Delivery modes and operating boundary
Sakura Internet’s Koukaryoku GPU-cloud family spans PHY bare-metal servers for dedicated high-performance clusters, VRT virtual machines, and DOK container-based GPU task execution; its official materials describe all as separate service options.
PHY is whole physical-server capacity, VRT provides virtualized GPU servers, and DOK runs container tasks without the same middleware/scaling operations burden. They have different tenancy, provisioning and workflow constraints; verify current availability and terms in Japan.
Computraq is prelaunch. Naming this provider here does not mean that Computraq hosts workloads on it, has tested its integration, or has a commercial relationship with it.
Buyer and procurement preparation
Use these questions to prepare a comparable, workload-specific evaluation. Product names describe the provider’s published offering; they do not establish availability for a particular account.
- Choose among PHY, VRT and DOK based on full-node control, VM flexibility or task-oriented containers; identify required GPU count, network and storage.
- Confirm current service status, reservation process, Japan data-location terms, capacity period, support and billing for the selected Koukaryoku tier.
- Ask how data is persisted or deleted and which cluster, middleware and scaling tasks the service provider owns.
Developer and operator evidence
Before selecting capacity, create a small acceptance plan that records the model and revision, input shape, runtime, GPU configuration, concurrency, target latency or job completion, and recovery expectations. Provider documentation is evidence of product design, not proof that this workload has been tested.
- For PHY, plan cluster-level installation and job orchestration; for VRT, verify guest driver, OS and network; for DOK, package the complete container task.
- Test Japanese-language support and data-location requirements with the applicable contract, not brand-level assumptions.
- Capture model/image versions, input/output paths, runtime limits and restart behavior in the workload runbook.