RSNA 2011: Observations by an Architect (November 27, 2011)

This past Sunday I had the opportunity to attend the annual RSNA trade show again. As in past years, most of my time was spent in the research exhibits, because from the perspective of an architect it is always interesting to not only see the move of technologies to the commercial space, but get a better understanding of what is inside the box. In the commerical space this year, I did happen to chance upon some whiteboard scribblings for the first time that were presumably intended to depict architecture to potential buyers (RSNA is heavily sales oriented), but it really just scratched the surface.

Unfortunately, due to my travels during the prior week of Thanksgiving I did not have as much energy available as in past years to speak with researchers, but I did manage to get a good demonstration from the researchers behind an exhibit entitled "Mobile Cloud-Computer-aided Diagnosis System for Colon Cancer Screening with Motion-based Virtual Colonoscopic Navigation". Somewhat reminiscent of one of my favorite exhibits from last year, "PACS 3D: Utilizing Cutting-edge Technologies to Create a Fully Interactive 3D PACS Interface", albeit the showcase this time revolved around a Microsoft Xbox 360 + Kinect rather than a Sony Wii, and no glasses were required.

This exhibit was much more sophisticated than its seeming counterpart last year, consisting of not just this Kinect user interface, but an iPad user interface as well as software hosted on an Intel cloud that virtually cleanses (prepares for analysis and viewing) the colon images served to these devices. As the researchers explained to me, this virtual cleansing took approximately 30 minutes prior to moving to the cloud, and now takes about 2.5 minutes. Their goal is to further reduce this time to about 20 seconds, which is the average time it actually takes to scan a patient to create the images.

After successfully performing extensive virtualization work for a client CTO this past year, and literally just completing a reading of "The Cloud at Your Service", by Jothy Rosenberg and Arthur Mateos, I was interested in hearing what these researchers thought about cloud security. They replied in the negative to my initial question as to whether a private cloud had been utilized, so I inquired as to whether they chose a public cloud perhaps because security is probably best left to specialists. Their reply was that security is not a concern of theirs, just the research, which makes sense, as long as the topic is revisited en route to the commercial space.

Cloud computing is actually one theme at RSNA 2011 that predictably intensified from last year. As expected, however, the definitions of "cloud computing" were all over the map. For example, one exhibit provided the graphic below that describes different means by which images can be accessed by users. According to the given definition, cloud computing "utilizes powerful servers which provide centralized data management and storage capabilities to facilitate collaboration and remote image access."

Not a very accurate definition. Clouds typically consist of commodity servers, which help enable cost efficient scalability, rather than "powerful servers". In addition, while it may be true that clouds provide centralized data management and storage, this aspect does not speak to any of the key differentiators that clouds provide. The table below, from the same exhibit (as always, sources are anonymous), seemingly provides additional information behind each of the image sharing methods in the above graphic, but is a bit awry.

Focusing on cloud computing, this table indicates, for example, that (1) cloud architecture is third-party, which assumes a non-private cloud, (2) again, "several powerful central servers" are involved, (3) something is secure,  not sure what – for example, that data at rest (data persistence in a database) and/or at motion (client communication with the cloud) is secure, and (4) cloud scalability is "limited by server power and storage space", which is always a limitation regardless of solution, but does not recognize one of the central benefits of using a cloud.

However, despite its flaws, this same exhibit reminded me one of the realities in the healthcare software space regarding competition: "Healthcare enterprises expend considerable effort to keep patients within their systems. Internally accessible integrated medical records that are not easily duplicated form a key competitive advantage. Open data networks that allow transfer of imaging or other records outside their enterprises erode this advantage, so competitive enterprises may not devote their scarce IT resources to implementing electronic image transfer." As someone who has experience in both the public healthcare space, and most recently, the private healthcare space, the former promoting standardized records, this is a dilemma with which I have needed to accomodate as a consultant. It will be interesting to see the degree to which this dilemma erodes in coming years due to available government funding.

Other notable common RSNA 2011 themes across commercial exhibits included "universal" this and that, which is marketecture to give the impression that said products play well with a wide variety of platforms, vaguely reminiscent of "DB2 Universal Database", which is actually multiple products, "innovation", which implies use of truly new ideas or methods, and "patient-centered" or "patient control", which imply that focus is on patients, even though the practical issue of competition just discussed seems to be the reality.

Somewhat disappointing was that the number of research exhibits in the information technology space mentioning open source software was less than in recent years, although when it was mentioned it seemed to have been presented better (for example, see the SIVIC graphic below), and it was a surprise to see that the number of research exhibits focusing on mobile platforms had decreased from last year. In fact, based on my experience at RSNA 2011, the amount of floor space dedicated to both commercial and research exhibits seemed to have decreased since last year.