Extreme Image Performance – Finding the Best Way for Image Delivery
Invitation to a video presentation
At recent Meetups on Web Performance, I had the enjoyable opportunity to talk about one of the biggest challenges in web and mobile application delivery — delivering ever-larger images faster. Finding the best possible optimization has a dramatic impact on the user experience, especially for small mobile devices with high-resolution screens and wireless connections.
You probably are aware that images increasingly make up the lion’s share of web application payload, and according to HTTParchive.org the average page has grown in size by two-thirds in just two years.
|Average page size||Average images on average page|
|May 15 2012: 1059 KB||Images: 672 kb|
|May 15 2014: 1772 KB||Images: 1,120 kb|
These recent talks at Web Performance Meetups have centered on the effectiveness and tradeoffs of the different approaches to image delivery, where Instart Logic has gone beyond known techniques and created new ones.
In the video, I cover—and compare:
- the basics of individual image optimization like removing metadata, using browser-specific formats, and resizing at the server for each requesting device
- progressive Image streaming
- split progressive image streaming
- streaming images in “stages of resolution” without increasing payload
- choosing the best optimization approach in each case
Without getting completely into the details of our secret sauce, I will just explain that on the server side, Instart Logic treats every web application, web page, and image as a unique player in the overall user experience. There is no one-size-fits-all concept operating; we spend some server cycles to determine the optimal way to treat every single image going to every end user. And it works — really well.
Rather than spelling it out and expecting you to read through it here, I invite you to dive into the video and watch and listen.
I think this video will change, in some way, how you think about content and applications that you see on mobile devices, and about the potential for highly visual content.
If you have questions or feedback about this topic, I’d be really pleased to hear from you at email@example.com.