am not familiar with other packages you mention above.
Kubernetes seems to have won the competition at the moment. With the current funders'push for cloud deployment of scientific resources and tools (e.g. NIH’s Data Commons, Australia Research Data Commons, European Open Science Cloud), there's some incentives to move projects to the cloud. However, most of these are research projects, not services. Cloud technologies are used as a way of getting access to remote data and/or computing resources.
Okey, thanks, what can you say about using Kubernetes on your server machines?
I think the main issue is not with k8s specifically but with cloud technologies in general. Cloud techs have been primarily developed to serve the needs of e-businesses to deliver web services. This is not the typical use case in biological research. While the cloud can be harnessed to address some computational needs in research, it involves rethinking software and workflows.