Object storage keeps data as discrete objects retrieved by name through an API — the model behind cloud stores like Amazon S3 — rather than as files in folders or blocks on a disk. Each object carries the data plus metadata and a unique key. For surveillance it is the natural home for the cold and archive tiers and for cloud-recorded video, because it scales almost without limit, is durable, and is priced cheaply per terabyte for storage at rest.
Its strengths are scale, durability, and metadata. Object storage can hold petabytes without the capacity ceilings of a single array, replicates data for high durability, and lets rich metadata travel with each clip, which suits long-term archives and large multi-site cloud systems. Lifecycle policies move objects to colder, cheaper classes by age and expire them automatically, which maps neatly onto retention policy.
The pitfalls are latency, retrieval cost, and fit. Object storage is reached over a network with higher latency than local disk and is not designed for the constant low-latency block writes of live recording, so continuous video usually lands on local block storage first and is tiered into object storage as it ages — never stream live recording straight into deep-archive object classes. Colder classes also charge retrieval fees and impose minimum-storage durations, so getting footage back for an investigation has a cost and a delay to budget for. Use object storage for archive and cloud retention with lifecycle automation, and keep live recording on storage built for sustained write.
