Secure payments
Standard refund policy
Trade Assurance
| Brand | Dell |
|---|---|
| Model | PowerEdge R960 |
| Form Factor | 4U Rack |
| CPU Model | Intel Xeon Gold 6416H (18C/36T) |
| Configuration | Dual Processors (2S) |
| Memory Config | 1TB DDR5 4800MHz (32x32GB) |
| Memory Bandwidth | 16-Channel Architecture |
| Storage Config | 4x 3.2TB NVMe PCIe Gen4 SSD |
| Storage Tier | High-Speed Cache Tier |
| Controller | PERC 12 (PMX5W4) Front Controller |
| NIC | Dual 100GbE QSFP56 OCP 3.0 Adapter |
| FPGA Support | FPGA Accelerator Support |
| Management | iDRAC9 Enterprise with Redfish API |
| Power Supply | Dual 2400W Platinum AC Redundant PSUs (1+1) |
| Workload Optimized | SAP HANA, Redis, MemSQL Ready |
| Latency Optimization | Low-Latency Memory Access |
| Condition | Brand New, Original |
Engineer's Note: The In-Memory Computing is a 4U Rack rack platform built around Intel Xeon Gold 6416H (18C/36T), with a scalable memory subsystem. It is sized for software-defined storage and backup targets where memory capacity is as decisive as raw compute.
1. Power resilience: Dual 2400W Platinum AC Redundant PSUs (1+1). Redundant, hot-swap power keeps a single PSU failure from becoming an outage — the cheapest insurance in the rack.
2. 4U Rack form factor: the chassis height sets the expansion budget — full-height PCIe slots for GPUs and HCAs, and drive bays for local boot or cache tiers. It determines how the platform evolves over its 5–7 year service life.
3. Compute density: built on Intel Xeon Gold 6416H (18C/36T). In a consolidated environment, per-socket core count determines how many VMs or containers each node can absorb before scheduling contention appears. Sizing the compute tier correctly the first time avoids the classic trap of overbuying "just in case" and paying for idle silicon.
4. Management: iDRAC9 Enterprise with Redfish API. Out-of-band control (health telemetry, remote console, firmware updates) means hardware issues are diagnosed without a datacenter visit, which directly improves fleet uptime.
5. Deployment fit: the In-Memory Computing is typically deployed as software-defined storage and backup targets. Its subsystem balance — compute, memory, storage, and networking in proportion — is what separates a purpose-built node from a general-purpose box that excels at nothing.
When the requirement is software-defined storage and backup targets, the In-Memory Computing is engineered so no single subsystem bottlenecks the rest. and the management plane keeps operational overhead flat as you scale. Each unit is fully tested before dispatch, and our team supports bulk deployments with configuration guidance and consolidated pricing.
Q: Can the In-Memory Computing be customized to my requirements?
A: Yes. Processor count, memory population, storage layout, and networking can be configured to order. Tell us your workload in the quote form and our engineers will propose a balanced configuration with room to grow.
Q: What is the power draw I should plan for?
A: Plan per the configured PSU rating plus headroom: a fully populated system can draw near the PSU maximum under load. For datacenter provisioning, size PDUs on max configured wattage, not idle draw.
Q: Can this server be shipped internationally?
A: Yes, we ship worldwide with export-compliant packaging. Transit insurance and appropriate power cables for your region are included on request.
Q: Which rails and rack accessories do I need?
A: The server requires model-matched rail kits for your rack type (round-hole, square-hole, or threaded). Confirm your rack specification when ordering — we can include the correct rail kit and cable management arm.
Q: Do you offer bulk pricing for multi-unit orders?
A: Yes — cluster deployments and multi-unit orders receive consolidated pricing and coordinated logistics. Request a quote with your target quantity for a same-week response.
