Data is spread across data centres and clouds and cannot be fed to GPUs fast enough; every new compute node means another data move and another storage build.
JDC (Qingxuan) is the storage and service foundation of the Jumborca Storage Power Center; it can also be read as an AI storage base. It holds PB-scale data, serves it over file, object and block protocols, places it in hot, warm or cold tiers, and moves the data that training and inference need to sit next to the compute. JDC is not sold as equipment: it is delivered as a service inside the Storage Power Center. JDC-365, our Microsoft 365 backup service, belongs to the same family.
Capabilities
- PB-scale capacity: one namespace for very large datasets, expanded as needed
- Multi-protocol access: file, object and block protocols on the same data, so training, inference and business systems each get what they need
- Hot-warm-cold tiering: data placed by access frequency, hot data close to compute, cold data kept long-term at low cost
- Data scheduling: works with the Storage Power Center scheduler to stage data beside a compute node or send the job to where the data is
- Delivered as a service: no equipment to buy; used under a Storage Power Center contract and operated by Jumborca