Add the Origin Pull feature to Cloud Files
Thanks for your votes. Cloud Files CDN does support origin pull today, where if content is not located on edge servers, it will fetch from the origin. That being said, it sounds like some of you may have a different definition of “origin pull”. We’ll take your votes and ideas into consideration. Thanks again!
17 comments
-
Richard Stephenson
commented
There definitely seems to be some confusion over the use of the term 'pull' as when I contacted tech support they said it was supported when the pull zone support I am thinking of definitely isn't!
I am thinking of the term pull zone as used by maxcdn, akamei etc.
It makes using CDN very easy straightforward without having to bother getting to grips with yet another API! -
Anonymous
commented
Here's another vote for *true* origin pull (from my servers - I don't want to manage pushing files up to Cloud Files). It's back to Voxel or Amazon Cloudfront until then.
-
James Sutherland
commented
It seems this is actually asking for the "custom origin" feature Amazon added to CloudFront back at the end of 2010 - and yes, I'd be delighted by this addition coming to Rackspace/Akamai as well. ESI would be nice too, but 'custom origin' would be enough to lure me off CloudFront now CNAMEs are there!
-
Michael
commented
It needs origin pull, otherwise it is useless to me, I've signed up for cloud files but I can't see how this will be useful to me presently
-
Scott
commented
Perhaps a version number built into the file names with auto removal of older versions when new files are pulled from the application server. For instance, if the page needs style.css?ver=2 and Cloud Files has style.css?ver=1 or no versions of this file, then Cloud Files will pull ver=2 from the application server, auto delete ver=1 if present and push ver=2 to the CDN. Done and done. :)
-
John Crenshaw commented
I think everyone means the same thing by "Origin Pull". The difference of opinion here is which source is *actually* the origin. For the core CSS/JS/image files on a site the application server is usually where the complete set of up-to-date files can be known to reside. In these cases the Cloud Files server is at best an intermediate cache with a hopefully up to date snapshot of the source files. The issue I have with this is that it forces application servers to know when new content is available and "push" the new data to Cloud Files. This basically defeats the purpose of the origin pull system. The need here is for some pulls (per collection basis?) to be able to reach back all the way to the *real* origin, rather than stopping midway at a location that is (at least in some cases) not actually the origin.
-
MikeTyson
commented
I want this service to. MaxCDN and NetDNA actually supports this. IE they are a full real CDN and a REAL cloud file service. Rackspace have let me down on this.
-
Robert (Jamie) Munro
commented
The origin pull you support is between the CDN and your cloud files storage servers. I think what people are wanting is between the CDN and their own servers, weather based in the Rackspace cloud or elsewhere.
-
phil
commented
why using API and all other sync effort when most cdn offer origin pull
If could file would offer origin pull, I would have signed today ...
-
rgb
commented
I withdrew my votes for this... it's pretty easy with the api to write a script that keeps a directory in sync with a cloudfiles container.
-
Anonymous
commented
This turned out to be a deal breaker for me. Had to cancel CloudFiles account and move to CloudFront.
-
Rajiv Aggarwal
commented
I was expecting that you will provide this feature after partnering with Akamai but was disappointing to learn that is not the case. The lack of this feature is forcing us to use Cloudfront for a number of projects.
-
Stefan
commented
If it is for static content this might help as a workaround:http://fileconveyor.org/
-
notmyname
commented
What you are describing is the ability to set an arbitrary origin server for the Akamai CDN. This is separate from the Cloud Files CDN integration. Currently, what is described here is exactly how Cloud Files + CDN works. That is, the CDN checks the edge cache, and, for a cache miss, pulls the content from your Cloud Files container.
-
Soroush
commented
Cloudfront support customer origin pull and Azure support origin pull of servers running on their cloud 'vm role', I hope that Rackspacecloud, at least, enable setting a cloudserver as origin pull.
-
Rory
commented
Akamai has support for this already, since you're now using Akamai, it shouldn't be too difficult to implement and would be very useful for your users.
-
Denouncer
commented
Undoubtably the next big feature needed on a CDN after CNAME support. Having to continually keep your content synced with the CDN is a nuisance and problematic. Origin pulls resolves this by checking for a local cache, and if not found, fetching from a specified source. i.e. your content server.
