Predictive Job Scheduling under Uncertain Constraints in Cloud Computing

Authors: Hang Dong, Boshi Wang, Bo Qiao, Wenqian Xing, Chuan Luo, Si Qin, Qingwei Lin, Dongmei Zhang, Gurpreet Virdi, Thomas Moscibroda

IJCAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our extensive experiments on three public, industrial datasets shows that CUC has great potential for supporting high reliability in cloud platforms.
Researcher Affiliation Collaboration 1Microsoft Research, China 2Microsoft Azure, United States 3The Ohio State University, United States {hangdong, boqiao, v-wenxing, chual, siqin, qlin, dongmeiz}@microsoft.com {v-boshiwang, gurvir, moscitho}@microsoft.com
Pseudocode Yes Algorithm 1: Training Procedures of CUC; Algorithm 2: Prediction Method in CUC; Algorithm 3: Greedy Construction Method in CUC; Algorithm 4: Testing Procedures of CUC
Open Source Code No The paper states that the pre-processed dataset 'will also be publicly available', but it does not explicitly state that the source code for the proposed CUC methodology is publicly available or provide a link to it.
Open Datasets Yes The experiments in this work are conducted on three public datasets, which are all collected from Microsoft Azure. In particular, two public datasets are introduced in the literature [Cortez et al., 2017] and each of both contains the representative VM workload of Microsoft Azure across 30 consecutive days; the other dataset is described in [Hadary et al., 2020] and includes the representative VM workload of Microsoft Azure across 14 consecutive days.
Dataset Splits Yes Each of the three datasets is divided into a training set and a testing set. For the first 5 days out of the 14 days, we only use the hourly available capacity for fitting our prediction model. Then a sliding window is adopted to generate samples, with each sample containing past hourly capacity for 5 days (i.e., 120 data points for prediction), future capacity for T = 24 hours, and the job requests in the next 24 hours.
Hardware Specification Yes In this work, all experiments were conducted on a machine with Intel Xeon E5-2690 v4 CPU, 112GB memory and NVIDIA Tesla P100 GPU.
Software Dependencies No The paper mentions using 'Gurobi, a professional solver for linear and non-linear optimization problems' but does not specify a version number. It also mentions other methods like 'time series decomposition based forecasting approach (TSDec)', 'automatic autoregressive integrated moving average (Auto ARIMA)', 'long short-term memory (LSTM)', 'un-observed component model (UCM)', and 'Prophet', but does not provide specific version numbers for any libraries or frameworks used to implement them.
Experiment Setup No The paper describes the overall design of the CUC algorithm and mentions setting 'the number of BO steps as k = 50', but it does not provide specific hyperparameters like learning rates, batch sizes, or optimizer settings for the prediction models or other components of the experiment.