Enterprise Imaging Performance Benchmarking

Systems performance is an effective way to benchmark the value of your Enterprise Imaging and teleradiology workflows. View our webinar recording of Dmitriy Tochilnk, President and CTO, Dicom Systems and Trevor Walker, Principal Systems Analyst, Stanford Health Care cover strategies for evaluating enterprise imaging with real, measurable impact on network performance, clinical workflows, and interoperability.


The excerpt from this webinar examines:

  • Managing DMWL Proxy and image routing
  • Handling traffic based on port redirection
  • Benchmarking technical systems for clinical results
  • Horizontal scaling vs. vertical scaling
  • Strategies for alleviating connectivity to PACS/MIMPS by leveraging the cloud

Note: This conversation has been edited for length and clarity.

Dicom Systems is known for its signature Unifier platform, as well as de-identification of medical images, integrating Artificial Intelligence (AI), intelligent, rules-based routing, fixing imaging workflows and optimizing teleradiology. Our work is recognized globally by top healthcare enterprises, government agencies, and partners for next-generation enterprise imaging interoperability. Thanks to our considerable experience we have gained access to insights and the practical concerns of those working in healthcare information infrastructure. One of our most notable clients is Stanford Healthcare, who have been kind enough to share their experiences with us in this webinar.

Stanford Healthcare Overview

Exterior of Stanford Health Care building. Source: Stanford Health Care
Stanford Health Care is a leading academic health system, delivering clinical innovation across its inpatient services, specialty health centers, physician offices, virtual care offerings, and health plan programs. As the only Level I trauma center between San Francisco and San Jose, Stanford Health Care provides compassionate, coordinated care, personalized for the unique needs of every patient. In November 2019, Stanford Health Care’s new 824,000-square-foot state-of-the-art hospital opened, bringing its total bed count to 605 and total operating rooms to 79. With nearly 3,023 medical staff and more than 1,400 residents and fellows, Stanford Health Care is committed to providing the highest quality care to patients and their families.

Data Benchmarking Background

Benchmarking is a way to measure the performance of an organization’s products, services, processes, and operations. In IT, benchmarking is defined as the comparison between vendor performance and designated benchmark organizations or indexes. When establishing parameters for benchmarking, theoretical performance boundaries are only marginally useful. While it couldn’t hurt to know how a system would perform under the most ideal, controlled conditions, too many variables influence performance in a real-world context to provide a cookie-cutter estimation or range. Our customers’ unique infrastructures invariably present idiosyncrasies that won’t fit within a standardized profile.

Stanford Healthcare was eager to check that the I/O and network speed would stand up to the traffic of its enterprise. During the webinar, Stanford’s Principal Systems Analyst Trevor Walker shared an overview of the Stanford Healthcare technology stack and the standards-based approach used to deploy vendor applications.

Trevor and Dmitriy went on to discuss the best practices for data benchmarking.
Both agreed that during implementation and the go-live, the importance of data benchmarking cannot be overstated. When benchmarking, consistency in measuring and judging results from the system in relation to requirements keyFor example, if you need to move an exam with a thousand slices CT with a full prior exam there’s a SATA disk with limited I/O throughput, so the transfer of the file takes over five minutes on disk speed alone, assuming an empty queue. Based on these findings, you might reasonably conclude an upgrade to a solid state drive is in order.
Be sure, in your testing process, to employ different tools to benchmark I/O memory and CPU depending on the time of day. This benchmarking data during pre- and post-go live is of great use in reference when checking down the line to make sure performance hasn’t degraded– for example, six months after deployment. Replicating tests to ensure that you can accurately pinpoint any problems in the system will make everyone’s jobs more straightforward.
The Dicom Systems Unifier is compatible with testing and staging scenarios for virtual machines as well as with physical ones.

Dicom Systems Unifier

Our Unifier platform is robust, flexible and lends itself perfectly for performance benchmarking. To repurpose an analogy originally used by our webinar guest Trevor Walker, we deliver the equivalent of an essential utility such as water at the right pressure, right temperature, and right time, optimizing data flow. We find that the Unifier is highly adaptable to different organizations and structures, but Stanford Healthcare mainly utilizes the Unifier to manage the various virtual machines they use for functions ranging from image routing, to modality worklist proxy, and QR proxy.

Data Benchmarking in Action

Stanford Healthcare puts a lot of stock in being able to assess detailed metrics in real time. With multiple imaging systems actively accessing their PACS/MIMPS and other systems, they need maximum visibility into their traffic. The Dicom Unifier layer in-between provides that for Stanford’s imaging support teams and troubleshooting teams. “Hey, my images were sent from the modality and they are not in the PACS/MIMPS.” Trevor Walker explained that when a query is made, the first step in their investigation is to check the Dicom Unifier, assessing the statistics and logs that the Unifier has recorded. For example, comparing the general performance to specific examples of performance can help Trevor’s team identify where issues of data transfer may be related to issues other than software limitations. It also enables the Stanford enterprise to compare peak and off-peak traffic, weekend traffic, and other valuable data comparisons. By using this granular access to networking data to identify bottlenecks in the system, any health IT specialist can massively improve their ability to optimize the system and eradicate inefficiencies.
Trevor emphasized, “Without the Dicom Unifier layer in the middle, we have very little visibility into our imaging traffic.This is an invaluable tool for our troubleshooting.”

Smart Image Routing

One of the Dicom SystemsUnifier platform’s other strengths is smart image routing. Systems like Stanford Healthcare, where there are discrepancies between the form of patient ID the archived PACS/MIMPS uses and the format the Radiology Information System supplies, leverage the Unifier platform to support a high volume of tag morphing. The smart imaging routing and retrieve proxy make any type of migration very straightforward for such enterprises.

Everyday Data Movement

With an organization as large as a Level 1 Trauma Center, many different departments including research, pediatrics, and clinical, require large quantities of data movement. Deploying the Unifier platform allows for the pressure on the PACS/MIMPS from ingestion and direct query retrieval to be offloaded. Especially when using a legacy PACS/MIMPS from 10 or 15 years ago to shift datasets for use with contemporary AIs, machine learning algorithms, and the like that need to be collaborated on by external research institutions, the antiquated system can struggle. The Unifier platform means that the data can be anonymized and pushed securely in transit and while at rest. (If the data needs to be identifiable, this can also be achieved after successful delivery.)

Future Planning for Peak Performance

Before opening the webinar to questions, some of Stanford Healthcare’s future plans and challenges were discussed. They included secure cloud storage, de-identification of images for biomedical research, and alleviating direct connectivity to PACS/MIMPS from ancillary imaging systems.
Of course, the larger the center and the more data that is being stored, the more cumbersome it becomes to migrate. Pulling individual studies out and pushing them to the cloud is manageable, but not at a rate that is fast enough for clinical or research purposes. Using the Unifier platform as the front end for modalities, with other secure, well-known technologies, allows for the possibility of avoiding migration.Instead, PACS/MIMPS replacement or enhancement of existing PACS/MIMPS can become an option.

Dicom Unifier and Enterprise Imaging Infrastructure Considerations For Performance Benchmarking

The Dicom Systems Unifier platform is Linux-based, which eliminates many of the Windows-related licensing, configuration and performance challenges. By deploying self-contained Linux-based appliances, the ecosystem is far more resilient and redundant than applications that depend on Windows for availability.
From an infrastructure perspective, the first element to consider is the choice between hardware deployment vs. virtualized environment (VMware of Hyper-V).
For physical servers, Dicom Systems typically deploys the Unifier platform on standard Supermicro hardware. It is important for our support engineers to rely upon a well-known, predictable spec. It is also much easier to ship standard hardware for smooth and quick field replacements rather than support a multitude of hardware platforms.
For virtualized appliances, Dicom Systems has a standard recommended configuration. Virtualized environments have quickly become a preferred method, as it is far quicker and easier to manage and update the resources of a VM than to upgrade or replace a physical server.
Additional Benchmarking Resources
As performance is paramount to our platform, we regularly benchmark the performance of our applications on various IT infrastructure.

We benchmarked the platform using applied infrastructure answers in collaboration with one of our original customers and high-volume Teleradiology end user who processes over 1.5 million studies a year and performed a series of tests to determine how transformation and workflow rules affect performance. The punchline, based on real-world benchmarking, is described in our White Paper titled “Can a single DICOM router process and route 4.1 Billion images per year?” (Click here to download the full analysis)

If you’d like to learn more about what the Unifier platform could do for your enterprise.
Request a Demo