Memory optimized because

Reserved instance because

Scenario 1: Web Application for Patient Records

AWS EC2 Instance Type:

Storage Optimized - Due to the large amount of data across the hospitals that performs a read-write operation

Pricing Model:

Dedicated Host - More cost-effective, steady traffic

Scenario 2: Real-time Health Monitoring System

AWS EC2 Instance Type:

Compute Optimized because it is suitable due to its high compute power, enhanced networking capabilities, and ability to handle real-time health data processing efficiently.

• Low Latency: Crucial for real-time data processing.

• High Throughput: To handle continuous data streams from wearables.

• Elasticity: To manage varying loads, particularly during peak usage times.

• High Network Performance: Essential for quick data transmission and responsiveness

Pricing Model:

On-Demand because it provides flexibility without long-term commitments, ideal for variable workloads typical in real-time health data processing.

Scenario 6: AI Model Training for Diagnostic Assistance

• Instance Type: Accelerated Computing

• Pricing: Dedicated Host

• Explanation: Accelerated computing instances can perform function such as floating-point number calculation, graphic processing, or even data pattern matching, with powerful GPUs which are essential for fast processing and efficient training of AI models in diagnostic assistance. Furthermore, using dedicated hosts ensures consistent performance and enhances security by providing isolated resources, crucial for handling sensitive healthcare data during AI model training.

Scenario 4: Medical Image Processing

Description:

A service for processing and analyzing large sets of medical images such as MRIs and CT scans using deep learning models.

Accelerated Computing Instances (P3):

• Description: These instances are designed for compute-intensive applications that require high-performance GPUs.

• Use Case Fit: Ideal for deep learning tasks. Specifically, P3 instances come with NVIDIA V100 and A100 GPUs, respectively, which are highly optimized for deep learning workloads.

  
  

Pricing Model

Combination of Reserved and Spot Instances:

• Reserved Instances: For the base workload that you know will consistently be required, reserving instances will reduce costs significantly.

  - Usage: Ideal for baseline capacity. Use Reserved Instances for critical and continuous tasks that require high availability.

  
  

• Spot Instances: For additional capacity to handle peak loads and batch processing, spot instances can be used to save costs.

  - Usage: Suitable for handling peak loads and non-critical tasks. Use Spot Instances to leverage cost savings for flexible workloads that can tolerate interruptions.

  
  

Scenario: Genomic Sequence Analysis

Amazon EC2 Instance Type: Either Accelerated or Memory. Accelerated because Genomic Analysis requires pattern recognition, and statistical analysis. However, we could also use a Memory-typed instance because loading datasets related to genomic research such as analyzing FASTA sequences requires huge amounts of memory.

Pricing: On-Demand, Spot Pricing

We plan to use On-Demand Pricing because the usage can peak at certain parts of the study. We wouldn't also need a dedicated server as it is only pure processing and we aren't serving anything to end-users. Furthermore, for a more budget-oriented option we could also opt to use Spot Pricing. However, this has the disadvantage of being outbid by other users of the instance.

Scenario 3: Drug Discovery Research

AWS EC2 Instance Type:

Storage Optimized

Pricing Model:

Reserved Instances

Explanation: Storage-optimized provides high-capacity, ensuring efficient handling of large datasets with low latency, critical for data-intensive chemical simulations. Using Reserved Instances is beneficial for predictable, long-term workloads because it offers significant cost savings, as it provide fixed discounted usage price.

.Instance Type: Accelerated Computing.

.Pricing : Dedicated host.

.Explanation: AI models must be trained by Global HealthTech Solutions in order to help diagnose illnesses using patient data and a variety of test results. This is a computationally expensive technique that includes analysing large amounts of data in order to identify patterns and generate precise predictions. Quick training and processing are essential to shortening the period between data collection and model deployment.AI models must be trained by Global HealthTech Solutions in order to help diagnose illnesses using patient data and a variety of test results. This is a computationally expensive technique that includes analysing large amounts of data in order to identify patterns and generate precise predictions. Quick training and processing are essential to shortening the period between data collection and model deployment.

To do this, the training process can be greatly accelerated by using high-performance computing resources on platforms like AWS EC2, such as GPU or TPU-based instances. These instances are made to meet the demands of deep learning and AI workloads in parallel processing, which guarantees quick and efficient model training and prompt and efficient diagnostic support.To do this, the training process can be greatly accelerated by using high-performance computing resources on platforms like AWS EC2, such as GPU or TPU-based instances. These instances are made to meet the demands of deep learning and AI workloads in parallel processing, which guarantees quick and efficient model training and prompt and efficient diagnostic support.

Name: Muhtasir Fuad Nehal

Id : 21000068