How to Choose HPE ProLiant Servers & Storage

Written and maintained by Haink's enterprise infrastructure team · Updated July 2026 · authorized-channel, serial-verified

Haink supplies HPE (Hewlett Packard Enterprise) servers, storage systems, and composable infrastructure to enterprises, data centers, and cloud providers in Hong Kong, Dubai, and Mainland China. Available product lines include HPE ProLiant DL and ML rack and tower servers, HPE ProLiant BL blade servers, HPE Synergy composable infrastructure, HPE Alletra storage, HPE Primera, HPE Nimble Storage, and HPE SimpliVity hyperconverged systems.

HPE is one of the world's largest enterprise hardware manufacturers, delivering compute, storage, and networking infrastructure used by organizations across financial services, telecommunications, cloud computing, healthcare, and government. Haink sources HPE equipment from authorized distributors and delivers to enterprise facilities and data centers across the Asia-Pacific and Middle East regions.

HPE Buyer's Guide — Choosing the Right Platform

Use this section to choose an HPE platform by workload, navigate the Gen11→Gen12 transition, see what is current versus end-of-life, decode HPE part numbers, and verify serial numbers before you pay.

Which HPE platform for your workload

Gen11 → Gen12: what to buy now

For new deployments that need the latest core counts, memory bandwidth and efficiency, choose ProLiant Gen12 (Intel Xeon 6). Gen11 (Intel Sapphire/Emerald Rapids and AMD EPYC 9004) remains widely available from stock and is frequently the better-priced choice for standardized virtualization and database builds where the newest Xeon 6 performance is not required. Gen12 is currently Intel-only — AMD EPYC ProLiant remains on Gen11. Where stock economics matter, Haink can quote Gen11-from-stock against Gen12-to-order side by side.

HPE lifecycle: current vs legacy / EOL

How to read an HPE part number

Fuller breakdown — the factory-only options that cannot be added later, and why the same DIMM carries both a -B21 and a -F21 number: HPE part numbers decoded.

HPE SKUs follow a consistent structure that often trips up buyers comparing quotes. An HPE part number is a base number plus a region/channel suffix, optionally followed by a localization option code:

• Base number — e.g. P28948 (newer parts start with “P”; older HP-era parts are all-numeric, e.g. 868703). This identifies the product or option kit.
• Suffix after the hyphen — the region/channel code. -B21 is the standard worldwide / most-regions code; -AA1 is China and -291 is Japan, and channel or direct variants such as -S21 or -425 also exist. The same server can appear under 3–7 different suffixes — the hardware is identical, only the sourcing region/channel differs.
• CTO vs BTO — a -B21 chassis marked “CTO” is a factory Configure-To-Order base you build up with option kits; a BTO SKU is a pre-defined, ready-built configuration.
• Trailing #code — an appended #ABA, #0D1 or #B19 specifies a localization / factory option (power cord, language, integration).

Practical takeaway: when you compare two HPE quotes, match the base number — a different suffix (e.g. -B21 vs -AA1) is the same box sourced through a different region, not a different product. Haink quotes the region-appropriate SKU and confirms it against HPE QuickSpecs before you order.

Serial-number verification before payment

The full buyer-side check — routes to market, the three tests that identify each, and the questions to put to any supplier: sourcing HPE.

Every HPE unit Haink supplies is serial-verifiable before payment. On request we provide the unit serial numbers so you can confirm — through the HPE Support Center warranty-entitlement and product-authentication check — that each server or array is genuine, factory-new (or explicitly identified as HPE Renew / refurbished), and carries valid HPE warranty for your region. This protects you from gray-market and counterfeit hardware, and it is a standard part of every HPE order we fulfil.

HPE ProLiant Servers

HPE ProLiant is the most widely deployed enterprise server platform in the world. ProLiant servers are available in rack, tower, and blade form factors and support workloads including virtualization, databases, enterprise applications, AI inference, and high-performance computing.

HPE ProLiant Gen12 (Intel Xeon 6) — Current Generation

HPE ProLiant Compute Gen12 is the current generation, built on Intel Xeon 6 (Granite Rapids) with DDR5, PCIe Gen5 and both air and direct-liquid cooling (DLC). Six models shipped from Q1 2025 — HPE ProLiant DL320, DL340, DL360, DL380, DL380a and ML350 Gen12 — with the four-socket DL580 Gen12 and HPE Synergy 480 Gen12 following in summer 2025. A dual-socket DL380 Gen12 scales to 288 cores and 8 TB of DDR5. Gen12 is currently Intel-only; the AMD EPYC ProLiant line remains on Gen11.

• HPE ProLiant DL360 Gen12 — 1U dual-socket Intel Xeon 6 rack server for high-density compute
• HPE ProLiant DL380 Gen12 — 2U dual-socket Intel Xeon 6 workhorse (up to 288 cores, 8 TB DDR5), air or DLC
• HPE ProLiant DL380a Gen12 — AI/GPU-optimized 2U variant (supports NVIDIA H200 NVL)
• HPE ProLiant DL320 / DL340 Gen12 — 1U / 2U single-socket for edge, storage and scale-out
• HPE ProLiant ML350 Gen12 — tower server for branch and edge
• HPE ProLiant DL580 Gen12 (4U, four-socket) and HPE Synergy 480 Gen12 composable module — available from summer 2025

The previous ProLiant Gen11 line (Intel Sapphire Rapids and AMD EPYC 9004) remains widely available from stock and is often the better-priced choice for standardized virtualization and database builds where the latest Xeon 6 core counts are not required.

HPE ProLiant DL Rack Servers

HPE ProLiant DL rack servers are designed for data center deployments requiring high density and scalability. Haink supplies the following HPE ProLiant DL server models:

• HPE ProLiant DL360 Gen11 — 1U dual-socket rack server for high-density compute environments
• HPE ProLiant DL380 Gen11 — 2U dual-socket rack server, the most widely deployed HPE platform for general-purpose workloads
• HPE ProLiant DL580 Gen11 — 4U four-socket rack server for memory-intensive workloads and large database environments
• HPE ProLiant DL20 Gen11 — 1U single-socket entry-level rack server for remote and branch deployments
• HPE ProLiant DL110 Gen11 — 1U single-socket storage-optimized platform
• HPE ProLiant DL560 Gen11 — 2U four-socket high-performance platform

HPE ProLiant Gen11 servers are powered by Intel Xeon Scalable processors (Sapphire Rapids) and AMD EPYC processors, with support for DDR5 memory, PCIe Gen5 expansion, and HPE iLO 6 integrated management.

AMD EPYC-based HPE ProLiant DL Servers

HPE also offers a full AMD EPYC-based ProLiant DL line, delivering high core density, large memory capacity, and competitive cost per core for cloud-native, virtualization, and scale-out workloads.

• HPE ProLiant DL325 Gen11 — 1U single-socket AMD EPYC 9004 platform for high-density storage and cloud-native workloads
• HPE ProLiant DL345 Gen11 — 2U single-socket AMD EPYC platform with high drive bay count for storage-intensive environments
• HPE ProLiant DL365 Gen11 — 1U dual-socket AMD EPYC platform for compute-dense virtualization requiring maximum core density in 1U
• HPE ProLiant DL385 Gen11 — 2U dual-socket AMD EPYC platform for high-core-count virtualization, containerized workloads, and GPU-accelerated deployments

HPE ProLiant ML Tower Servers

HPE ProLiant ML tower servers are designed for office and edge deployments where rack infrastructure is not available. Haink supplies HPE ProLiant ML350 Gen11 and ML110 Gen11 tower servers for enterprise and SMB environments.

HPE ProLiant BL Blade Servers (Legacy — succeeded by HPE Synergy)

HPE ProLiant BL blade servers and the HPE BladeSystem c-Class enclosure are a sunsetting (end-of-life) platform — HPE's successor for modular, composable compute is HPE Synergy (including the Synergy 480 Gen12 module). Haink supports the installed base with BL460c Gen10 blades, interconnects and enclosure spares from stock; new modular deployments should standardize on HPE Synergy.

HPE Synergy Composable Infrastructure

HPE Synergy is a composable infrastructure platform that allows organizations to dynamically allocate compute, storage, and networking resources as software-defined pools. HPE Synergy is used by enterprises and cloud providers requiring flexible, programmable infrastructure at scale.

HPE Synergy components available through Haink include:

• HPE Synergy 480 Gen11 Compute Module — dual-socket compute for virtualization and cloud workloads
• HPE Synergy 660 Gen11 Compute Module — dual-socket high-memory platform for in-memory databases
• HPE Synergy 12000 Frame — modular enclosure supporting up to 12 compute modules
• HPE Synergy Image Streamer — stateless boot and rapid provisioning
• HPE OneView management software — composable infrastructure management

HPE Servers for AI and High-Performance Computing

HPE supplies GPU-accelerated server platforms designed specifically for AI training, AI inference, and high-performance computing workloads. Haink delivers HPE AI infrastructure to organizations building machine learning platforms and large-scale compute environments in Hong Kong, Dubai, and Mainland China.

HPE AI and HPC platforms available through Haink include:

• HPE Cray XD670 — 5U AI server with 8× NVIDIA H100 or H200 SXM5 GPUs and dual 5th-gen Intel Xeon, for large-scale training and fine-tuning
• HPE Cray XD685 — dense GPU server with 8× NVIDIA B200 (liquid-cooled) or 8× AMD Instinct MI300X / MI325X on AMD EPYC, for frontier-model training
• HPE ProLiant DL380a Gen12 — 4U AI/GPU server (Intel Xeon 6) supporting up to 10 double-wide GPUs (L40S, H100 NVL, H200) or 8× RTX PRO 6000 Blackwell Server Edition, for inference and mixed AI
• HPE ProLiant DL380 Gen12 with NVIDIA GPU options — general-purpose AI inference and machine learning
• HPE Apollo 6500 Gen10 Plus (legacy) — prior-generation 8-GPU server; superseded by the Cray XD and DL380a Gen12 platforms (spares / refurb from stock)

HPE Storage Systems

HPE manufactures a range of enterprise storage platforms covering primary storage, all-flash arrays, hybrid storage, and backup infrastructure. Haink supplies HPE storage systems to data centers and enterprise environments across the APAC and MENA regions.

HPE Alletra Storage

HPE Alletra is HPE's cloud-native storage portfolio, designed for organizations requiring scalable, cloud-managed all-flash storage with predictive analytics and automated management.

• HPE Alletra Storage MP B10000 — the current flagship: a disaggregated, scale-out block array delivering mission-critical performance and 100% data availability at midrange economics; it anchors HPE GreenLake for Block Storage and is the platform HPE Primera, Nimble and 3PAR migrate to
• HPE Alletra 9000 — prior tier-1 all-flash array (mission-critical, sub-millisecond latency), now migrating to Alletra Storage MP B10000
• HPE Alletra 5000 / 6000 — Nimble-derived arrays for mixed-workload and secondary storage. Discontinued: last orders for the 6000 closed 30 November 2025 and the 5000 reaches obsolescence 31 October 2026 on the published schedule, with support to 2030. Both refresh to the B10000 — see what happened to 3PAR, Primera and Nimble

HPE Primera (Legacy — under the HPE Alletra umbrella)

HPE Primera is a tier-1 all-flash storage platform designed for enterprise databases, ERP systems, and business-critical applications requiring guaranteed availability and consistent low latency. HPE Primera delivers 100% availability guarantees backed by HPE's support commitments. New tier-1 deployments now use HPE Alletra Storage MP B10000; existing Primera arrays migrate to it non-disruptively with Peer Motion.

HPE Nimble Storage (Legacy — under the HPE Alletra umbrella)

HPE Nimble Storage provides adaptive flash storage with HPE InfoSight predictive analytics, which uses machine learning to predict and prevent infrastructure issues before they impact operations. Nimble Storage is widely deployed for mixed virtualization and database environments, and is now consolidated under the Alletra umbrella; new deployments use HPE Alletra Storage MP B10000 — the Alletra 5000 and 6000 are themselves discontinued — with non-disruptive Peer Motion migration from Nimble.

• HPE Nimble Storage AF-Series — all-flash arrays for performance-sensitive workloads
• HPE Nimble Storage HF-Series — adaptive flash arrays for mixed workload environments
• HPE Nimble Storage CS-Series — secondary storage for backup and disaster recovery

HPE StoreEasy NAS

HPE StoreEasy is HPE's network-attached storage (NAS) platform for file storage workloads including home directories, shared file storage, and archive. StoreEasy runs Windows Storage Server and integrates with Active Directory and Microsoft environments.

• HPE StoreEasy 1460 / 1560 / 1660 — 4U NAS appliances for SMB and departmental file storage environments
• HPE StoreEasy 1860 — high-capacity NAS for large-scale unstructured data and media storage

HPE MSA Storage

HPE MSA (Modular Smart Array) is an entry-level SAN and direct-attached storage platform designed for small and mid-size organizations. HPE MSA 2060 and MSA 1060 arrays are available through Haink for organizations requiring affordable block storage.

HPE StoreOnce

HPE StoreOnce is a backup appliance platform providing deduplication-based backup storage for enterprise backup and disaster recovery environments. HPE StoreOnce integrates with HPE RMC, Veeam, Commvault, and other leading backup software platforms.

HPE Edgeline Edge Servers

HPE Edgeline is HPE's portfolio of edge computing platforms designed for industrial, IoT, and operational technology (OT) environments where servers must operate outside traditional data center conditions.

• HPE Edgeline EL300 — compact converged edge system combining x86 compute with I/O expansion for industrial control and edge analytics
• HPE Edgeline EL4000 — modular edge computing platform for high-demand industrial AI and data acquisition workloads

HPE SimpliVity Hyperconverged Infrastructure

HPE SimpliVity is a hyperconverged infrastructure (HCI) platform that combines compute, storage, networking, and data protection into a single node. SimpliVity is built on HPE ProLiant hardware and is optimized for VMware vSphere environments.

HPE SimpliVity delivers built-in data deduplication and compression, VM-centric backup and recovery, and WAN optimization across distributed sites. Haink supplies HPE SimpliVity nodes for enterprise branch deployments, remote offices, and consolidated data center environments.

Related Resources

• Sourcing HPE — the four routes hardware reaches you, what a complete quote contains, and what to ask a supplier before committing
• Aruba vs Juniper under HPE — neither line is being killed, so which to standardise on
• HPE Alletra explained — what happened to 3PAR, Primera and Nimble, and the lifecycle dates
• HPE part numbers decoded — regional suffixes, factory-only options, and the -B21/-F21 memory trap
• HPE stock, lead times and pricing — current availability and firm quotes
• Dell PowerEdge vs HPE ProLiant — the two-way comparison of the incumbents
• Lenovo vs Dell vs HPE — the current generation compared three ways
• VMware exit — the hardware decision — what can be carried across and what has to be rebought
• Server refresh and consolidation — what a 4:1 consolidation actually saves
• Sizing an HCI cluster — resiliency overhead, rebuild reserve and platform floors
• Buying 20–60 servers as one project — how an order of this size is actually structured
• Stock vs BTO/CTO — what each route costs in time and money
• Gray-market channel risks — why serial verification before payment matters
• Enterprise storage selection — where Alletra sits against the market
• Dell PowerEdge selection guide · Lenovo ThinkSystem selection guide · all brand guides
• Sovereign cloud deployment — 40 nodes, 7,680 vCPUs, national data residency

How to Size a ProLiant Host — the Numbers

Sizing is mostly cores and RAM. Rules of thumb: ~1 physical core per 4–6 vCPUs at moderate load, 4–8 GB RAM per VM plus ~10–15% host overhead, and 20–30% headroom for N+1.

Worked example — a 150-VM consolidation at ~4 vCPU / 8 GB each: ~100–150 usable cores and ~1.3–1.5 TB RAM. A 2-socket DL380 Gen12 (2× Intel Xeon 6, up to 288 cores and 8 TB DDR5 in the pair) covers it with room; for HA you'd run two DL360 Gen12 1U hosts (N+1). Databases push toward more RAM and fewer VMs; AI toward GPU (DL380a Gen12) or Cray XD, and away from oversubscription. Rules of thumb, not guarantees — we size it precisely for your estate.

When HPE Isn't the Answer

We stay neutral, because a wrong standardization is costly:

• If you're standardized on Dell ProSupport or Lenovo, a single-vendor estate can beat mixing — HPE wins on iLO management, Synergy composability and GreenLake, not on being different for its own sake.
• For commodity GPU inference at scale, Supermicro and Dell XE are often cheaper per GPU; HPE's real AI strength is Cray XD670/XD685 for HPC and large training.
• If you want opex, not capex, HPE GreenLake (pay-per-use) may beat an outright purchase — we can quote both.
• Budget-constrained? ProLiant Gen11 from stock frequently beats new Gen12 on price for standardized workloads.

Five Expensive Mistakes We See in HPE Procurement

1. Buying Gen12 when Gen11 from stock was the better deal. For standardized virtualization and database builds that don't need the latest Xeon 6 core counts, Gen11 ships this week and costs less.
2. Designing new builds on BladeSystem c-Class. It's sunsetting — HPE's modular successor is Synergy (480 Gen12). Buy Synergy for new composable infrastructure; keep c-Class only for spares.
3. Ignoring the part-number suffix. A different suffix (e.g. -B21 vs -AA1) is the same box for a different region — match the base number, not the whole SKU (see the decoder above).
4. Skipping serial verification. Gray-market ProLiant may have no HPE warranty or support entitlement. Verify serials before payment.
5. Over-buying the storage tier. Primera and Nimble are legacy under the Alletra umbrella; so is Alletra 5000/6000. The current platform for both tier-1 and mixed/secondary is Alletra Storage MP B10000, sized by controller count rather than by model.

Frequently Asked Questions

Gen12 or Gen11 — which ProLiant should I buy now?

Buy Gen12 (Intel Xeon 6) when you need the latest core counts, memory bandwidth or efficiency, or for new AI/GPU builds (DL380a Gen12, Cray XD). Buy Gen11 from stock when the workload is standardized virtualization or databases — it ships this week and is usually the better-priced choice. Note Gen12 is currently Intel-only; AMD EPYC ProLiant remains on Gen11. We'll quote both side by side.

How do I read an HPE part number and avoid buying the wrong regional SKU?

An HPE part number is a base (e.g. P28948) plus a region/channel suffix: -B21 is the worldwide/most-regions code, -AA1 is China, -291 is Japan. The same server appears under several suffixes — identical hardware, different sourcing region. When comparing quotes, match the base number, not the suffix. We quote the region-appropriate SKU and confirm it against HPE QuickSpecs (full decoder above).

Where do you ship HPE from, and how fast?

ProLiant DL360/DL380 Gen11 configurations ship from Hong Kong stock (free port, no duty), often within days; Gen12, DL380a GPU, Synergy, Cray XD and Alletra are typically 2–12 weeks to order depending on configuration. We also ship from Dubai for the GCC/MENA and handle import documentation and customs for Mainland China.

What HPE is in stock now, and what are realistic lead times?

Popular ProLiant DL360/DL380 Gen11 configurations are typically in stock in Hong Kong (days–1 week); Gen12, GPU (DL380a, Cray XD), Synergy and Alletra Storage MP are usually 2–12 weeks to order. Stock moves — see the live table on our HPE supply page and request current availability for your models.

What is the difference between HPE ProLiant DL360 and DL380?

The HPE ProLiant DL360 Gen11 is a 1U single-processor-width rack server optimized for high-density compute environments where rack space is constrained. The HPE ProLiant DL380 Gen11 is a 2U platform offering greater storage capacity, more PCIe expansion slots, and support for higher GPU counts, making it the more versatile platform for general-purpose data center workloads, virtualization, and AI inference.