Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CCO: predict the future spot interruption rates #41

Open
ilya-kolchinsky opened this issue Apr 17, 2022 · 1 comment
Open

CCO: predict the future spot interruption rates #41

ilya-kolchinsky opened this issue Apr 17, 2022 · 1 comment

Comments

@ilya-kolchinsky
Copy link
Collaborator

Different spot instance types are characterized by different interruption rates which are subject to change over time. When considering the reliability of the proposed deployment (#40), the optimizer should consider the future projected interruption rates. To that end, a time series forecasting solution utilizing historical data to predict future trends has to be implemented.

@ilya-kolchinsky
Copy link
Collaborator Author

Next, examine the mechanism by which the optimizer will not necessarily offer the cheapest instance, but the most efficient one - which its cost will be greater per hour of work, but it will work in less time than the cheapest instance.

This is not related to this issue and should be stated as a separate one.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant