Use the storage architecture guidelines in this section to configure your storage infrastructure for an enterprise medical imaging system.

A pair of Cisco Nexus switches provides redundant paths for IP traffic from compute to storage, and for external clients of the medical imaging system image viewer:

Two Cisco UCS fabrics through different fabric interconnects provide two failure domains. Each fabric is connected to both IP networking switches and to different FC networking switches.

Identical service profiles for each Cisco UCS blade are created as per FlexPod best practices to run VMware ESXi. Each service profile should have the following components:

The VMware ESXi host cluster runs workload VMs. The cluster comprises ESXi instances running on Cisco UCS blade servers.

Each ESXi host includes the following network components:

In healthcare organizations, medical imaging systems are critical applications and well-integrated into the clinical workflows that begin from patient registration and end with billing related activities in the revenue cycle.

The following diagram shows the various systems involved in a typical large hospital; this diagram is intended to provide architectural context to a medical imaging system before we zoom into the architectural components of a typical medical imaging system. Workflows vary widely and are hospital and use- case specific.

The figure below shows the medical imaging system in the context of a patient, a community clinic, and a large hospital.

A typical medical imaging system uses an N- tiered architecture. The core component of a medical imaging system is an application server to host various application components. Typical application servers are either Java runtime- based or C# .Net CLR- based. Most enterprise medical imaging solutions use an Oracle database Server or MS SQL Server or Sybase as the primary database. Additionally, some enterprise medical imaging systems also use databases for content acceleration and caching over a geographic region. Some enterprise medical imaging systems also use NoSQL databases like MongoDB, Redis, and so on in conjunction with enterprise integration servers for DICOM interfaces and or APIs.

A typical medical imaging system provides access to images for two distinct set of users: diagnostic user/radiologist, or the clinician or physician that ordered the imaging.

Radiologists typically use high- end, graphics- enabled diagnostic viewers that are running on high- end compute and graphics workstations that are either physical or part of a virtual desktop infrastructure. If you are about to start your virtual desktop infrastructure journey, more information can be found here.

When hurricane Katrina destroyed two of Louisiana’s major teaching hospitals, leaders came together and built a resilient electronic health record system that included over 3000 virtual desktops in record time. More information on use cases reference architecture and FlexPod reference bundles can be found here.

Clinicians access images in two primary ways:

The medical imaging system can provide image access to a community of physicians or to CIN-participating physicians. Typical medical imaging systems include components that enable image interoperability with other health IT systems within and outside of your healthcare organization. Community physicians can either access images via a web-based application or leverage an image exchange platform for image interoperability. Image- exchange platforms typically use either WADO or DICOM as the underlying image exchange protocol.

Medical imaging systems can also support academic medical centers that need PACS or imaging systems for use in a classroom. To support academic activities, a typical medical imaging system can have the capabilities of a PACS system in a smaller footprint or a teaching- only imaging environment. Typical vendor- neutral archiving systems and some enterprise- class medical imaging systems offer DICOM image tag morphing capabilities to anonymize the images that are used for teaching purposes. Tag morphing enables healthcare organization to exchange DICOM images between different vendor medical imaging systems in a vendor- neutral fashion. Also, tag morphing enables medical imaging systems to implement an enterprise- wide, vendor- neutral archival capability for medical images.

Medical imaging systems are starting to use GPU-based compute capabilities to enhance human workflows by preprocessing the images and thus improving efficiencies. Typical enterprise medical imaging systems take advantage of industry- leading NetApp storage efficiency capabilities. Enterprise medical imaging systems typically use RMAN for backup, recovery, and restore activities. For better performance and to reduce the time that it takes to create backups, Snapshot technology is available for backup operations and SnapMirror technology is available for replication.

The figure below shows the logical application components in a layered architectural view.

The figure below shows the physical application components.

The logical application components require that the infrastructure support a diverse set of protocols and file systems. NetApp ONTAP software supports an industry- leading set of protocols and file systems.

The table below lists the application components, storage protocol, and file system requirements.