The HPE ProLiant Buying Guide

Buying guideUniqcli TeamMarch 28, 202612 min read

HPE ProLiant servers power data centers, government agencies, hospital systems, and university campuses across virtually every vertical. The portfolio has been continuously developed over three decades, and as of mid-2026 it spans two active generations — Gen11 and Gen12 — covering single-socket edge appliances through quad-socket scale-up systems. That breadth is genuinely useful, but it also means choosing the wrong model is easy when you are navigating dozens of SKUs, three processor families, and multiple storage architectures at once.

This guide is designed to give procurement teams, IT architects, and department heads a vendor-honest, decision-focused framework. We cover naming conventions, form factors, processor tradeoffs, workload alignment, security architecture, and total-cost factors — everything you need to walk into a quoting conversation with confidence. Where relevant, we note differences between Gen11 and Gen12 so you can make an informed generational choice for your budget and timeline.

Understanding the ProLiant Naming Convention

Before evaluating specific models, it helps to understand how HPE encodes information in the product name itself.

DL (Density Line): Rack-optimized servers designed for 19-inch racks. The digits indicate socket count and tier: DL3xx is typically 1–2 socket mid-range rack, DL5xx is 4-socket scale-up.
ML (Maximum Line / Tower): Tower servers built for sites without dedicated rack infrastructure — branch offices, clinics, schools, and wiring closets. Most ML towers can be optionally racked.
MicroServer: Ultra-compact single-socket platforms aimed at the edge, lightweight virtualization, or cost-constrained deployments.
Synergy: Composable infrastructure housed in HPE Synergy frames. Technically a distinct product family, but it shares the same ProLiant firmware, iLO management, and security architecture.

The generation number — Gen11 or Gen12 — is arguably the most consequential part of the name. It determines which processor families are supported, which version of iLO manages the system, and what security capabilities are available at the silicon level. Both generations are currently active and orderable; neither is end-of-life.

Gen11 vs. Gen12: What Changed at the Platform Level

HPE introduced the ProLiant Compute Gen12 family in February 2025, initially with Intel Xeon 6 processors. The first AMD-based Gen12 models (DL325 and DL345 Gen12 using 5th Gen AMD EPYC) followed in mid-2025. For buyers evaluating a refresh or new deployment today, here is what materially changed between the two generations:

Capability	Gen11	Gen12
Management firmware	iLO 6	iLO 7
Silicon Root of Trust scope	Firmware chain	Extended to option cards via SPDM
Quantum-resistant cryptography	Not included	NIST / CNSA 2.0 compliant
Secure Enclave / FIPS 140	FIPS 140-2	FIPS 140-3 Level 3 Secure Enclave
Primary Intel platform	4th/5th Gen Xeon Scalable	Intel Xeon 6 (P-core and E-core)
Primary AMD platform	EPYC 9004 / 9005 (Genoa / Turin)	EPYC 9005 (Turin)
PCIe generation	Gen5	Gen5
Max GPU density (DL380a)	4 GPUs (DL380 Gen11)	8 GPUs (DL380a Gen12)
Federal security posture	Strong	Quantum-resistant, TAA-upgradable

Key takeaway: Gen11 remains an excellent platform for most enterprise, SLED, and healthcare workloads and will be under active HPE support for years. Gen12 is the right choice when you need the most future-proof security posture (federal, CMMC, FedRAMP High), the latest Intel Xeon 6 architecture, or maximum AI inference density. For cost-sensitive refreshes with a 3–5 year horizon, Gen11 often delivers better price-to-performance at current street prices.

Explore currently available configurations on our HPE ProLiant compute page.

Form Factor Selection: DL Rack, ML Tower, or MicroServer

The form-factor decision is upstream of everything else — it constrains cooling, power delivery, drive bay count, and GPU expansion capacity.

Choose a DL rack server when:

You have an existing or planned 19-inch equipment rack
The server will live in a data center, colocation cage, or wiring-closet rack
You need high drive counts, multiple GPUs, or shared rack cooling infrastructure
Remote lights-out management matters more than acoustics

Choose an ML tower server when:

The deployment site has no rack (classroom, exam room, branch office)
Acoustics matter — tower chassis run significantly quieter than rack units under load
Budget is constrained and the simpler chassis reduces upfront cost
You want the flexibility of an ML350 Gen11/Gen12, which converts between tower and rack-mount chassis

Choose a MicroServer when:

Minimal physical footprint is the primary constraint — retail kiosk, healthcare bedside, IoT gateway
Workloads are lightweight: file sharing, domain controller, agent-based monitoring, light virtualization

Processor Family: Intel Xeon vs. AMD EPYC

Both processor families are first-class citizens across the ProLiant portfolio, and HPE maintains active investment in AMD EPYC models. The right choice depends on workload type, software licensing model, and memory-bandwidth requirements.

Intel Xeon Scalable (Gen11) / Intel Xeon 6 (Gen12)

Best for: VMware vSphere, Microsoft Hyper-V, SQL Server, Oracle Database, and other workloads where commercial per-core software licensing incentivizes lower-core-count CPUs
Gen11 models: DL320, DL360, DL380, DL560, DL580 Gen11 — supporting 4th and 5th Gen Intel Xeon Scalable processors
Gen12 models: DL320, DL340, DL360, DL380, DL380a, ML350, DL580 Gen12 — powered by Intel Xeon 6, supporting up to 144 E-cores or 86 P-cores per socket
Memory: Up to 8 TB DDR5 at up to 5600 MT/s on dual-socket Gen11 and Gen12 platforms

AMD EPYC (Gen11 and Gen12)

Best for: High core count requirements, memory-bandwidth-intensive workloads (HPC, EDA, in-memory databases, Redis), Linux-native environments, and scenarios where per-VM or subscription licensing eliminates per-core penalties
Gen11 models: DL325 Gen11 (1P, 1U), DL345 Gen11 (1P, 2U), DL365 Gen11 (2P), DL385 Gen11 (2P) — supporting AMD EPYC 9004 (Genoa) and 9005 (Turin) series up to 160 cores per socket
Gen12 models: DL325 Gen12 and DL345 Gen12 — featuring 5th Gen AMD EPYC with approximately 2x the memory capacity versus Gen11 predecessors
Core density highlight: The DL385 Gen11 in a dual-socket configuration can reach extremely high core and thread counts, making it a standout for virtualization consolidation projects where per-VM licensing is in play

Matching ProLiant Models to Common Workloads

The table below maps frequently encountered workload types to the ProLiant models most commonly deployed for each, along with a brief rationale.

Workload	Recommended Model(s)	Why
General virtualization (VMware, Hyper-V)	DL360 Gen11/Gen12, DL380 Gen11/Gen12	Proven ecosystem support, balanced compute and I/O
VDI / desktop virtualization	DL345 Gen12, DL385 Gen11	High memory bandwidth; AMD EPYC economics with per-VM licensing
AI inference / GPU workloads	DL380a Gen12	Up to 8 NVIDIA GPUs; PCIe Gen5 fabric
SQL Server / Oracle OLTP	DL360 Gen11, DL380 Gen11/Gen12	Intel Xeon; favorable per-core licensing economics
HPC, simulation, EDA	DL385 Gen11, DL365 Gen11	High EPYC core count and memory bandwidth
Branch office / ROBO	ML350 Gen11/Gen12, MicroServer Gen11	Tower-to-rack convertible; quiet operation
Edge / healthcare bedside	MicroServer Gen11, DL145 Gen11	Low power, compact, suitable for edge deployments
High-capacity NVMe storage	DL380 Gen11 (EDSFF config)	Up to 36 E3.S NVMe drives; excellent IOPS density
Scale-up ERP / in-memory DB	DL580 Gen12, DL580 Gen11	4-socket platform; massive memory capacity

For regulated workloads (federal, healthcare, SLED), also factor in the security requirements described in the next section before finalizing a model.

Storage Architecture: Drive Types, Controllers, and External Options

Storage configuration is the decision buyers most often regret when they get it wrong — drive bay counts and controller choices are difficult to change after deployment.

Drive Form Factors

LFF (Large Form Factor, 3.5"): Higher raw capacity per drive slot; best for NL-SAS or large spinning media tiers where $/TB matters more than IOPS
SFF (Small Form Factor, 2.5"): More bay slots in the same chassis depth; standard for all-flash SAS or SATA SSD configurations
EDSFF / E3.S: The emerging enterprise NVMe SSD form factor; the DL380 Gen11 supports up to 36 E3.S drives in the front bays, delivering extremely high sequential throughput and IOPS density for database and AI training workloads
NVMe vs. SAS/SATA: NVMe dramatically reduces storage latency and is the right choice for any database, VDI, or AI training workload where I/O is a bottleneck; SAS/SATA remains cost-effective for bulk and archival tiers

RAID and Storage Controllers

HPE SmartArray controllers: The standard choice; integrates natively with iLO for health monitoring, predictive failure alerting, and Compute Ops Management baselines
HPE MR controllers (Broadcom MegaRAID-based): Available for environments that require specific RAID feature sets or third-party management compatibility

External and Shared Storage

ProLiant servers integrate equally well with HPE's external storage portfolio — HPE MSA arrays for entry-level SAN/NAS, HPE Alletra MP for enterprise block and file, and NVMe-oF fabrics for disaggregated storage architectures. See our Uniqcli storage guides for HPE storage selection guidance.

Security Architecture for Regulated Buyers

If you support federal, DoD, SLED, or healthcare workloads, the security architecture of the server platform matters as much as raw performance. HPE ProLiant has genuine, hardware-rooted differentiators in this area.

HPE Silicon Root of Trust anchors server firmware integrity to a physically unclonable fingerprint embedded in the iLO ASIC during manufacturing. This fingerprint cannot be replicated by software updates or altered by supply-chain tampering. Even if an attacker compromises BIOS firmware or an option card, the server detects the violation and can recover to a known-good state. Silicon Root of Trust has been a ProLiant feature since Gen10; Gen11 extended it through SPDM authentication of peripheral devices, and Gen12 extends it further with iLO 7's Secure Enclave.

Gen11 security highlights (iLO 6):

Silicon Root of Trust with runtime firmware verification (continuous, not just at boot)
FIPS 140-2 validated cryptographic modules
Secure System Erase aligned with NIST SP 800-88 sanitization standards
TPM 2.0 for BitLocker, measured boot, and remote attestation workflows
SPDM-based component authentication for PCIe devices

Gen12 additional security capabilities (iLO 7):

FIPS 140-3 Level 3 Secure Enclave for cryptographic key storage and operations
Quantum-resistant cryptography compliant with NIST Post-Quantum standards and CNSA 2.0
Enhanced Silicon Root of Trust extended to option cards at boot and runtime
HPE Trusted Supply Chain option (servers assembled in U.S. facilities by vetted personnel) for TAA compliance — critical for federal procurement under FAR/DFARS

For CMMC Level 2/3, FedRAMP High, or HIPAA-covered environments, specifying Gen12 with the Trusted Supply Chain option gives you the most defensible security posture available from any major server vendor today.

Explore your options or get a quote with compliance requirements noted.

Lifecycle Planning: Gen11, Gen12, and Long-Term Support

A common concern during ProLiant evaluations is whether purchasing Gen11 today means buying into an aging platform. The short answer is no. Both generations are active product lines with full hardware warranty and HPE Pointnext support availability. HPE typically supports each server generation for a multi-year lifecycle after commercial introduction, so Gen11 hardware ordered in 2025–2026 will remain under active support well into the next decade.

That said, Gen12 provides a longer runway to end-of-engineering, which matters for organizations with 7–10 year asset lifetimes (common in government and healthcare). It also provides forward compatibility with Intel Xeon 6's roadmap and the iLO 7 management architecture that will underpin future HPE software investments.

HPE Compute Ops Management — HPE's cloud-based fleet management SaaS — supports both Gen11 and Gen12, enabling mixed-generation fleets to be managed, patched, and baselined from a single dashboard without on-premises management servers. This makes phased migration from Gen11 to Gen12 operationally straightforward.

Platform-specific availability note: At time of writing, the DL580 Gen12 (4-socket) and HPE Synergy 480 Gen12 blade modules are the most recently available additions to the Gen12 family. If your workload requires 4-socket scale-up, verify current availability and lead times with your Uniqcli account team before committing to a configuration.

Total Cost of Ownership: Beyond List Price

List price is rarely the right procurement metric for server infrastructure. These TCO levers frequently change the rank-order of options:

Power and cooling efficiency: Gen11 rack servers deliver measurably better performance-per-watt versus Gen10 predecessors. At scale, power savings compound significantly over a 3–5 year asset life — relevant for federal agencies and healthcare systems with sustainability mandates or constrained power envelopes.
Software licensing interaction: Workloads running commercial software with per-core licensing (SQL Server Enterprise, Oracle SE2, vSphere Foundation) can see software costs vary dramatically between a 32-core and a 64-core platform even when hardware prices are similar. Model your software costs before finalizing CPU selection.
HPE GreenLake: HPE's as-a-service consumption model allows compute capacity to be provisioned on a metered, pay-as-you-grow basis — reducing upfront capital expenditure and shifting to operational expense. Uniqcli can structure GreenLake agreements alongside outright hardware purchases.
Trade-in programs: HPE offers trade-in credit for retiring Gen9 and Gen10 assets. Factoring residual value into the refresh business case often meaningfully reduces the net cost of Gen11 or Gen12 investment.
Care Pack and Pointnext support tiers: Options range from next-business-day parts replacement to 24x7 4-hour onsite response with a proactive care advisor. For healthcare and federal environments with uptime SLAs, underestimating support costs is a common mistake.

When you request a quote through Uniqcli, our team models these variables — not just the hardware line items — so you have an accurate TCO picture before the purchase order is signed.

How Uniqcli Helps

Uniqcli is an authorized HPE and HPE Aruba Networking partner with experience configuring ProLiant servers for federal agencies, SLED institutions, healthcare systems, and enterprise IT organizations. We do not start from a product catalog — we start from your workload profile, compliance requirements, existing infrastructure, and budget horizon.

If you know exactly what you need, start with our HPE ProLiant compute catalog or visit the Uniqcli shop for current pricing and configurations. If you are still scoping a project — comparing Gen11 versus Gen12, Intel versus AMD, or evaluating GreenLake against a traditional capital purchase — contact our team or request a detailed quote. We will build a bill of materials that covers processors, memory, storage, networking, and support, with no surprises after the purchase order is signed.