AIPrimer.AI
  • 🚦AI Primer In Transportation
  • CHAPTER 1 - INTRODUCTION TO MACHINE LEARNING
    • Machine Learning in Transportation
    • What is Machine Learning?
    • Types of Machine Learning
      • Supervised Learning
      • Unsupervised Learning
      • Semi-supervised Learning
      • Reinforced Learning
    • Fundamental concepts of machine learning
      • Model Training and Testing
      • Evaluating the Model’s Prediction Accuracy
      • The Underfitting and Overfitting Problems
      • Bias-Variance Tradeoff in Overfitting
      • Model Validation Techniques
      • Hyperparameter Tuning
      • Model Regularization
      • The Curse of Ddimensionality
    • Machine Learning versus Statistics
  • CHAPTER 2 - SUPERVISED METHODS
    • Supervised Learning_Complete Draft
    • K-Nearest Neighbor (KNN) Algorithm
    • Tree-Based Methods
    • Boosting
    • Support Vector Machines (SVMs)
  • CHAPTER 3 - UNSUPERVISED LEARNING
    • Principal Component Analysis
      • How Does It Work?
      • Interpretation of PCA result
      • Applications in Transportation
    • CLUSTERING
      • K-MEANS
      • SPECTRAL CLUSTERING
      • Hierarchical Clustering
    • REFERENCE
  • CHAPTER 4 - NEURAL NETWORK
    • The Basic Paradigm: Multilayer Perceptron
    • Regression and Classification Problems with Neural Networks
    • Advanced Topologies
      • Modular Network
      • Coactive Neuro–Fuzzy Inference System
      • Recurrent Neural Networks
      • Jordan-Elman Network
      • Time-Lagged Feed-Forward Network
      • Deep Neural Networks
  • CHAPTER 5 - DEEP LEARNING
    • Convolutional Neural Networks
      • Introduction
      • Convolution Operation
      • Typical Layer Structure
      • Parameters and Hyperparameters
      • Summary of Key Features
      • Training of CNN
      • Transfer Learning
    • Recurrent Neural Networks
      • Introduction
      • Long Short-Term Memory Neural Network
      • Application in transportation
    • Recent Development
      • AlexNet, ZFNet, VggNet, and GoogLeNet
      • ResNet
      • U-Net: Full Convolutional Network
      • R-CNN, Fast R-CNN, and Faster R-CNN
      • Mask R-CNN
      • SSD and YOLO
      • RetinaNet
      • MobileNets
      • Deformable Convolution Networks
      • CenterNet
      • Exemplar Applications in Transportation
    • Reference
  • CHAPTER 6 - REINFORCEMENT LEARNING
    • Introduction
    • Reinforcement Learning Algorithms
    • Model-free v.s. Model-based Reinforcement Learning
    • Applications of Reinforcement Learning to Transportation and Traffic Engineering
    • REFERENCE
  • CHAPTER 7 - IMPLEMENTING ML AND COMPUTATIONAL REQUIREMENTS
    • Data Pipeline for Machine Learning
      • Introduction
      • Problem Definition
      • Data Ingestion
      • Data Preparation
      • Data Segregation
      • Model Training
      • Model Deployment
      • Performance Monitoring
    • Implementation Tools: The Machine Learning Ecosystem
      • Machine Learning Framework
      • Data Ingestion tools
      • Databases
      • Programming Languages
      • Visualization Tools
    • Cloud Computing
      • Types and Services
    • High-Performance Computing
      • Deployment on-premise vs on-cloud
      • Case Study: Data-driven approach for the implementation of Variable Speed Limit
      • Conclusion
  • CHAPTER 8 - RESOURCES
    • Mathematics and Statistics
    • Programming, languages, and software
    • Machine learning environments
    • Tools of the Trade
    • Online Learning Sites
    • Key Math Concepts
  • REFERENCES
  • IMPROVEMENT BACKLOG
Powered by GitBook
On this page
  1. CHAPTER 7 - IMPLEMENTING ML AND COMPUTATIONAL REQUIREMENTS
  2. High-Performance Computing

Deployment on-premise vs on-cloud

There are two ways to deploy a High-Performance Computing System. HPCs can be deployed on-premise where the organization sets up its own data center with clusters as per its demand. Another method of deployment is to avail the service of a cloud service provider to set up a virtual HPC on the cloud. To choose from among these types of deployment several factors such as location, setup costs, operational costs, etc. must be considered.

Setup Costs: The setup costs include the overall costs incurred in creating the hardware and software infrastructure for HPC development. These costs are relatively lower in the case of cloud-driven solutions compared to on-premise solutions. For example, Amazon Deep Learning and Amazon Marketplace Instances come completely configured with GPUs. To counter this on-premise HPC solutions from Dell or NVIDIA also aim to reduce these setup costs with their popular frameworks and open-source HPC stacks.

Operating and Capital Costs: In the early phase of projects, when experimentation of the model is the major focus the workload is relatively less. During this period the operational costs for both on-premise and on-cloud HPC are comparable. Advancing into further stages, when the workload increases, more intensive use of cloud services can spike the cost significantly. Hence, agencies are advised to conduct a Total Cost of Ownership analysis (TCO) for their project to choose the mode of deployment.

Data and Application Locality: Any machine learning application to be run on HPC will be handling large datasets in the size range of terabytes. Data needs to be ingested, cleaned and processed before use in training the model. For these processes, it is beneficial if the application and data reside at the same location. This reduces latencies, improves quality of service and reduces costs. However, if the teams employed to work with the data are distributed across the world, on-cloud HPC solutions can provide better accessibility.

Security and Privacy: Most Cloud service providers promise complete security and confidentiality of the user’s data. However, they are vulnerable to attacks when the system is online. Users can decide based on the sensitivity of their data whether they are comfortable with hosting it on-premise or on-cloud HPC.

Workload Demand: Projects, where there might be temporary surges in workload demand, can choose to deploy on-cloud. Cloud services offer flexibility to choose capacity and pay as per usage. Availing this facility is a cost-effective strategy for organizations as it prevents them from managing more compute capacity than needed.

Performance: The computing bandwidth offered in cloud service is equivalent to the computing capability of on-premise infrastructure. Latency in cloud services is generally low. However, if the user is accessing from a long distance it significantly degrades performance due to an increase in latency.

PreviousHigh-Performance ComputingNextCase Study: Data-driven approach for the implementation of Variable Speed Limit

Last updated 1 year ago