Skip to content

SAP HANA Hosting
on certified nodes

SAP HANA Hosting from data centres in Germany: 4 size classes are approved — HANA S, HANA M, HANA L and HANA XL — on 4 of 5 host classes. Licensing is per host and month, not per core; the datastore is billed separately. Approved means: an audit report according to SAP HWCCT (Hardware Configuration Check Tool) is available for exactly this configuration — but not for any other combination.

Nodes from €2,313.00$2,683.08 net per month: host, platform operations and licence. The datastore — the storage cluster where the HANA volumes reside — is billed separately. Approval at 2 locations. A node failure (P1) receives an initial response in platform operations within 30 minutes, around the clock; questions and changes go via the ticket in the account to support according to the SLA, in the base tier with a response within 8 h to a critical case, available Mon–Fri 08:00–18:00.

Approved sizes
4
HANA S to HANA XL — a separate report per class
Approved host classes
4 of 5
1 class is explicitly not approved
Memory
256 GB – 1 TB
physical per node, without overbooking
Log write latency
310 µs
measured · guaranteed ≤ 1000 µs

The database size determines the class. The class determines the volume layout, datastore quantity and price — calculated using the same catalogue values as the data sheet below.

Database size (column store) in GB

32 GB512 GB

What your system currently occupies in the column store. SAP calculates the class at double this amount, because HANA needs the second half of the memory for the delta merge (incorporating new data into the column store) and for intermediate results.

Reserve for growth

Requirement: 160 GB — HANA L supports up to 256 GB.

How much the database can grow without you having to change the class. Changing the class requires a new certification.

Processor

Certified on Intel Xeon: HANA M, database up to 128 GB. AMX are matrix math units in the processor core, designed for inference on the CPU.

Location

Only locations with a certified datastore — certification is only possible there. The location does not change the price.

Net monthly · HANA L€3,411.88$3,957.78Result

HANA L

PC-L128

EPYC 9554 · S/4HANA with connected BW, mixed OLTP and OLAP load

Memory
512 GB
vCPU, 1:1 to the core
64
Database up to
256 GB
Location
Düsseldorf · dus1

Volume layout

/hana/data
512 GB
/hana/log
256 GB
/hana/shared
512 GB
Total volume
1,280 GB
free up to the minimum commitment
768 GB
Node (host, platform operations, licence)
€3,188.65$3,698.83
Datastore NVMe Performance, 2,048 GB billed
€223.23$258.95
Net monthly
€3,411.88$3,957.78

All amounts net, plus 19 % VAT. Excluding operations tier and recovery — both are listed below under operations.

The 4 classes form a ladder, each with its own protocol. Under the certification, a 1:1 ratio applies: one vCPU, i.e. one virtual processor core, per physical core — without overprovisioning.

HANA S

PC-M64

S/4HANA development, quality assurance, small secondary systems

Memory
256 GB
of which allocatable
232 GB
vCPU 1:1
48
Database up to
128 GB

€2,313.00$2,683.08

per month, net · host, platform operations and licence; datastore separate

HANA M

PC-M64i

Productive S/4HANA in SMEs, Intel certification with AMX

Memory
256 GB
of which allocatable
232 GB
vCPU 1:1
48
Database up to
128 GB

€2,383.00$2,764.28

per month, net · host, platform operations and licence; datastore separate

HANA L

PC-L128

S/4HANA with connected BW, mixed OLTP and OLAP load

Memory
512 GB
of which allocatable
472 GB
vCPU 1:1
64
Database up to
256 GB

€3,188.65$3,698.83

per month, net · host, platform operations and licence; datastore separate

HANA XL

PC-X256

BW/4HANA with large columnar storage, consolidated corporate systems

Memory
1 TB
of which allocatable
960 GB
vCPU 1:1
384
Database up to
512 GB

€7,083.18$8,216.49

per month, net · host, platform operations and licence; datastore separate

Data sheet of the four HANA size classes: host class, processor, cores, clock speed, memory, allocatable vCPUs and monthly price per node, net
ClassHostProcessorCoresBase clockRAMallocatablevCPUper month
HANA SPC-M64EPYC 9454P482.75 GHz256 GB232 GB48€2,313.00$2,683.08
HANA MPC-M64iXeon 6741P482.50 GHz256 GB232 GB48€2,383.00$2,764.28
HANA LPC-L128EPYC 9554643.10 GHz512 GB472 GB64€3,188.65$3,698.83
HANA XLPC-X2562× EPYC 99653842.25 GHz1 TB960 GB384€7,083.18$8,216.49

Limit per host class

A NUMA node is a processor socket with the memory directly attached to it.

PC-M64HANA S
One socket, one NUMA node: VMs over 232 GB RAM no longer fit on this class. For wider single VMs, PC-L128 or PC-X256 is the right fit.
PC-M64iHANA M
Live migration is only guaranteed between hosts of the same CPU generation. A mixed cluster of PC-M64 (AMD) and PC-M64i (Intel) does not migrate during operations.
PC-L128HANA L
The boost ends at 3.75 GHz — the class wins on width, not on clock speed. Loads that depend on single threads run faster on PC-S32 with 5.7 GHz.
PC-X256HANA XL
Two sockets mean two NUMA nodes. VMs over 480 GB RAM or over 192 vCPU exceed one node and lose around 12% memory bandwidth — with licences per core, this host also pays for 384 cores.

Why allocatable memory is below the class designation

The hypervisor, the virtualisation layer of the host, requires memory itself: with PC-M64, 232 GB of 256 GB remain for the guest. SAP's volume rules still calculate using the physical configuration. The database itself may only occupy half, here 128 GB: HANA needs the other half for intermediate results and the delta merge, the incorporation of new write operations into the column store.

Certification possible in 2

  • fra1Frankfurt am Main€2,313.00$2,683.08
  • dus1Düsseldorf€2,313.00$2,683.08

The price per location shows HANA S with location factor, net and without datastore.

The configuration is certified, not the product

A node is considered accepted if an HWCCT report is available for exactly this configuration: host, memory, datastore and volume slice. You receive it upon handover and anew after every change.

  1. 01

    Define configuration

    Class · location · datastore

    You choose one of the 4 classes, one of the 2 locations and the certified Datastore NVMe Performance. All of this is listed in the catalogue on this page.

  2. 02

    Delivery and HWCCT run

    on exactly this configuration

    The node is delivered as defined and tested with the HWCCT. 4 storage metrics are measured against fixed limits — for example, ≤ 1000 µs for log write latency.

  3. 03

    Report upon handover

    per node

    You receive the report upon handover. From then on, the storage metrics, or KPIs for short, are contractually guaranteed — not just measured.

  4. 04

    Change means new acceptance

    RAM · datastore · location · class

    If the memory, datastore, location or class changes, a new acceptance is required. Until the new report is available, the guarantee is suspended.

  • Acceptance protocol according to SAP HWCCT per node
  • Storage KPIs contractually guaranteed instead of just measured
  • 1:1 mapping of vCPU to physical core, no memory overcommit
  • Separate volumes for /hana/data, /hana/log and /hana/shared

ENTRONYX SAP Certified

per host and month

Certified HANA nodes, licensed per host — without overcommitment.

Licence model
€1,290.00$1,496.40 per host and month
Nodes per environment
1 to 8
Approved host classes
4 of 5
vCPU per core
1:1
Management
vCenter plus SAP Host Agent, LaMa connection on request
Maintenance windows
By arrangement per environment, 21 days lead time, outside your closing periods

The licence is tied to the host, not the number of cores. A node with 384 cores costs the same in terms of licensing as one with 48.

Fully populated memory banks of an open server in bright light; each bank is fitted with a memory module.

Exactly 4 combinations are accepted:

  • HANA SonPC-M64
  • HANA MonPC-M64i
  • HANA LonPC-L128
  • HANA XLonPC-X256

Another combination is not prohibited, just not documented. The ENTRONYX CLOUD guarantee does not apply to it.

One host class is not approved

The delivered configuration is always approved, not a single metric. That is why the justification is per class.

PC-S32

PC-S

3 Gbit uplink without a separate storage fabric. Shared NFS or iSCSI datastores are not permitted on this class so that replication and guest traffic do not share the same line. Without a shared datastore, there is no automatic HA restart.

Three volumes, three rules, no interpretation

SAP dictates the layout, not us. The three file systems follow from the memory of the class — and thus the quantity ordered on the datastore.

/hana/data
1 × RAM
At the savepoint, HANA writes the column store completely here. The volume therefore holds the entire memory.
/hana/log
min(0.5 × RAM; 512 GB)
The redo log, the log of all write operations for the restart, is overwritten cyclically. More than the cap does not provide an additional recovery point.
/hana/shared
min(1 × RAM; 1,024 GB)
Program files, trace files and profiles. The volume grows with the instance, but not with the data — hence the cap.
Database up to
RAM ÷ 2
The column store may occupy around half. HANA needs the other half for intermediate results and the delta merge, the incorporation of new write operations into the column store.

Volume per class, in GB

  • HANA S

    640 GBbilled 2,048 GB

    /hana/data 256 · /hana/log 128 · /hana/shared 256

  • HANA M

    640 GBbilled 2,048 GB

    /hana/data 256 · /hana/log 128 · /hana/shared 256

  • HANA L

    1,280 GBbilled 2,048 GB

    /hana/data 512 · /hana/log 256 · /hana/shared 512

  • HANA XL

    2,560 GBbilled 2,560 GB

    /hana/data 1,024 · /hana/log 512 · /hana/shared 1,024

3 of the 4 classes are below the minimum commitment and pay the same amount. Only HANA XL exceeds it.

Volume layout and storage costs per size class. Quantities in GB so that the sum remains verifiable; the calculation is below each value. Net amounts per month on the Datastore NVMe Performance.
ClassRAM/hana/data/hana/log/hana/sharedTotalbilledper monthwith snapshots
HANA S256 GB256 GB1 × 256128 GBmin(128; 512)256 GBmin(256; 1,024)640 GB2,048 GBMinimum commitment€223.23$258.95€266.24$308.84
HANA M256 GB256 GB1 × 256128 GBmin(128; 512)256 GBmin(256; 1,024)640 GB2,048 GBMinimum commitment€223.23$258.95€266.24$308.84
HANA L512 GB512 GB1 × 512256 GBmin(256; 512)512 GBmin(512; 1,024)1,280 GB2,048 GBMinimum commitment€223.23$258.95€266.24$308.84
HANA XL1,024 GB1,024 GB1 × 1,024512 GBmin(512; 512)1,024 GBmin(1,024; 1,024)2,560 GB2,560 GB€279.04$323.69€332.80$386.05
NVMe drive shelf from the front: all carriers are occupied, a green status light is on at each one.
The Datastore NVMe Performance charges at least 2,048 GB. Each GB costs €0.1090$0.1264 per month. Snapshots cost an additional €0.0210$0.0244 per GB and month.

The evidence on which the acceptance is based

SAP checks 4 storage metrics in the HWCCT run. The limit value is guaranteed; the measured value comes from the certification run and proves it. A new measurement is taken after every change to the configuration or datastore.

Limit, measured value and the gap between them

Log write latency, 4 KB
guaranteed≤ 1000 µsmeasured310 µs
3.2 ×Reserve
Log write latency, 16 KB
guaranteed≤ 1000 µsmeasured420 µs
2.4 ×Reserve
Data write throughput, 16 MB
guaranteed≥ 400 MB/smeasured2140 MB/s
5.4 ×Reserve
Data read throughput, 16 MB
guaranteed≥ 400 MB/smeasured3060 MB/s
7.7 ×Reserve
Rear of a server with plugged-in fibre optic cables, neatly bundled into strands.
The volumes are not located in the node, but in the datastore — connected via NVMe/TCP or NFS 4.1. Triply replicated across two fire zones of the location. This connection carries the guaranteed p99 latency of 0.4 ms: 99% of all accesses are faster.

4 datastore classes can be booked, one is certified

Datastore NVMe Performance

All 4 classes can be created under the SAP stack. This is useful for templates, exports and powered-off secondary systems. The three volumes of a productive instance, however, belong on the Datastore NVMe Performance.

Datastore classes under the SAP stack — certification, performance commitment and price
FeatureNVMe PerformanceNVMe/TCP · NFS 4.1NVMe StandardNFS 4.1 · iSCSICapacityiSCSI · NFS 4.1ArchiveNFS 4.1
SAP acceptanceHWCCT log per node−−−
IOPS per datastoreAccesses per second: 4 KB random, 70/30 read/write120,00045,0009,000900
Throughput6,000 MB/s2,400 MB/s1,100 MB/s350 MB/s
Latency p9999% of all accesses are faster than this value0.4 ms0.9 ms4.0 ms25.0 ms
Price per GB and month€0.1090$0.1264€0.0720$0.0835€0.0290$0.0336€0.0068$0.0079
Minimum commitment2,048 GB1,024 GB8,192 GB20,480 GB

Backup windows, redundancy and snapshots of all classes can be found under Private Cloud storage.

Row of identical server nodes in a bright aisle, each slot equipped identically.

An environment comprises 1 to 8 nodes. Each node is accepted individually and receives its own report. Acceptance in Frankfurt am Main (fra1) and Düsseldorf (dus1).

Maintenance is not on our calendar, but on yours

An approved node is not maintained in the fixed weekly window. Every change to firmware, hypervisor or datastore affects the approval. Therefore, it is agreed individually.

Maintenance windows

By arrangement per environment, 21 days lead time, outside your closing periods

Nodes per environment
1 to 8
Change lead time
8 to 48 h
Management
vCenter plus SAP Host Agent, LaMa connection on request

The maintenance window is the agreed period during which work may be performed on a node. The lead time depends on the booked operations level. You define closing periods, inventories and payment runs once; no window will be scheduled during these times.

Operations levels with initial response per incident class (P1: highest urgency, P2: second highest), availability, change lead time and net price per month — per environment and per host
TierInitial response P1Initial response P2AvailabilityLead timeper environmentper host
Platform operations30 min4 hP1 around the clock, P2 and P3 Mon–Fri 08:00–18:0048 hincludedincluded
Platform and guest operations15 min2 hP1 and P2 around the clock, P3 Mon–Sat 08:00–20:0024 h€690.00$800.40€39.00$45.24
Full operations15 min1 hAround the clock via fixed phone number, named technical contact8 h€2,490.00$2,888.40€89.00$103.24

Prices apply per environment, plus a rate per host. Operations effort grows with the number of systems, not just with the number of nodes. The complete model can be found under Managed dedicated hosts.

Limit per operations level

Platform operations
Escalation: NOC → Platform engineer on duty → Infrastructure on-call management
Firmware maintenance requires a host evacuation. Without a second host in the cluster, this means downtime of up to 45 minutes per maintenance window.
Platform and guest operations
Escalation: NOC → System engineer on duty → Named Service Manager → Infrastructure on-call management
Guest patches are only applied to images delivered by ENTRONYX. Custom-built images remain your responsibility and are excluded from the availability commitment.
Full operations
Escalation: Fixed phone number → Technical Account Manager → Operations Management → Technical Management
Application operations include the services listed in the operations manual. In-house developments not listed are monitored but not modified.

Common questions before purchase

9 questions that regularly come up before an enquiry. The answers use the same catalogue values as the sections above.

Can't find your question? Send us your configuration.

Is the certified HANA hardware located in Germany?

Yes: The approved host classes and the certified datastore are located in data centres in Germany, across 2 locations. Acceptance testing is performed on exactly this equipment on site; location, host class and datastore are stated in the report.

Limits and measured values of the metrics

How quickly does support respond if the node fails?

An outage (P1) is covered by the 'Platform operations' tier, which is included with every node: initial response in 30 minutes, around the clock. Questions and change requests are handled via a ticket in the account to support according to the SLA; the basic tier is included in the price, responds to a critical case within 8 h and is available Mon–Fri 08:00–18:00. Initial response means: acknowledged, classified, status reported — not a promise of resolution.

Operation tiers and initial response times

What is the term, and how do I cancel?

Without a term agreement, the contract runs for one month and renews for another; terms of 12 or 24 months are paid in advance and you receive a 5% or 8% term discount respectively. You can cancel monthly contracts with one day's notice to the end of the term, contracts for 12 or 24 months with 30 calendar days' notice — otherwise they renew for twelve months. After the contract ends, your data remains accessible for 30 calendar days; after that, it is irretrievably deleted.

What does “accepted” mean?

An SAP HWCCT test report is available for exactly this configuration — host, memory, datastore and volume layout. From handover, the storage metrics are contractually guaranteed, not just measured.

Limits and measured values of the metrics

Can I run a size other than the 4 classes?

Yes, it is not forbidden. However, there is no report for such a combination, and the guarantee of the storage metrics does not apply to it.

Request a different size

Why do HANA S, HANA M and HANA L pay for the same amount of storage?

The certified Datastore NVMe Performance charges at least 2,048 GB. The three volumes of these classes add up to 640 to 1,280 GB and are therefore below this. Only HANA XL exceeds the minimum allocation.

Volume layout in the picture

What happens if there is a change to RAM, datastore or location?

Any such change requires a new acceptance test on the modified configuration. Until the new report is available, the guarantee of the storage metrics is suspended.

The path to acceptance

At which locations is acceptance possible?

In Frankfurt am Main (fra1) and Düsseldorf (dus1): The approved host classes and the certified datastore are located there.

How is it licenced?

Per host, not per core: A host with 384 cores costs the same in terms of licences as one with 48. An environment comprises 1 to 8 nodes, and each node carries its own licence.

Licence per host and month, net€1,290.00$1,496.40

Determine size class, schedule acceptance

The class determines the volume layout, datastore and price. The report is generated on the delivered configuration and is provided to you upon handover. A change to the class, location or datastore requires a new acceptance.

Size classes
4
Nodes from
€2,313.00$2,683.08
vCPU per core
1:1
Log write latency
310 µs