7 Digital evidence storage has become one of the largest and least glamorous line items in a modern police budget. Body-worn cameras, in-car video, interview room recordings, phone extractions, drone footage, CCTV pulled from businesses and doorbell video submitted by the public all land in the same place: an evidence repository that has to keep every file intact, findable and defensible in court for years, sometimes decades. The challenge is not just volume. Evidence storage has to satisfy prosecutors, defense counsel, records officers, auditors and the CJIS security officer at the same time. A file that is lost, altered or impossible to retrieve on time can sink a case. This article walks through what agencies actually store, how the storage layer is usually built, and what to look for when the current system starts running out of room. What counts as digital evidence today A decade ago most agencies thought of digital evidence as a few DVDs in a property room. Today the mix is far broader: Body-worn camera (BWC) and in-car video, usually the single largest source by volume. Interview and interrogation room recordings, often required by state law for certain offenses. Mobile device extractions from forensic tools, which can run from a few gigabytes to hundreds of gigabytes per device. Third-party video: CCTV exports from businesses, residential doorbell clips, traffic and transit cameras. Drone and aerial imagery, including high-resolution stills and 4K video. Computer forensic images of seized laptops, servers and storage media. Audio: 911 calls, radio traffic and jail calls. Almost all of it is unstructured, large and write-once. Very little of it is ever edited, but all of it may need to be retrieved quickly, copied for discovery and proven unchanged. How much digital evidence storage an agency really needs Video dominates every capacity plan. Industry guidance commonly cites roughly 0.8 GB per hour for standard-definition body camera video, around 3 GB per hour at 720p and roughly 12 GB per hour at 1080p, with each officer generating on the order of 1 GB per shift on average once real recording habits are factored in. Those numbers vary widely by device, codec and policy, so treat them as starting points rather than fixed rules. Our companion piece on body-worn camera storage requirements works through the math in detail. The more important drivers are usually policy, not pixels: Retention rules decide how long footage lives. Non-evidentiary video may be kept for weeks or months, while evidence tied to serious cases can be held for years or indefinitely. Recording policy (always-on versus event-triggered, pre-event buffers, resolution settings) can swing volume by several times. Duplication for discovery, appeals and redaction creates additional copies that also need governance. Because retained evidence accumulates year over year, many agencies find that storage demand compounds rather than grows linearly. Planning for petabyte-scale capacity is no longer unusual for large city or state agencies. The general principles in storage capacity planning apply directly here. The storage architectures agencies use Most evidence platforms separate the application (the digital evidence management system, or DEMS) from the storage underneath it. That storage typically takes one of four shapes. Vendor-hosted cloud Many body camera vendors bundle storage with their evidence software as a per-officer subscription. This removes hardware from the agency’s plate and simplifies procurement. The trade-offs are long-term cost as retained data grows, dependence on one vendor’s platform, and the effort required to move years of evidence elsewhere if contracts change. On-premises NAS or SAN Traditional file servers are familiar and fast for small agencies, but they become difficult to scale past a few hundred terabytes, and forklift upgrades every few years mean migrating evidence that must remain provably intact. On-premises object storage Object storage stores each file as an object with metadata and a unique identifier, scales horizontally by adding nodes, and supports features such as versioning and write-once retention locks. Many DEMS and video platforms can write to an S3-compatible target. For agencies keeping large archives on their own premises or in a government data center, this has become a common foundation. Hybrid A growing number of agencies keep recent, frequently accessed evidence close to investigators and move older material to a lower-cost tier, or keep a second copy at another site. The decision between models is covered in cloud DEMS vs on-prem evidence storage. Requirements that make evidence storage different Integrity and chain of custody Courts care whether a file is exactly what it claims to be. Evidence systems typically compute a cryptographic hash when a file is ingested and record every access, copy and export in an audit log. The storage layer supports this by preventing silent modification, keeping its own access records and periodically verifying object integrity against stored checksums. Under Federal Rule of Evidence 902(14), effective December 1, 2017, data copied from an electronic source can be self-authenticated when a qualified person certifies a matching hash value, which is one reason hashing has become standard practice. Our guide to digital evidence chain of custody goes deeper. Immutability Write-once controls, such as object lock in compliance mode, prevent anyone (including administrators) from deleting or overwriting evidence before its retention period ends. That protects against both ransomware and insider mistakes. The mechanics are explained in S3 object lock: immutability and WORM and immutable storage. CJIS security controls Digital evidence frequently contains criminal justice information, which brings the FBI CJIS Security Policy into scope. As of 2026 the policy is in its 6.x series (version 6.0 was released in December 2024 and 6.1 in June 2026), and it requires encryption of CJI at rest and in transit outside physically secure locations, multi-factor authentication and audit logging. Version 6.1 raised minimum encryption strength for several controls from 128-bit to 256-bit. Practical storage implications are covered in our CJIS compliant storage checklist. Retention and defensible deletion Holding everything forever is expensive and can create its own legal exposure. Agencies need storage that can apply retention by category, place legal holds and delete on schedule with a record of what was removed (the practice known as data deletion verification). Common failure points Agencies that struggle with evidence storage tend to hit the same problems: Capacity cliffs when a fixed-size array fills and a new one has to be bought and migrated mid-budget cycle. Retrieval delays when older evidence sits on slow or offline media and discovery deadlines loom. Ransomware exposure when evidence shares credentials and networks with general IT systems. Single-site risk when every copy of the evidence lives in one building. Lock-in when exporting years of evidence from a hosted platform turns out to be slow or costly. Cloud exit planning is worth doing before signing, not after. Checklist: evaluating digital evidence storage Can capacity grow in small increments without a forklift migration? Does the storage support write-once retention that administrators cannot override? Are hashes, access logs and deletion records kept and exportable for court and audit? Is data encrypted at rest and in transit with key management that meets current CJIS guidance? Can evidence be replicated to a second site or kept in an air-gapped copy? Does the DEMS or video platform support the storage target natively (for example over the S3 API)? How quickly can a large case file be retrieved and copied for discovery? What does it cost to exit, including bulk export of retained evidence? Who controls the encryption keys and where does the data physically reside? Putting it together Digital evidence storage is a long-term commitment. The footage recorded this year may still need to be produced in court a decade from now, intact and with a clean chain of custody. That favors storage that scales without disruptive migrations, enforces immutability, records every access and keeps the agency in control of where its evidence lives. Whether the answer is a vendor cloud, an on-premises platform or a hybrid of both, the decision should be driven by retention policy, legal defensibility and total cost over the full life of the evidence, not just the first contract term. Frequently asked questions Is cloud storage allowed for digital evidence? Yes, as long as the provider and configuration meet the CJIS Security Policy and state requirements. Agencies still need to confirm encryption, personnel screening, audit access and data location in the contract. How long should digital evidence be kept? It depends on state law, offense type and case status. Non-evidentiary video may be kept for 60 days or so in some states, while evidence in serious felonies can be held for many years or indefinitely. What storage features matter most for court? Integrity (hashing and immutability), complete audit logs, and the ability to produce exact copies quickly. Without those, defense counsel can challenge authenticity. Why do agencies move evidence to object storage? Object storage scales to very large capacities, supports write-once retention and metadata, and works with many evidence platforms over the S3 API, which makes it a practical foundation for long-term archives. Does digital evidence need to be encrypted? When it contains criminal justice information and leaves a physically secure location, the CJIS Security Policy requires encryption. Most agencies encrypt evidence at rest everywhere as a matter of course. Further reading For more depth, see body-worn camera storage requirements, digital evidence retention, chain of custody for digital evidence, cloud DEMS vs on-prem evidence storage and our CJIS compliant storage checklist.
