By Tim Kaschinske
@TimKaschinske
Part 3 of 3
This is part 3 of 3 in a blog series intended to provide best practice information on using file-level protection within a hospital environment. For full information and context you can read Part 1 and Part 2, posted on Feb 4 and Feb 11 respectively.
In this entry, I will cover best practices for applying common archive policies such as “protect and keep” and “protect and stub” to your PACS repository without disrupting workflow. I will also share how to use policies to force immediate archive and utilize free space monitoring all designed to enhance your protection levels.
Protect and Keep
PACS applications are typically designed to ensure that images be accessed in less than two seconds. To support this retrieve performance, PACS vendors typically recommend policies that keep the files on their primary storage. Protect (archive) and keep policies do this by writing the files to multiple storage locations while keeping the original files on the primary storage for the application. Since the files are kept on primary storage there is no effect on the application or its performance. When the files on primary storage are lost or corrupted, then they can be recovered quickly from the archive.
Protect and Stub
Protect and stub is an alternative policy that can be used when protecting files. However, care must be taken to ensure that this approach does not affect application performance. With this policy, files are protected to multiple storage locations, but the original file is replaced with a “stub” that takes up less storage space. If the stub is accessed by the PACS application, the original file is restored transparently from the archive without any special action being required by the application or its user. In this way, the PACS application continues as if the file was always on primary storage.
This policy can be effective for reducing the cost and storage space required for PACS images. Stubbing ensures only the most frequently used files are stored on primary storage, and removes other files to secondary, less expensive media. Stubbing can also reduce the amount of time it takes to back up primary storage. An appropriate approach to stubbing is necessary to ensure that the files can be quickly restored so as not to affect application performance. For example, the performance requirement of the application will require that files be written to onsite disk storage with this policy.
Immediate File Protection
Scans for file protection are often implemented as batch jobs that run periodically. The scanning process used to find the files that must be archived is run as a batch job, periodically and not immediately after new files are created. Because scans are periodic, there can be a gap between when a file is created and when it is protected into the archive.
To avoid a large gap between file creation and its protection, you can take the approach of implementing a policy which can detect when a file a new file is created – and then take immediate action to protect it. This approach requires the use of an agent that can monitor access to the file system and detect when a file is created or modified, and then immediately trigger the archive job. This approach can dramatically improve file protection and reduce the risk of image loss.
Free Space Monitoring
As discussed previously, a policy of protect and stub can help to optimize the amount of free storage that is available to an application. While it is best to keep the data locally for performance reasons, organizations often struggle to maintain enough disk space to hold all of their images in primary storage. Purchasing cycles do not always line up well with storage needs. Often, IT teams purchase more storage than may be required to ensure that there is enough capacity for the application. A policy of protect and stub can help in these situations.
A simple alteration in the stubbing policy, however, can help it to operate better to address this particular problem. As an alternative to running a scan job that protects and stubs all files based upon created date or modified date, consider a scenario where the system can monitor the available free space and take action when a threshold is reached. Since the system has already protected the files using a policy of protect and keep, it can also keep track of the file attributes so that it is aware of the oldest, largest, and least frequently accessed files. A policy can be implemented when a threshold is reached to replace those files with a stub and reclaim needed storage space. This way, files are only stubbed when it is necessary to maintain free storage space, and can be restored when more capacity is added at a later date.
This best practice not only protects the files used by the application, but also against the application being down due to insufficient storage space. Image availability is the result.
