Various Datasets Accompanying the Paper "Infusing Formal Methods into Ops-related Practices: Case Studies in Resource Allocation in the Cloud"

Various datasets accompagnig  the paper Infusing Formal Methods into Ops-related Practices: Case Studies in Resource Allocation in the Cloud. The code used for generating them and how to use them is available at here.

TrainingDatasets.zip contains datasets for 4 use cases. More details about the properties of the dataset is in the paper, Section 3.3.

GNNModels.zip contains datasets, for each of the 4 use cases, with graph neural networks trained using the above datasets with the architectures and hyperparameters as specified in the paper, Sections 4.1 and 5.2.

SMT-LIB.zip contains datasets, for each of the 4 use cases, with constrained optimization problems in the SMT-LIB format augumented with soft constraints extracted from GNN predictions as discussed in the paper, Sections 4.2 and 5.3.

StatDataFromSMT-LIB.zip contain datasets, for each of the 4 use cases, used for drafting the conclusions for the 3 research questions in the paper. The data is extracted from the SMT-LIB files above and contains information like Samples, Epochs, Batchsize, Offers, Type, Time, Price which is then used to draw the conclusions for RQ1-3.