IDC MVP hosted collections will be stored in storage buckets accessible to the users via the Requester Pays feature.
We will also be hosting those collections in a DICOM store accessible via DICOMWeb interface. However, the main purpose of that DICOM store in the MVP will be to support image visualization and not replicating of data at a user resource. It is currently the case that DICOM stores do not support requester pays. IDC MVP will be using a throttling proxy to provide free restricted DICOMWeb access to the hosted content for the purposes of limited image viewing, limiting maximum download size per IP and total per day.
Post MVP, we plan to explore if we need to implement a less restrictive solution to support DICOMWeb access to IDC hosted content.
If you would like to see a mechanism that would allow unrestricted DICOMWeb access to IDC data content (noting that unrestricted access will come at some cost to the user downloading the data), please describe your use case here and upvote this feature.
It would be nice to have free public DICOMweb access to IDC-hosted data so that data sets could be retrieved for demos, tutorials, etc. via an open, vendor-neutral interface.
Thanks for the feedback. The current IDC DICOMWeb interface is throttled by a proxy to prevent download of the data collection out of the cloud, as that involves egress charges. One possible solution would be to provide researchers with their own proxy server projects that can access our backend Google Healthcare DICOMWeb endpoint, and attach the researcher’s billing account to the project (in a manner similar to the way Broad Terra handles user’s Google billing).
I think the use cases @lassoan is thinking about might be addressable if we create a public read only dicomweb endpoint that only has a few dicom sample files (maybe one from each collection or something). That would limit the costs since there wouldn’t be much to download. We could still throttle it just to be sure a billion people don’t all try to access the same day.
The reason to do this is that we (IDC) are trying to promote the use of the dicom standard but for software testing or demos the current methods are to either pay for a google store (and deal with logins) or spin up a bunch of (generally insecure) web infrastructure for dcm4chee or similar.
Yeah, a tiny example endpoint with a very limited set of data is a great idea. I would not even put a proxy on it, since that will probably cost more money than letting everybody have a go at it.
That’s great, we can start collecting some images. @lassoan and I discussed today that having some dicom conformance examples freely available for testing would be very helpful.
At a future meeting we can discuss how best to set things up.
This thread was not updated, but the statement above is no longer accurate effective November 2021.
Effective ~ November 2021, and as of today, thanks to the partnership of IDC with Google and Amazon Web Services public data program, egress of IDC data from storage buckets (either Google Storage or AWS S3) is free of charge, and does not require sign in.
In light of the increasing interest in DICOMweb access, in particular, for digital pathology applications, we amended IDC proxy policy to now allow egress without restricting to IDC viewer only. The per-IP daily quota still applies. Please see the updated proxy policy here: Proxy policy - IDC User Guide. I hope this helps those users interested in using DICOMweb for accessing IDC data!