A framework for monitoring workload schedules in compute clusters
- Type: Bachelor Thesis
- Status: Completed
- ID: 2021-010
- Student: Julius Hinze
Cloud computing has become very high in demand in the recent years, enabling companies to rapidly develop software products. Due to the high demand, the number of cloud providers is growing larger by the day, making it hard for users to find just the right service for just the right price. Consequently, the need for price estimation tools for cloud users is greater than ever. While such price estimation tools already exist, they are mostly either limited to a single cloud provider or limited in functionality; that is, they are limited in estimation accuracy and mostly only work for simple infrastructure as a service. Tech start-ups and companies would benefit if such tools could also estimate prices for any cloud provider, for any kind of cloud computing service, given only a collection of software and hardware requirements. It turns out that this problem is solvable and satisfactory results are achievable using a model-as-code approach.