Problem: A city wants to deploy 100 security cameras across downtown for traffic monitoring and public safety. Each camera streams H.264 video at 1080p resolution. Calculate the required network bandwidth, choose appropriate connectivity, and estimate monthly data costs.
Step 1: Calculate Per-Camera Bandwidth
H.264 video compression produces variable bitrates depending on motion: - Low motion (static street scene): 2–4 Mbps - Medium motion (normal traffic): 4–6 Mbps - High motion (crowded intersection): 6–8 Mbps
Use 5 Mbps average per camera as a conservative estimate.
Step 2: Calculate Total Bandwidth for 100 Cameras
Using the formula \(B_{total} = N \times R_{cam} \times U\) where \(N\) is camera count, \(R_{cam}\) is per-camera bitrate (Mbps), and \(U\) is utilization:
Peak bandwidth (all cameras streaming simultaneously): \[B_{peak} = 100 \times 5 \times 1.0 = 500\text{ Mbps}\]
Sustained bandwidth (80% utilization factor): \[B_{total} = 100 \times 5 \times 0.8 = 400\text{ Mbps}\]
Step 3: Determine Connectivity Options
Option A: Fiber Optic Backbone
- Capacity: 1 Gbps fiber link
- Sufficient for 500 Mbps peak with 50% headroom
- Cost: $500–1,000/month for 1 Gbps dedicated fiber
- Installation: $10,000–20,000 one-time for fiber runs
- Verdict: Best for centralized control room receiving all streams
Option B: 4G/5G Cellular per Camera
- Capacity: 10–50 Mbps per cellular connection
- 5 Mbps per camera is achievable on 4G LTE
- Cost: $20–40/month per camera for 100 GB data plan
- Total: $2,000–4,000/month for 100 cameras
- Verdict: Prohibitively expensive for 100 cameras
Option C: Wi-Fi Mesh Network
- Capacity: Wi-Fi 5 (802.11ac) = 867 Mbps theoretical per AP
- Real-world: 400–500 Mbps with 10–15 cameras per access point
- Requires: 7–10 outdoor APs with fiber/Ethernet backhaul
- Cost: $300 per AP x 8 APs = $2,400 hardware + fiber backhaul
- Monthly: $500–800 for backhaul links
- Verdict: Viable if fiber backhaul is available to AP locations
Chosen Solution: Fiber to central hub + Wi-Fi mesh for camera connectivity
Step 4: Calculate Monthly Data Volume
\[D_{month} = \frac{B_{total} \times 10^6 \times 86400 \times 30}{8} \quad [\text{bytes}]\]
Cameras record 24/7: - Data per camera per day: 5 Mbps x 86,400 s/day = 432,000 Mb / 8 = 54 GB/day - Data per camera per month: 54 GB x 30 days = 1,620 GB/month = 1.62 TB/month - Total for 100 cameras: 1.62 TB x 100 = 162 TB/month
Step 5: Calculate Cloud Storage Costs
Option 1: AWS S3 (Standard tier)
- Storage: $0.023/GB/month
- 162 TB = 162,000 GB x $0.023 = $3,726/month storage
- Retrieval + bandwidth out: $0.09/GB (if downloading evidence footage)
Option 2: Intelligent Tiering (7-day hot / 23-day archive)
- Hot storage (7 days): 162 TB / 30 x 7 = 37.8 TB @ $0.023/GB = $869/month
- Cold storage (23 days): 124.2 TB @ $0.004/GB (Glacier) = $497/month
- Total: $1,366/month (63% savings vs standard)
Option 3: Local NAS with Cloud Backup
- Local: 200 TB NAS ($8,000 hardware) stores 1 month + live buffering
- Cloud backup: Upload alerts/incidents only (5% of footage) = 8.1 TB/month @ $0.023/GB = $186/month
- Total: $186/month + $8,000 one-time (payback in 21 months vs full cloud)
Step 6: Total Cost of Ownership (5-Year Analysis)
Fiber + Wi-Fi Mesh + Local NAS:
- Hardware: $2,400 (APs) + $8,000 (NAS) + $3,000 (switches/cabling) = $13,400
- Installation: $15,000 (fiber runs, AP mounting)
- Fiber backhaul: $600/month x 60 months = $36,000
- Cloud backup: $186/month x 60 months = $11,160
- NAS replacement (year 3): $8,000
- 5-year total: $81,560
All-Cellular:
- Hardware: $5,000 (100 x $50/camera LTE modem)
- Cellular service: $3,000/month x 60 months = $180,000
- Cloud storage: $1,366/month x 60 months = $81,960
- 5-year total: $266,960 (3.3x more expensive than the fiber solution)
Step 7: Bandwidth Optimization Strategies
Optimization 1: Motion-triggered recording
- Record only when motion is detected (vs continuous 24/7)
- Reduces data by 60–80% for low-traffic cameras
- New bandwidth: 500 Mbps x 0.3 = 150 Mbps average
- Storage: 162 TB x 0.3 = 48.6 TB/month (saves $2,580/month on cloud)
Optimization 2: Edge AI analytics
- Cameras detect events locally (license plate recognition, person counting)
- Upload only metadata + 10-second clips when events occur
- Reduces upload by 95%: 162 TB x 0.05 = 8.1 TB/month
- Bandwidth: 500 Mbps x 0.05 = 25 Mbps (enables a cheaper internet tier)
Optimization 3: Adaptive bitrate
- Lower resolution/bitrate during nighttime (less scene detail visible in darkness)
- Nighttime (8pm–6am, 10 hours): 2 Mbps vs 5 Mbps daytime
- Weighted average: (5 Mbps x 14 h + 2 Mbps x 10 h) / 24 h = 4.08 Mbps
- Saves ~18% bandwidth and storage
Final Optimized Design:
- Fiber backbone + Wi-Fi mesh
- Motion-triggered recording with 30-second pre-buffer
- Edge AI for event detection (upload events only)
- Adaptive bitrate (day/night)
- Result: 48.6 TB x 0.05 = 2.43 TB/month cloud storage
- Cloud cost: 2.43 TB x 1,024 GB/TB x $0.023/GB = $57/month (vs $3,726/month unoptimized)
Key Insights:
- Bandwidth scales linearly with cameras: 100 cameras = 10x the bandwidth of 10 cameras
- Storage costs exceed bandwidth costs for always-on video systems
- Edge processing (motion detection, AI) provides 20–50x cost reduction vs cloud-only
- Cellular is 3–5x more expensive than fiber for fixed deployments, but necessary for mobile applications
- Design for peak, not average: the network must handle all cameras streaming simultaneously during incidents