
Kit Wai Chan - Fotolia
How can flash storage aid medical imaging archive needs?
Expert Brien Posey breaks down why flash storage may not be the best choice for medical imaging archiving in terms of speed and cost, and how it may be useful for other purposes.
On the surface, it would initially seem that flash storage is not the best choice for a medical imaging archive. After all, the most appealing thing about flash storage is its speed, and medical record archives are by definition static. Most of the data residing within the archives is likely to remain there, unused, until a healthcare organization's data lifecycle management policy, state law or HIPAA eventually allows that data to be purged.
There is also the issue of storage capacity. Medical imaging data tends to consume a significant amount of space because it includes high-resolution imagery that is necessary to help radiologists and physicians diagnose illnesses. Typical hospitals may generate 10-20 TB of image data each year, all of which must be stored digitally somewhere.
Although manufacturers have introduced solid-state drives (SSDs) in recent years with capacities that rival those of spinning media, the cost per gigabyte tends to be much higher for SSD. A medical imaging archive is probably not the best use for such a high-priced storage tier.
However, even though an all-flash array is probably a poor choice for use as medical records archive storage, it is possible that flash storage could be employed in another, more practical way.
A medical imaging archive could conceivably benefit from using flash storage as a read/write cache. Using flash-based caching provides faster writes and reads of data. Also, when clinical use is heavy, medical imaging archive systems can produce more storage-related IOPS. IOPS is a common performance measurement for storage devices. Flash-based cache approaches can flatten I/O levels and boost the system's responsiveness.
So how does flash cache work in for imaging? If a clinician needed to read data from medical image archives, for example, that data could be automatically copied to an SSD-based read cache. This would allow subsequent reads of the data to occur much more quickly than would be the case if the image were being pulled from archive storage. When the data is eventually deemed by the system to be "cold" -- in other words, the data is used less frequently and thus does not need high-performing storage -- it is moved out of the cache to make room for other data.
Similarly, flash storage can also act as a write tier. This setup would allow data to be written to the medical imaging archive at SSD-like speeds. The cached data would then be gradually moved by the system off of the write cache and onto the high-capacity tier.
Given how greatly the world of medical imaging -- particularly its digital aspects -- has changed over the last two decades, healthcare organizations will need to periodically re-evaluate the storage of such information. For example, it remains to be seen how AI might affect any of the flash-based options above, as AI brings up the possibility of real-time analytics of medical imaging data, which may challenge the notion of an image remaining static. Also, greater interoperability among organizations may see image data being shared more frequently between different storage systems.
SearchHealthIT editorial director Scott Wallask contributed to this article.