… better than LTO or a File System that is. And, these are my top five reasons why:
Scaling … that’s future proof
You might be able to add a tape to an LTO set-up or a new JBOD to a SAN, but can you do that in five-year’s time or will that require a tech swap out? With good object storage you can scale old with new. It’s a far cry from SAN’s that only scale with like-for-like hardware or LTO5…6…7…. Is it even readable anymore?
Transportability you never knew you needed
An object transfers with its metadata from location to location. Simple. Whereas a file based solution may well have to have its metadata in a database or in a format that is not very user friendly and possibly a little bit “niche”, difficult to keep in sync as data moves from location to location. Data migration is a fact of life and will always be a pain point. Lose a DB and that can pretty much make the files unfindable.
But nowadays, object storage is more than just an enabler for reliable “replication”. Modern hybrid workflows in an API connected world can make use of the metadata to enable workflows that go from on-prem to cloud to another on-prem and back again. Object storage, with its metadata and APIs enables it all.
Fast and distributed asset access: Can I hear “AI”?
Re-using assets should be the main reason of keeping them in the first place, yet it doesn’t always feel that way. At least historically in media libraries, that “re-use” was often more of a “WORN” experience… write-once, read… never! However, now, sports teams are reselling libraries of obscure matches, news-rooms are reselling news clips and productions are pulling shots from their libraries instead of needing to re-shoot. And AI comes into this, because now we need quick and distributed ways to re-read and re-analyse media so that it can be found and re-used. Distributed read access is very important because libraries can be extremely large and paralysed access from multiple clients allows analysis to become practical. And of course, object storage provides parallel access in spades.
Metadata that’s searchable … and extensible
Good object storage lets you search the metadata that is kept with the objects – this is hugely advantageous over keeping the metadata in a database. Databases get out-of-date, out-of-sync, require management and aren’t highly transportable at a file-by-file level. Keeping metadata in a file is ok.. but if you want to change that metadata then you need to rewrite the file. This is doable, but not really very practical (e.g., imagine keeping a read counter or an audit in the file). Much better to have a system that makes all of this easy, because to quote an old adage: if you can’t find it, you don’t have it.
Digital Content Governance: that’s lifecycle management to some
Digital Content Governance, or DCG, means keeping content with the right level of redundancy, with audits, and security nailed. DCG is concerned with fine grained user access rights, the immutability of the content (e.g., for a period of time) and ensuring the content stays bitwise authentic. All of these services can be built into good object storage solutions but are unlikely to exist in LTO or SAN/NAS solutions.