When delivering content directly from an EC2 Windows instance, the network interface can indeed become your primary bottleneck. Standard EC2 instances typically have network performance that scales with instance size:

// Example: Checking network bandwidth in Windows EC2
Get-NetAdapter | Select-Object Name, LinkSpeed

// Typical output for t3.medium instance:
Name     LinkSpeed
----     --------
Ethernet 100 Mbps

Amazon S3 operates fundamentally differently from EC2 in terms of bandwidth:

• Automatic scaling to handle virtually unlimited requests
• Distributed architecture across Availability Zones
• No single-point bandwidth limitation

For scenarios where same-origin policy prevents S3 usage:

// Option 1: Reverse proxy configuration (Nginx example)
server {
    listen 80;
    server_name yourdomain.com;
    
    location /videos/ {
        proxy_pass https://your-bucket.s3.amazonaws.com/;
        proxy_set_header Host your-bucket.s3.amazonaws.com;
    }
}

// Option 2: Signed URLs with temporary access
// (AWS SDK for JavaScript example)
const AWS = require('aws-sdk');
const s3 = new AWS.S3();

const params = {
    Bucket: 'your-bucket',
    Key: 'video.mp4',
    Expires: 3600 // URL expires in 1 hour
};

s3.getSignedUrl('getObject', params, (err, url) => {
    console.log('Temporary URL:', url);
});

For 100 concurrent users streaming 5MB videos:

Use this PowerShell script to measure actual throughput:

# Windows EC2 Network Test Script
$testFile = "C:\test\100MB.bin"
$webServer = "http://your-server.com/test"

Measure-Command {
    Invoke-WebRequest -Uri $webServer -OutFile $testFile
} | Select-Object TotalSeconds

# Calculate MB/s: (100 / TotalSeconds)

For dynamic content that requires same-origin access:

// PHP example for hybrid delivery
function getVideoStream($videoId) {
    if (shouldUseS3($videoId)) {
        header("Location: " . generateS3SignedUrl($videoId));
    } else {
        readfile("/local/videos/" . $videoId);
    }
}

// Helper function to determine delivery method
function shouldUseS3($videoId) {
    $popularVideos = ['intro.mp4', 'demo.mp4'];
    return in_array($videoId, $popularVideos);
}

Your observation about the 100Mbps network limit on your Windows EC2 instance is correct. AWS EC2 instances have instance-type-dependent bandwidth caps. For example:

# AWS CLI command to check instance network specs
aws ec2 describe-instance-types \
--instance-types t3.large \
--query "InstanceTypes[].NetworkInfo" \
--output table

Typical bandwidth allocations:

• t3.nano: Up to 5Gbps (burst)
• t3.large: Up to 5Gbps (burst)
• m5.large: Up to 10Gbps
• c5n.4xlarge: Up to 25Gbps

With your scenario (100 users × 5MB × 20 videos):

# Bandwidth calculation
total_bandwidth = users × (video_size × videos_per_user) × 8 bits/byte
                = 100 × (5MB × 20) × 8
                = 80,000Mbps (or 80Gbps) - theoretical max

In reality, you'd need to consider:

• Concurrent users (not all 100 may be active simultaneously)
• Video compression (H.264/HEVC can reduce sizes)
• Progressive loading (users rarely download all 20 at once)

S3 provides:

• No per-instance bandwidth limits
• Multi-AZ redundancy
• Built-in CDN via CloudFront

For your AJAX CORS issue, consider:

// Example S3 CORS configuration
[
  {
    "AllowedHeaders": ["*"],
    "AllowedMethods": ["GET"],
    "AllowedOrigins": ["https://yourdomain.com"],
    "ExposeHeaders": []
  }
]

Workaround for same-domain requirement:

// Nginx reverse proxy config example
server {
    listen 80;
    server_name yourdomain.com;

    location /videos/ {
        proxy_pass https://your-bucket.s3.amazonaws.com/;
        proxy_set_header Host your-bucket.s3.amazonaws.com;
    }
}

To monitor EC2 network usage:

# CloudWatch metrics to watch
aws cloudwatch get-metric-statistics \
--namespace AWS/EC2 \
--metric-name NetworkOut \
--dimensions Name=InstanceId,Value=i-1234567890abcdef0 \
--statistics Average \
--period 60 \
--start-time 2023-01-01T00:00:00Z \
--end-time 2023-01-01T23:59:59Z

For immediate relief:

1. Enable EC2 Enhanced Networking (requires HVM AMI)
2. Use Elastic Network Adapter (ENA) drivers
3. Consider placement groups for low-latency

When ready to migrate from EC2:

# CloudFront+S3 setup via Terraform
resource "aws_cloudfront_distribution" "video_dist" {
  origin {
    domain_name = "${aws_s3_bucket.videos.bucket_regional_domain_name}"
    origin_id   = "S3-${aws_s3_bucket.videos.id}"
  }

  enabled             = true
  default_root_object = "index.html"

  default_cache_behavior {
    allowed_methods  = ["GET", "HEAD"]
    cached_methods   = ["GET", "HEAD"]
    target_origin_id = "S3-${aws_s3_bucket.videos.id}"

    forwarded_values {
      query_string = false
      cookies {
        forward = "none"
      }
    }
  }
}