A New Kind of Backend Image Storage
We built LIS (Large Image Storage) — a backend system designed to run in your private cloud (AWS, Azure, GCP) that dramatically reduces image storage costs while serving images on demand. LIS works with all major image types (eg: JPEG , AVIF , WEBP , etc )
LIS is not a CDN or edge service. It sits alongside your existing object storage (S3, GCS, Azure Blob) and replaces raw image files with a compressed representation that's orders of magnitude smaller. When your application requests an image, LIS decompresses and returns it in under 200 milliseconds.
Key capabilities:
- Lossless or lossy mode — choose based on your quality requirements
- GeoTIFF support — works with geospatial imagery, including returning slices of an image on request
- Drop-in replacement — your application requests images the same way it always has
The Numbers: 1 Petabyte of Images
Assume a corpus of 1 petabyte of raw images — roughly 200 million images at 5 MB average (typical high-res JPEG). Here's how LIS compares to standard formats in lossless mode:
| Method | 1 PB Raw Becomes | Reduction | Cost @ $0.023/GB/mo |
|---|---|---|---|
| JPEG (q85) | 100 TB | 10x | $2,300/mo |
| WebP | 67 TB | 15x | $1,541/mo |
| AVIF | 50 TB | 20x | $1,150/mo |
| JPEG XL | 40 TB | 25x | $920/mo |
| LIS | ~190 GB | ~5,200x | $4.37/mo |
That's not a typo. LIS compresses 1 PB of images down to approximately 190 GB — a 5,200x reduction. Monthly S3 storage drops from $2,300 to under $5.
At Scale: 100 Petabytes
For organizations managing massive image libraries — satellite imagery providers, medical imaging archives, e-commerce platforms — the savings scale linearly:
| Method | 100 PB Raw Becomes | Monthly S3 Cost |
|---|---|---|
| JPEG (q85) | 10 PB | $230,000/mo |
| WebP | 6.7 PB | $154,100/mo |
| AVIF | 5 PB | $115,000/mo |
| JPEG XL | 4 PB | $92,000/mo |
| LIS | ~19 TB | $437/mo |
At 100 PB, you're saving over $229,000 per month compared to JPEG — or $2.7 million per year.
What LIS Does Not Do
LIS reduces storage costs, not egress. When a client requests an image, LIS decompresses it and serves the full image over the network. If your bottleneck is bandwidth rather than storage, LIS isn't the right tool — look at our Web Infrastructure solution instead.
Get Started
LIS deploys as a container in your cloud environment. We handle setup and tuning for your specific image corpus.
Contact us to discuss your image storage requirements, or check our API documentation for integration details.