hybrid architecture

Hello guys!

We have the following situation,

We are testing Orthanc’s solution to migrate our current DICOM service to this solution.

We have the following architecture, an ubuntu server with Orthanc 1.11.2 with the following plugins:

  • s3-storage
  • postgresql-index
  • postgresql-storage
  • dicom-web

We have this installed in vmware inside the datacenter in our hospital.
On the other hand, we have installed the same architecture in AWS, the server is an ec2 instance and Postgres in RDS. Both architectures store the studies in S3.
Doing tests we have noticed that there is not much difference in performing queries and retrieving studies between both solutions. So we chose to use the AWS architecture for all its benefits.

Testing the response times to retrieve studies from Orthanc are much slower than doing it on our current system, that is logical because the current system runs the Postgres database and the application on the same server in the LAN of our hospital as well storage is on local disk. We currently store 1TB of studies per month.
Is there any way to handle a “hybrid” architecture?

It would be an excellent solution for us to be able to store at least the last 60 days of studies on local disk and after that time the studies are migrated to s3. When making the queries from Osirix within these 60 days the response time would be very fast, and after those days the access to the studies would be towards s3 accepting the slowest response time. Is there any way to be able to do this?

Thank you!

Regards

Hi,

We have stored about 40TB on a on-premise orthanc installation
In my experience you need optimize your postgresql configuration files and store the database on fast storage (ssd / nvme).
We store the dicom files on disk.

I don’t know if it’s possible to do storage tiering within Orthanc but most commercial storage solutions have tiering built in. If you google Linux automated tiered storage you will also find some solutions for linux.

AWS also has tiered storage https://aws.amazon.com/s3/storage-classes/intelligent-tiering/

With kind regards,

Martin

Hi,

I’m just finishing an hybrid storage mode for the S3 plugin. The use case is a hospital with 50 TB of data that can not scale its local storage anymore and who wants to store the least accessed files on S3. There’s only one Postgres DB in the LAN and one Orthanc instance in the LAN as well.

New files are stored on disk and they are moved to S3 through a Rest API call (we assume that each user will have its own rules to move files from disk to S3 so, right now, everyone should script its own rules).

Would that meet your needs ?

Alain.

Thank you both for your answers!

Alain, that’s exactly what we need! our situation is the same as that of that hospital.

when will that version of the plugin be available?

Thank you!

Hi Damian,

The v 2.1.0 has been released and is part of the osimis/orthanc:22.10.2 docker images.

Documentation is available here.

Best regards

Alain.

Hi Alain!

Thanks for the info. We are not using docker at the moment. Can I directly download the plugin from AWS S3 and compile it for a linux ubuntu? would it work with version 1.11.2 of Orthanc?

Will this hybrid architecture be available for Linux? (not docker)

Thank you!!