Exploring Latent Features for Memory- Based QoS Prediction in Cloud Computing Yilei Zhang, Zibin Zheng, and Michael R. Lyu Department of Computer Science & Engineering The Chinese University of Hong Kong Hong Kong, China School of Computer Science National University of Defence Technology Changsha, China SRDS 2011, Madrid, Spain, Oct , 2011
Outlines Introduction System Architecture QoS Prediction Approach Experiments Conclusion 2
Cloud Computing Cloud computing provides a model for enabling convenient, on- demand network access to a shared pool of computing resources : Networks Servers Databases Services 3
Cloud Applications Building on a number of distributed cloud components Large-scale Complicated Time sensitive High-quality Case 1: New York Times Used EC2 and S3 to convert 15 million scanned news articles to PDF (4TB data) 100 Linux computers 24 hours Case 2: Nasdaq Uses S3 to deliver historic stock and fund information Millions of files showing price changes of entities over 10 minute segments 4
Performance of Cloud Components High-quality cloud applications rely on the high-quality of cloud components. remote network access Location independence Personalized performance evaluation on cloud components is essential. Method 1: evaluating all the components to obtain their QoS performance. Impractical: time-consuming, expensive, thousands of components. Method 2: collaborative filtering approach Predicting component QoS by employing usage experiences from similar users. 5
System Architecture 6
Example User-component matrix: m × n, each entry is a QoS value. – Sparse – Prediction accuracy is greatly influenced by similarity computation. 7
Latent Features Learning 8 Latent-component matrix HLatent-user matrix V u1 u2 u3 u4 c1 c2 c3 c4 c5 c6
Similarity Computation Pearson Correlation Coefficient (PCC) Similarity between users: Similarity between components: 9 Latent-component matrix H Latent-user matrix V u1 u2 u3 u4 c1 c2 c3 c4 c5 c6
Neighbors Selection For every entry w i,j in the matrix, a set of similar users towards user u i can be found by: A set of similar items towards component c j can be found by: 10
Missing Value Prediction Similar User-based: Similar Component-based: Hybrid: 11
Experiments QoS Dataset Metrices : the expected QoS value. : the predicted QoS value N : the number of predicted values. 12
Experimental Results 13
Experimental Results 14
Experimental Results 15
Experimental Results 16
Conclusions and Future Work Conclusions: A collaborative approach for personalized cloud component QoS value prediction A large-scale real-world experiment A publicly released real-world QoS dataset Future Work: Investigation of more QoS properties Experiments on different kinds of cloud components 17
Thank you! 18