ExaByte File Archive
Services section
More than just Object Storage
-
Contextual metadata
Every data object carries a multitude of metadata layers, that carry rich context—not just labels. Default layers cover identity, access; integrity and clear content descriptors; access rights and integrity. User defined layers offer maximum flexibility through custom APIs. The result: An Onion of metadata layers that powers discovery and analytics.
-
Bit rot resistant
Our system is "bit rot"1 resistant. Each object includes per-chunk checksums in the default metadata layer. Background data crawlers2 and real time access endpoints compare these; if a mismatch appears, cross site replicas repair the affected chunk3 before delivery or re-storage. This protects data against silent bit rot over its lifetime. Client applications can also store object checksums to validate integrity after retrieval.
-
Dynamic ExaFile storage
Object storage scales from a handful of megabytes to multi-exabytes; once stored, objects are immutable (but removable) with practically unlimited versions. New versions write only modified or new chunks (deltas), and chunks can include headroom to absorb future edits without rewriting the whole object. Data crawlers track new or changed objects, run checks or analysis, post notifications or summaries without reprocessing the entire version and avoid repeated scans over potentially terabytes of data.
-
Disaster recovery
Our object store is disaster resilient. After hardware is reinstalled or replaced, specialised recovery crawlers repair or duplicate lost data and rebuild indices and databases as needed. Even after major events, the system recovers all data from undamaged disks and restores service to the fullest possible state. During recovery, not all data may be continuously available, but priority requests are handled first.
Custom Hardware
To provide storage in remote locations with poor internet, we designed a quarter-standard transport container that can hold up to 1.3 exabytes of data. It's built from modular pods, each with one Mac mini and sixteen 20 TB disks; one pod has already been tested as a proof of concept. The container. includes its own power supply and climate control, and can be enhanced with a UPS and solar panels for greater autonomy.
The design prioritises energy efficiency. Using Apple's CoreStorage framework, we can power disks on or off with fine granularity, reducing consumption when full throughput isn't required.
When the container reaches its maximum capacity, it can be swapped out easily and transported to a central data centre.
Get in touchProblems we solve
We provide flexible, virtually unlimited data storage for finance and government, research and education, and for individuals and small- and medium-size businesses for general data-storage needs.
-
Finance, Banking & Government
- Safe, long-lasting, unchangeable data storage for records and contracts
- Analysis of large numbers of documents, ingest, classify, search and extract
- Contract management with version links, reminders, and audit trails
-
Research, Science & Education
- Safe storage for medical records and consent files
- Storage for huge scientific datasets, high-energy physics, astronomy/space, remote-sensing, ecological data
- Safe and uneditable laboratory notes for reproducible results
- Ability for scientists to develop custom data crawlers to analyse enormous amounts of data and notify of interesting or new findings
-
Small & Medium Businesses
- General company cloud storage
- Backup of company data
- Safe, encrypted storage of sensitive HR, finance, and other data
- Access for well defined user groups with read/write/ modify rights at the individual or group level
-
Individual users
- Storage for large amounts of media files (photos, movies, audio)
- Safe and encrypted storage for important personal files (passports, medical records, legal documents), with the ability to share them with relatives, guardians, or lawyers, without anyone else gaining access
- Storage for backup applications