Your documents live on independent cloud storage providers, not on our servers. Every document is currently replicated to two providers, and reads are served from whichever one is available, so a provider outage never takes your documents offline.
Why two providers
Single-provider storage is a single point of failure. Outages happen. Regions go offline. With a single provider, their downtime is your downtime. Replicating across unrelated providers means the failure modes do not overlap.
What you get
- No data loss from a provider failure. A copy on a separate provider always exists.
- No downtime from a provider outage. If one provider is unreachable, reads are served from the other. Document URLs keep working.
- Nothing to configure or run. Replication happens on every upload, automatically, with no queues or jobs for you to monitor or schedule.
Where your documents live
Today pdrive replicates every document to Akamai Cloud and Cloudflare R2. Both are major S3-compatible providers on different networks, and we configure each one to store documents in a different region. A problem at one does not propagate to the other. Documents are encrypted in transit and at rest at both providers.
Neither provider is AWS. When a major AWS outage cascades across the internet, pdrive stays reachable, and so do your documents.
We may add or change providers over time. When that happens, nothing changes for you. Document URLs stay the same, your documents keep loading, and you do not need to migrate or reconfigure anything.
Caching
All document reads are cached on our servers, so repeat requests stay fast. The first read of a document pulls it from the storage provider. Subsequent reads come straight from cache while the document is being accessed.
Caching is about speed, not durability. The offsite providers remain the source of truth. If the cache is lost, reads fall back to the providers on the next request.