Is it possible that zoom requests don't cache well since most of the URLs are unique?


When a customer accesses a zoom image on your site, the first request that the viewers make is a context request. This request is a lightweight non-cacheable request to the origin server. Essentially, the image server responds with a unique ID that is tied only to the image being zoomed. If the image is changed, the ID changes as well.

After the context is received, the viewer makes repeatable requests using the context ID as one of the arguments. The requests are repeatable, since the viewer requests the series of tiles that it requires. Another customer zooming in on the same image will likely make the same requests. Even if there is panning on the image, there is a finite number of requests that the viewer makes. All of these requests are cacheable by our CDN provider.

If you look at the requests by two users panning the same image, you see that a number of the requests are the same. The viewer requests large repeatable regions.

In our hosted environment, our content delivery network provider serves over 90% of the traffic.

The e-catalog viewers add a unique id to every map request. So, updates to the image maps don't require a clearing of the Akamai cache.


이 작업에는 Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License의 라이센스가 부여되었습니다.  Twitter™ 및 Facebook 게시물은 Creative Commons 약관을 적용받지 않습니다.

법적 고지 사항   |   온라인 개인 정보 보호 정책