Cutting Cloud Costs for Security Camera Networks

By Nuruzzaman Faruqui

·

Big organisations watch their perimeters with grids of HD cameras, and streaming all that video to the cloud is expensive. This research makes camera networks smarter about what they send, cutting bandwidth and storage costs.

31.29%less bandwidth
22.43%less storage
HDIoT camera grids

The problem

HD video needs far more bandwidth and storage than text, audio or images. For large organisations with many cameras, cloud network and storage costs climb quickly.

What the research did

We combined Machine Vision at the IoT Edge (Mez), which adjusts video frames according to network delay, with a new Grid Sensing (GRS) algorithm. GRS automatically ranks how much bandwidth each camera in the grid needs, so the system can optimise the whole grid’s traffic.

What we found

Experiments showed a 31.29% reduction in bandwidth and a 22.43% reduction in storage requirements, leading to significant cloud cost savings.

Who benefits

  • Large organisations running perimeter surveillance
  • Cloud and security providers
  • Campuses, airports and industrial sites

Limits and what comes next

Savings depend on scene activity and network conditions. Reducing video data must not reduce the ability to detect real security events, which needs ongoing checks.

The team

I led this research with Sandesh Achar, Sandeepkumar Racherla and Prathyusha Sripathi (Amazon), Vineet Dhanawat (Meta), Md. Monirul Islam, Jia Uddin (Woosong University), Manal A. Othman (Princess Nourah bint Abdulrahman University), Md Abdus Samad and Kwonhue Choi (Yeungnam University).

Frequently asked questions

Why is cloud video expensive?

HD video uses a lot of bandwidth and storage, and cloud providers charge for both.

What is Mez?

Machine Vision at the IoT Edge, a technology that adapts video frames to network conditions.

How much did the system save?

31.29% of bandwidth and 22.43% of storage in experiments.

About the author

Nuruzzaman Faruqui is an award-winning AI educator and researcher who has taught artificial intelligence and machine learning at university level since 2018. A two-time winner of the Excellence in Teaching Award (2023, 2024) and winner of the Best Researcher Award (2024) at Daffodil International University, he is now a doctoral researcher in AI at Monash University, Australia. He has published 53 papers and six books, and writes these articles to make AI research easy for everyone to understand. His teaching · About Zaman · Get in touch

Keep reading