Recovery Protocol

How Many Backup Fragments Should You Keep? The Mathematics of Disaster Quotas

How to calculate the optimal number of threshold fragments for personal and enterprise backups, balancing loss tolerance, storage complexity, and operational friction.

YourKeep Team3 min read
#Fragment Quotas#Threshold Strategy#Redundancy Planning#Erasure Coding

How Many Backup Fragments Should You Keep? The Mathematics of Disaster Quotas

When configuring a threshold recovery backup system, one of the most critical engineering decisions is determining the total number of fragments ($N$) and the minimum reconstruction threshold ($K$).

Generating too few fragments (such as 2-of-2) leaves you with zero redundancy—a single failed hard drive or locked cloud account causes catastrophic total data loss. Conversely, generating too many fragments (such as 20-of-30) introduces massive cognitive overhead, increases the difficulty of endpoint management, and inflates bandwidth costs during initial uploads.

Finding the optimal quota requires balancing mathematical failure probabilities against real-world human operational discipline.


The Core Trade-Off Triangle in Threshold Systems

Every threshold scheme balances three competing constraints:

                  [ Loss Tolerance (N - K) ]

                            / \
                           /   \
                          /     \
  [ Storage Efficiency (K/N) ] ─── [ Operational Simplicity ]
  1. Loss Tolerance ($N - K$): How many simultaneous endpoint failures you can survive without losing data.
  2. Storage Efficiency ($K / N$): The percentage of distributed storage dedicated to actual data versus mathematical parity.
  3. Operational Simplicity: The number of distinct physical devices, cloud accounts, and custodians you must track over time.

Evaluating Standard Threshold Configurations

Configuration ($K$-of-$N$) Storage Overhead Node Loss Tolerance Recommended Use Case
2-of-3 1.50x (150%) 1 node (33%) Ephemeral projects, single-user low overhead
3-of-5 1.67x (167%) 2 nodes (40%) Personal documents, small family estates
4-of-7 1.75x (175%) 3 nodes (43%) Medium-tier legal archives, developer teams
6-of-10 (YourKeep Default) 1.67x (167%) 4 nodes (40%) Master lifetime archives, multi-cloud redundancy
8-of-12 1.50x (150%) 4 nodes (33%) Enterprise bulk cold storage

Why 6-of-10 Is the Optimal Sweet Spot for Personal Archives

YourKeep selects 6-of-10 as the standard default for master archives because it maximizes survivability without overwhelming the user:

1. High Multi-Category Failure Tolerance

In a 6-of-10 setup, you can lose four independent storage nodes simultaneously. This covers realistic multi-vector disaster scenarios:

  • Provider A permanently suspends your cloud account (1 node lost).
  • An old external hard drive develops mechanical failure (1 node lost).
  • A flash drive is lost during relocation (1 node lost).
  • An S3 bucket subscription payment fails due to an expired credit card (1 node lost).

Even after these four concurrent failures, you still possess 6 healthy fragments across other endpoints, allowing you to execute a 100% loss-free restore in seconds.

2. Manageable Node Topology

Distributing 10 fragments across a balanced matrix is intuitive:

  • 2 Local hardware nodes (Workstation SSD + Local NAS)
  • 4 Cloud storage nodes (Google Drive, OneDrive, Dropbox, Proton)
  • 2 Object storage nodes (Backblaze B2, Cloudflare R2)
  • 2 Physical cold media nodes (Home safe USB + Family member safe USB)

Summary Decision Rule

  • For lightweight, temporary backups with low maintenance: use 2-of-3.
  • For uncompromised lifetime preservation of irreplaceable family and business records: use 6-of-10.