I sat in on a webcast Wednesday, July 20, 2011, and noticed how the concepts of Disaster Recovery and Business Continuity were frequently paired…in the bullets and in the presentation…with the inference that if you had a DR solution, you had a Business Continuity solution. That’s not the case. A Disaster Recovery solution is basically a second copy of the data, usually stored remotely, so it can be accessed whenever it becomes necessary to replace the first copy of the data that might be lost or temporarily unavailable. A DR solution could be as simple as an off-line digital tape cartridge or a DVD disk stored in a safe room. It could be a near-line solution such as a tape library system or an on-line solution like a separate spinning disk system, either of which is directly attached to the PACS. A second SAN or NAS storage solution paired with its front end gateway server could be located in a remote data center in order to increase its chances of avoiding whatever disaster might take down the PACS and/or its primary copy of the data. The DR solution could be as elaborate as a Cloud-based storage solution, which often feature multiple data centers located in distant states.
In all of the above examples however, the original PACS application or a standalone display application like a web server is required to access and display that back-up image data. If the PACS application or any of the display applications are down, or in any way unavailable, the second copy of the data cannot be accesses and it cannot be displayed. In this case, there is no Business Continuity.
A significant number of installed Radiology PACS are based on a single instance of the data management and display application, but they are configured with both a primary and a secondary storage solution, each located in separate data centers. Depending on how geographically separate those data centers are, the secondary storage solution represents a solid Disaster Recovery solution.
On the other hand, very few Radiology PACS are configured with two instances of the data management application or the display applications. And because there is only one instance of these applications, such a configuration does not have a Business Continuity solution. With few exceptions, if the PACS application is unavailable, neither the primary nor secondary copy of the data is accessible. If the display applications are unavailable, neither the primary nor the secondary copy of the data can be displayed.
A true Business Continuity solution requires two instances of the data management application, which can access either the primary or secondary storage solutions, and two instances of whatever display application the user prefers for accessing and displaying the data. These two paired applications…data management and data display…should be geographically separated, so either of them can survive the disaster that might befall the other.
Since few PACS can be configured with multiple instances of its data management and display application, the current strategy for building a Business Continuity solution has shifted to deploying a dual-sited Vendor Neutral Archive and a dual-sited UniViewer. The two separate instances of the VNA and the two separate instances of the UniViewer back each other up in the event of a disaster. In the event that the only instance of the PACS and/or its only instance of the display application becomes unavailable, the new studies are probably interpreted at the modalities, and the priors are retrieved from whichever VNA is available and displayed using whichever UniViewer is available. That is a true Business Continuity solution. Take a look at my previous post regarding Failover Strategies, if you want to get a better idea of how primary and secondary subsystems back each other up.