This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
Video is everywhere, and growing exponentially. According to a recent report, 35 billion video ads were viewed in December, representing year-over-year growth of more than 100%. And every industry is seeing video growth, which creates a problem for data managers because video challenges storage management in four ways:
If, however, you move video to specialty storage, you'll achieve five amazing benefits:
* Higher Performance. Video has different characteristics and performance requirements from traditional corporate data. A single video file is often up to a TB or more. Collaborative video processing applications (e.g., post production) may or may not be IOPS- or bandwidth-centric, but they often require almost zero latency.
This combination of large file and latency sensitivity is a challenge for traditional storage architectures. Storage arrays typically utilize intelligent memory caches to deliver performance; having written data to the cache, the application is freed to perform its next operation. Read transactions from memory buffers can occur at much higher speed than a read from disk. For traditional data, this model works well. But these software-managed caches are not large or flexible enough to manage unpredictable video data streams.
Video files overflow storage caches, causing the array to pause while it pages data to and from the disk. Meanwhile, the latency-sensitive video application continues to send data. This results in "frame drop"; writes which are not serviced rapidly enough are "dropped" by the storage system. This results in a performance problem and may, in some cases, cause data to be lost.
When video causes this data traffic jam, other transactional or productivity data which is better suited for this architecture isn't getting serviced either. Resources are being used but nobody's getting good performance.
Field engineers from the best array vendors will anticipate these problems and deal with them by turning the caches off and overprovisioning disk heads in the storage system to address the needs of the video/image data. But, besides adding expense, this leaves traditional data without the tools needed to address performance.
Segregating out large, latency-sensitive video data — and applying specialty storage which is engineered to match its needs — gives both sets of applications better data access and performance.
* Faster, smoother, cheaper backups. In addition to being super-sized, video data is unique in a second way: these files are static. While users may clip pieces of these files to make new files, the original files aren't modified. For this reason, they can create an unnecessary burden in your backup process. Smart incremental backup technologies will avoid backing up these files, but when you need to do a full backup, these large static files will get copied — over and over again, using lots of bandwidth and storage.
Sign up for Computerworld eNewsletters.