Image Management at Scale
- Pramendra Gupta
Issues
Images are #1 performance blocker
Increasing demand
increase in number of devices and platforms
Reduce operation cost
Reduce infrastructure cost
Requirements
Optimization
such as lossless compression, optimize resolution
Conversion
such as from single source to multiple source (jpeg, webp)
Resizing, Cropping
such as thumbnail, cropping from center and face detection
Features
Fast thumbnail
AI powered
face detection, noise detection, image optimization
Machine Learning
reduce batch process, meta info optimization
Serverless
Rules
Minimize unique file size
Minimize image types
Minimize reinventing wheel
Minimize processing time (TTM)
Maximize stability
Maximize security
MapR Cluster
JP US
S3
CDN
👤
customer
Jenkins
Image
Manager
asynchronous
bidirectional
replication
👤
admin
Glacier
Architecture
Road Map
Phase I
Phase II
Image Management Solution
Akamai, Cloudinary
CDN
Akamai
Image Management Solution
Image Management Application, MapR
CDN
Akamai
References
Image resizing on the fly
Insanely fast JPEG/ JPG thumbnail scaling
Image Management at Ebates

Image management at scale