Block Storage
HDD, SSD, and NVMe-backed volumes that attach to Cloud Compute or Bare Metal, resize on demand, and deliver their full performance profile with no IOPS surcharges.
Storage that scales with you
Grow a volume from gigabytes to terabytes without downtime or re-provisioning.
Every volume is replicated locally within its metro, so a single drive failure does not result in data loss.
Each tier delivers its full performance profile as standard. There are no provisioned-IOPS or throughput add-ons to layer on top.
Mount volumes to Cloud Compute or Bare Metal instances in the same metro, in any of our locations.
Provisioned per volume, attached over the network
Every volume is a raw block device attached over the network to your Cloud Compute or Bare Metal instance. It is formatted and used exactly like a local disk, with the added benefit of being resizable and replicated.
Select HDD, SSD, or NVMe based on the performance your workload requires. Tiers can be mixed and matched across the same instance.
Volumes attach to any Cloud Compute or Bare Metal instance in the same metro and mount as a standard block device.
Expand capacity live, with no downtime and no need to re-provision or migrate data to a bigger volume.
Pick your performance tier
Three tiers, all provisioned per volume and all replicated within their metro. Mix and match them across the same instance as different parts of your workload require.
No IOPS surcharges: unlike AWS EBS, which layers separate charges for provisioned IOPS and throughput on top of the storage itself, a Parallax volume delivers its tier's full performance profile as standard. Tell us the capacity and tier you need and we will quote the volume, not a matrix of add-ons.
The right tier for every workload
SSD or NVMe volumes for MySQL, Postgres, and other transactional workloads that need consistent low-latency I/O.
Persistent volumes for virtual machines and containerized workloads that need to survive redeploys and reboots.
Bulk storage for large video, image, and asset libraries where capacity matters more than raw IOPS.
Long-lived log retention and archival data that is written once and read rarely.
A cost-effective landing zone for scheduled backup jobs and snapshot exports off primary volumes.
Latency-sensitive services such as queues, caches, and real-time analytics, where every millisecond of I/O matters.