Implications of Public Cloud Resource Heterogeneity for Inference Serving

Keyphrases

• Cloud Resources 100%
• Public Cloud 100%
• Inference Serving 100%
• Latency 66%
• Resource Type 66%
• Model Heterogeneity 66%
• Low Latency 33%
• Confounding 33%
• Application Domain 33%
• Response Latency 33%
• Increasing Trends 33%
• Facial Recognition 33%
• Design Implications 33%
• Application Requirements 33%
• Prediction System 33%
• Billing 33%
• Personal Assistant 33%
• Application Constraints 33%
• Self-managing 33%
• Resource Model 33%
• Deployment Cost 33%
• Machine Learning Models 33%
• Machine Learning Prediction 33%
• Resource Characteristics 33%
• Holistically 33%
• Machine Learning Inference 33%

Computer Science

• Public Cloud 100%
• Learning System 100%
• Machine Learning 100%
• Application Domain 33%
• Face Recognition 33%
• Application Requirement 33%
• Personal Assistant 33%