Welcome Guest, Not a member yet? Register   Sign In
Caching and MD5 collisions
#1

[eluser]slowgary[/eluser]
Hi all,

I developing a multi-site application and am looking ahead at caching my pages. I looked into the output class and see that cached files are saved using an MD5 hash of the full URL of the cached page. This is nice since I may have similar segments on multiple sites, like so:
Code:
http://www.first-site.com/products/baby/rubber-baby-buggy-bumpers
http://www.other-site.com/products/baby/rubber-baby-buggy-bumpers
My concern now is that since the MD5 algorithm can generate collisions, is there any possibility that any of the pages could have a duplicate cache file name? It would be VERY bad if a visitor hit one of our sites and sees a page from a different site.

Can anyone confirm if this is even a remote possibility? I've read a little about how people have created duplicate hashes from different files, but I'm not sure enough about the technical details to know if this could happen from 2 similar URLs.
#2

[eluser]WanWizard[/eluser]
This will make an interesting read: http://www.accessdata.com/media/en_us/pr....en_us.pdf
#3

[eluser]slowgary[/eluser]
Thanks WanWizard!

My gut feeling is that collisions in this case will NEVER happen, as it seems that very specific circumstances are required accomplish such a collision. This is how I felt before I posted as well.

Because my understanding of this topic is still limited (even after reading the article), it would still be great if someone more knowledgeable than myself could explain why it will never happen with something like 2 similar URLs.




Theme © iAndrew 2016 - Forum software by © MyBB