Genome Tape vs Cloud Cold Storage

A simplified model for ~200 whole genomes per year: cold archive on LTO tape, with local HDD/SSD working storage, compared to cloud cold/archive storage with and without re-read/reanalysis costs.

Workload

Genome size, GB Starting genomes Genomes added/year Total years

Cloud read/reanalysis frequency per year Auto-zoom graph to show break-even

Cloud cold storage rates

Cold storage $/TB/month Read/restore $/TB read

Separate cloud rows show: pure offsite cold backup with zero reads, and cloud cold storage where the archive is read back for reanalysis.

Tape cold storage: base package

Tape robot/library cost Slots in robot TB per LTO tape Cost per tape

Local working storage

HDD drive size, TB HDD cost $/TB HDD drives in RAID5 SSD scratch TB SSD cost $/TB Amortize hardware over years

Defaults: Quantum SuperLoader 3 style package: 8 slots, LTO-8, 96 TB native full, quoted at $10k for the unit. LTO-8 tapes default to $100 each and 12 TB native.

Final archive

Tapes needed

Robot full capacity

Break-even size

Tape vs cloud cold storage

Cloud backup, no readsCloud with reads/reanalysisTape + local storage upfrontBreak-even

Selected local storage capacities

Storage area	Raw / purchased	Usable	Cost	Notes

Cost summary

Option	Upfront	Recurring/read cost	Total over years	Effective $/TB-year	Effective $/genome-year

Tape and local hardware details

Item	Quantity	Capacity	Unit cost	Total	Notes

Year-by-year cloud vs tape

Year	Genomes	Archive size	Cloud no-read cumulative	Cloud with reads cumulative	Tape/local cumulative

Reference links

Cloud pricing: AWS S3, Azure Blob, Google Cloud Storage. Tape references: LTO Program, Quantum SuperLoader 3.