Orthanc as a research VNA

Hi,

We are looking research VNA for our university research department

Can we use Orthanc as a research VNA?
Can we add users and assign users to different project / patients?
Is it possible to setup different / multiple storage points?

We are also looking at Xnat as a possible solution

With kind regards,

Martin Dijkhuizen
University of Groningen.

Hi,

Yes Orthanc is made for PACS research,

No you can’t assign User to different Projet/patient, at least not in Orthanc itself, but that is something i’m implementing in OrthancToolsJS (software built on the top of orthanc).

No there is one single storage point for Orthanc. Then you can export your data with not limit using Orthanc APIs if you want to make copy in remote storage.

Best regards,

Salim

Hi Salim,

Thank you for your reply.
I think I didn’t state my storage question clear enough. What I meant was can Orthanc handle multiple filepaths (where the dicom files are stored)
for example /storage/storage1 , /storage/storage2 to separate the divide the storage OI and load.

Do you have a idea when the implementation of the project/ user management will be ready?

Best regards,

Martin

Hello,

Regarding this part of your question:

I think I didn’t state my storage question clear enough. What I meant was can Orthanc handle multiple filepaths (where the dicom files are stored)
for example /storage/storage1 , /storage/storage2 to separate the divide the storage OI and load.

Orthanc does not provide this feature out-of-the-box, but you have multiple possibilities:

  • Under GNU/Linux, use Linux LVM in order to aggregate a set of physical partitions into a single logical volume, then use this volume as the “OrthancStorage” configuration option of Orthanc (this is the most direct solution):

https://docs.min.io/docs/distributed-minio-quickstart-guide.html

  • If you are a DB expert, you might also consider using the PostgreSQL storage area plugin to store the DICOM files, and configure PostgreSQL to use several volumes for its database (using tablespaces, horizontal sharding…), but this is not straightforward:

https://book.orthanc-server.com/plugins/postgresql.html

  • You can develop your own storage area plugin for Orthanc by using the “OrthancPluginRegisterStorageArea2()” function in the Orthanc SDK, then share back your contribution to the community:

https://sdk.orthanc-server.com/group__Callbacks.html#ga801b92161b0bfbfa6a3dd8862b09036a
https://hg.orthanc-server.com/orthanc/file/Orthanc-1.9.3/OrthancServer/Plugins/Samples/StorageArea

Regards,

Sébastien-

As a complement to my previous answer, you might also have an interest in having a look at distributed filesystems such as Lustre and Ceph, either as mount points for the “OrthancStorage” folder, or through as a custom Orthanc plugin (the S3 plugin might already be compatible with Ceph, though I am unsure):
https://en.wikipedia.org/wiki/Lustre_(file_system)
https://en.wikipedia.org/wiki/Ceph_(software)

https://docs.ceph.com/en/latest/radosgw/s3/

Sébastien,

Thank you for your reaction. I hoped for a more windows like answer as I am windows administrator :slight_smile:
I’m going to check the storage solutions out.

Best regards,

Martin

The Microsoft Windows equivalent of LVM is called “Storage Spaces”:
https://docs.microsoft.com/en-us/windows-server/storage/storage-spaces/overview

The other solutions I mentioned are also usable under Microsoft Windows. In particular, instructions to download the relevant binaries of MinIO are available on the following page:
https://docs.min.io/docs/minio-quickstart-guide.html

I will not elaborate further, as I personally don’t use proprietary software such as Microsoft Windows.

Sébastien-

Hi Sébastien,

Thank you for your answer again. My last question. Do you have experience with large datasets on Orthanc?
I’ve read in the Orthanc docs that Orthanc supports up to 15TB storage and users report even around 28 TB Storage.

The idea is that we at least need to host 60 TB and this will grow.

Best regards,

Martin

Hello,

It is incorrect to say: “Orthanc supports up to 15TB storage”.

The correct assertion is: “Nobody in the community has ever reported a strong limitation of Orthanc with respect to large datasets”, provided the recommended setup is used:
https://book.orthanc-server.com/faq/scalability.html#recommended-setup-for-best-performance

There is nothing in the source code of Orthanc that would prevent the DICOM database from growing to dozens of TB. The pressure would actually be more on the database engine (note that PostgreSQL is highly recommended) than on Orthanc itself, especially since release 1.9.2 introduced the possibility of having multiple Orthanc servers sharing the same database (which allows to deploy a cohort of Orthanc servers in order to lighten the workload on each of the servers).

We need testimonials from the community (including from your team) in order to refine this information.

In either case, you really should be considering switching from Microsoft Windows to GNU/Linux if you are considering such a large-scale deployment.

Regards,
Sébastien-