In Lustre, a logical object volume serves as an umbrella for multiple Object Storage Clients (OSCs), each of which connects to an OST. Because of this structure, a single file may be split across several OSTs, in a technique called striping. Striping allows parts of files to be stored on different OSTs in a RAID-0 pattern. This is done to enable fast parallel I/O, just as in a RAID-0 disk array.

• The number of objects is called the stripe_count.
• Objects contain "chunks" of data that can be as large as stripe_size.

With striping, the performance of a Lustre file system potentially scales with the number of OSTs available. Furthermore, the capacity of a Lustre file system equals the sum of the capacities of the storage targets.

These qualities result in two benefits:

1. I/O rate to a file is the aggregate of I/O rates to objects.
2. Max file size is not limited by the size of a single target.

For TACC systems, the total amount of Lustre storage is measured in petabytes, and the nominal aggregate data rate exceeds 150 GB/s. But this speed is split among all users of the system. Metadata access can be a bottleneck, so the MDSs need to have especially good performance (e.g., solid state disks on some systems).