Minimum Hardware Requirements

Minimum CPU, memory, storage, and GPU capacity needed to run the BullSequana AI platform.

This page defines the minimum hardware baseline for a full BullSequana AI platform deployment.

Treat this baseline as the practical floor for getting started with BullSequana AI. Scale beyond it as production usage grows.

Minimum Cluster Sizing

The complete platform is validated on the following cluster shape:

Node pool	Nodes (min–max)	vCPU per node	RAM per node	Pool total (minimum)
System	2	4	16 GiB	8 vCPU / 32 GiB
Workers	7–10	8	32 GiB	56 vCPU / 224 GiB

That gives a cluster minimum of:

64 vCPU / 256 GiB RAM at the minimum worker count (7 workers)
72 vCPU / 288 GiB RAM at the typical initial worker count (8 workers)

Only the aggregate capacity and the per-node shape matter. Nodes with more vCPU and RAM per node are equally valid as long as the cluster totals are met and no single node is smaller than the worker shape above.

How The Platform Uses This Capacity

At steady state the platform reserves about 36 cores and 86 GiB of RAM, and can consume up to roughly 95 cores and 169 GiB under load:

Platform layer	CPU (steady state)	CPU (under load)	RAM (steady state)	RAM (under load)
Cluster infrastructure	9 cores	24 cores	28 GiB	43 GiB
AI serving	17 cores	55 cores	42 GiB	111 GiB
Data and ML	11 cores	15 cores	16 GiB	16 GiB
Total	~36 cores	~95 cores	~86 GiB	~169 GiB

The cluster minimum above is intentionally larger than the steady-state figures. The headroom covers load peaks, background jobs, rolling upgrades, and continued operation when a node is unavailable.

Storage

The platform provisions roughly 1.2 TiB of persistent storage:

Consumer	Storage
Strimzi (Kafka)	300 GiB
Rook-Ceph cluster backing	260 GiB
MLflow	210 GiB
Grafana	100 GiB
Milvus	~97 GiB
Other components	~200 GiB

Plan at least 1.5–2 TiB of usable storage to leave operational headroom. If Ceph replication is configured above a replication factor of 1, multiply the raw disk requirement accordingly.

The cluster must expose valid RWO and RWX storage classes before deployment. See Prerequisites for the storage-class inputs.

GPU requirements

GPU capacity is workload-dependent. BullSequana AI supplies KServe vLLM runtime profiles for CPU and 1, 2, 4, or 8 NVIDIA GPUs. Select a profile only after checking the model's weight size, precision, context length, concurrency, and tensor-parallel requirements against the available hardware.

Speech and Docling run on CPU by default and can use optional GPU worker settings. A deployment therefore has no universal three-GPU minimum; the required count is the sum of the runtime profiles and accelerated services enabled for that site.

Workload	GPU behavior
KServe vLLM	Selected runtime requests 1, 2, 4, or 8 NVIDIA GPUs; CPU profiles request none
Speaches speech service	CPU by default; optional CUDA mode requests one GPU
Docling	CPU by default; each configured GPU worker requests one GPU

Keep enough CPU, RAM, local or persistent model cache, and scheduling headroom for every selected profile. A GPU with sufficient VRAM can still fail if its node lacks the host memory or ephemeral storage needed to start the model.

GPU node configuration

GPU scheduling is driven by node metadata, not by VM SKU. GPU nodes must provide:

node label workload=gpu
node taint nvidia.com/gpu=present
a compatible NVIDIA driver and container runtime
the nvidia runtime class and NVIDIA device plugin, supplied by the cluster owner or the optional bsqai helm-apply gpu-operator bootstrap

By default the platform uses preferred (soft) GPU affinity. Set common_enforce_affinity to true to make GPU placement a hard scheduling requirement.

Summary

Dimension	Minimum
CPU	64 vCPU across 9 nodes (72 vCPU recommended)
RAM	256 GiB (288 GiB recommended)
Storage	1.5–2 TiB usable, with `RWO` and `RWX` storage classes
GPUs	Site-specific; 0 for CPU-only, otherwise the sum of selected KServe and optional speech/Docling GPU workloads