Enterprises are going visual, a trend that's beginning to pose a serious storage challenge to organizations in a number of fields.
Here are just a few examples of image data storage needs across various industries:
- Semiconductor manufacturers save images of wafers and other components for manufacturing purposes, as well as to meet compliance and liability obligations.
- Insurance firms store vehicle and structural images to verify damage claims.
- Businesses and law firms store images of contracts and various other types of documents.
- Companies with recognized brands store significant amounts of creative content for advertising and other marketing purposes.
- Online shopping services need to store massive amounts of product photos.
- Shipping services frequently store images to confirm deliveries.
- Healthcare researchers and service providers need to save an array of medical images.
- Commercial photographers routinely save large RAW image files.
- Emerging technologies, such as autonomous vehicles and genomics, use large numbers of small image files.
- Virtually every enterprise stores PDFs, PowerPoint images, employee photo IDs and a range of other visual documents.
Best image data storage practices
A classification strategy should be at the top of the image data storage best practices list, said Jim Kemp, director of operations at document management technology provider ColumbiaSoft, based in Portland, Ore.
"This is especially true if images are used in business processes and related to other documents, such as reports, emails or business records," Kemp said.
Image classification determines what types of values are associated in terms of metadata -- the facts that describe attributes of each image.
"Depending on the needs of the business process, these facts may include everything from a date to [geographic information system] coordinates to a description of the image itself," Kemp said.
Proper classification is not only valuable when searching images, it's essential for cross-referencing files with related content. It can even drive automated image naming, filing location, security and records policies services, Kemp noted.
Storage approaches and tools
High-resolution images are typically huge files that soak up large amounts of storage. Image data also needs to be readily available and well preserved. The best way to meet these image data storage requirements is with a network system such as a SAN.
Hector RuizPresident, Corporate Shields
"SANs are enterprise-ready storage solutions capable of holding a large amount of data while offering reliability and protection," said Hector Ruiz, president of Corporate Shields, an Orlando-based IT consulting firm.
A scale-out file server is frequently the choice for saving image data as primary storage. Meanwhile, object storage, the public cloud and tape drives are all appropriate targets for data protection and archiving.
The public cloud can potentially supply high performance with reasonable latency and provide the durability of object storage without the upfront capital costs. However, it's potentially more expensive over time when users account for storage and egress costs.
Minimize image data storage volume
Organizations that stockpile many images can free a limited amount of storage space with software.
"Some enterprise storage solutions offer image compression technology and file deduplication, cutting down the storage amount needed to store the files," Ruiz said.
Collaboration between IT and business units is the best way to minimize the amount of outdated data that wastes storage space.
"It's not IT's responsibility to determine the business value of storing image data," said Jordan Winkelman, field solutions CTO for the Americas at storage technology provider Quantum, based in San Jose. "Sometimes, data is just maintained forever because business units are unsure of what can be deleted and, therefore, just store and back up everything."
Fortunately, numerous media and digital asset management tools enable rapid classification of data sets. The tools can tag metadata that IT needs to enter manually. They can also harvest metadata from binaries or even use AI to index images.