Performance test DICOM forward to Orthanc server setup

Performance test [login to view URL] proxy infrastructure Background: CMRAD is deploying a component to local hospitals which acts as a proxy and anonymizer/pseudonymizer of DICOM data being sent from the hospital to the central CMRADCORE orthanc instance. The proxy installation is a barebone orthanc server running either on windows or linux, managed by the hospital IT department. It stores nothing more than its own configuration making sure the component is non-vital, stateless and can be replaced at any time and does not need robust protection (backups, continuous availability, etc). Objective: Make sure that our setup can scale gracefully and manage the parallelism which will occur in an individual proxy instance and within the proxy-to-CMRADCORE network. Ensure that the upload guarantees data integrity end to end between Proxy and CMRADCORE (no studies are broken apart, all files are transferred, the study either arrives intact or fails completely - there is no in-between option) An average bandwidth of minimally 20Mbps is achieved at each proxy site during the full load test. Setup: Install 10 proxy instances in non-AWS-frankfurt location(s). 5 on windows, 5 on linux. configure them with CMRADs standard proxy scripts and point them to our CMRADCORE DEV server in AWS Frankfurt Prepare 100 DICOM study test cases for sending through this infrastructure. 10 studies per proxy. Use a mix of different source (Modalities, filesizes, file counts, manufacturer, etc.) to get a good sample of what the industry uses. Execution: In an iterative fashion slowly increasing the load, until we are sending in parallel all 10 studies on all 10 proxies: 1. Record bandwidth used during execution for each proxy instance (both the bandwidth and the % of total available) 2. Record the CPU and memory used by the Proxy orthanc instances during each execution 3. Detect failures in the proxy anonymization and forwarding/upload and suggest changes to the CMRAD scripts or Orthanc installations (both proxy and CMRADCORE) to correct these. 4. Detect bottlenecks appearing either at the Proxy or in CMRADCORE and suggest changes to remediate the bottleneck. (adding hardware is the LAST resort to such fixes, first we try to be smarter) 5. Advice on other robustness/scalability improvements to the setup found during the execution. Are there autoscaling or multi-orthanc advice to be given to assure scalability? The effort will be lead by the Freelancer, however CMRAD will assist with granting access and source information, etc. CMRAD guarantees to delete all test data after the test from our servers. If you are interested, let's setup a time to discuss the requirement and how to best execute it before you create a quote. And let's be smart about it, how to do this efficiently.