Skip to content

SAP SOLUTIONS

SAP HANA Infrastructure

High-memory, dedicated servers for your S/4HANA and BW/4HANA databases. Separate data and log volumes on NVMe, a synchronous secondary with HANA System Replication, and automatic failover with Pacemaker; all in our Trabzon data center, in-country and monitored 24/7.

  • 6 TB+memory classes
  • HSRHA with Pacemaker
  • NVMedata · log separation
  • Trabzonlocal data · KVKK

WHY DEDICATED HANA INFRASTRUCTURE?

An in-memory database needs unshared resources

SAP HANA keeps all working data in memory and writes only the log to disk on each transaction. That is why the amount of memory, CPU-to-memory proximity and log write latency translate directly into user experience.

Guaranteed memory, no overcommitment

The memory allocated to production HANA is never shared with another workload. CPU and memory placement follows the NUMA layout, so queries are not affected by other customers' load.

Verified storage performance

Data and log volumes are kept separate on NVMe. After installation, storage is measured with SAP's HANA hardware check tool (HCMT) and the results are reported to you.

High availability ready

With a synchronous secondary HANA, a Pacemaker cluster with fencing, and a virtual IP, a primary server failure is handled within minutes without data loss.

Predictable latency

Application servers and HANA communicate within the same facility over a dedicated low-latency network. Replication traffic flows over a separate network card and VLAN.

Monitoring that understands HANA

Memory consumption, log volume utilization, savepoint durations, replication status and the backup chain are tracked by the 24/7 NOC; when a threshold is exceeded, we step in before you notice.

Classes ready for growth

When your memory needs grow, moving to the next class up is done via HSR in a planned way with a short outage. Temporary capacity can be added for the test phases of a transformation project.

ARCHITECTURE OPTIONS

The right HANA layout for every system

Production, analytics and non-production systems have different needs. We choose the architecture based on system role, data volume and SAP's current support conditions.

Recommended for production

Scale-up · physical server

The entire database runs on a single high-memory server. It is the simplest, fastest and most common architecture for transaction-heavy systems such as S/4HANA.

Memory
256 GB – 6 TB+
Socket
2 · 4 sockets
Protection
HSR + Pacemaker
Large analytics

Scale-out · multi-node

Data is distributed across multiple servers, with an optional standby node. Designed for BW/4HANA systems whose data volume exceeds several terabytes.

Node
2 – 8 workers + standby
Network
Separate internal HANA network
Storage
Shared /hana/shared
DEV · QAS · Training

Virtual HANA

Development, test and training systems run on supported virtualization configurations; capacity is provisioned quickly and system copies are easily refreshed.

Memory
64 GB – 1 TB
Setup
From a template · within hours
Copy
Planned system refresh

Production systems are installed on enterprise servers listed in SAP's certified hardware lists for HANA, with the relevant SAP Note settings. Virtualized production HANA is considered only in configurations supported by SAP and within size limits.

HANA MEMORY CALCULATOR

Today's database, tomorrow's HANA memory

Enter the size and type of your current database to get a rough view of the HANA memory you will need once compression, working space and growth are factored in, plus the recommended class and storage volumes.

2 TB

Database size used on disk (excluding log and free space).

Source database type

Estimated compression: 5×

Set the compression ratio manually

Use this if your sizing report produced a different ratio.

Workload

A single server (scale-up) is recommended for transaction-heavy systems.

Planning horizon

%15

HANA memory requirement by year

  • estimated requirement
  • recommended class

How is it calculated? Compressed data = source size ÷ compression ratio. HANA memory is the sum of twice the compressed data (data + query working space) and a fixed allowance for code and system tables. This is a rough rule of thumb and does not replace SAP's sizing reports.

MEMORY CLASSES

HANA servers from 256 GB to over 6 TB

Each class is configured with a balanced mix of memory, CPU and NVMe data/log volumes. Pricing is quoted based on HA/DR scope and management model.

HANA · S

256 GB

Small S/4HANA production systems, Solution Manager and large QAS copies.

Processor
2 sockets · 32 cores
Data / log
1 TB / 256 GB NVMe
Typical use
~50–150 users

HANA · M

512 GB

Mid-sized S/4HANA production systems and small BW/4HANA installations.

Processor
2 sockets · 48 cores
Data / log
2 TB / 256 GB NVMe
Typical use
~150–300 users

HANA · XL

2 TB

High-volume S/4HANA and mid-sized BW/4HANA production systems.

Processor
2 sockets · 96 cores
Data / log
8 TB / 512 GB NVMe
Typical use
~600–1,200 users

HANA · 2XL

3 TB

Multi-company holding structures and large analytics workloads.

Processor
4 sockets · 112 cores
Data / log
12 TB / 512 GB NVMe
Typical use
1,000+ users

HANA · 4XL+

6 TB+

The largest S/4HANA systems and scale-out BW/4HANA nodes; custom-designed per project.

Processor
4 – 8 sockets
Data / log
NVMe / SAN depending on the project
Typical use
Enterprise scale

All HANA servers are hosted in our Trabzon data center in compliance with KVKK (Turkish Personal Data Protection Law). Virtual HANA options for DEV/QAS start at 64 GB.

SYSTEM REPLICATION SIMULATOR

What happens when the primary goes down?

Choose the replication mode, then start a failure scenario. Follow step by step how Pacemaker brings the secondary into service and how application servers reconnect through the virtual IP.

Normal operation

APPLICATION LAYER

PASAAS 1AAS 2

connected · VIP 10.20.0.50

VIP → hana01
HALL APRIMARY

hana01

running · read/write

HALL BSECONDARY

hana02

replica · ready in memory

ISOLATED ZONEDR COPY

hana03

asynchronous replica

> pacemaker · event log

Cluster is healthy. Select a failure scenario.

The simulation is sped up; the times on screen represent the sequence in a typical cluster, and the values are examples.

STORAGE & PERFORMANCE

Every commit depends on log write speed

HANA processes data in memory, but to make each transaction durable it writes to the log volume and takes savepoints at regular intervals. Log latency determines the response time users feel, while the speed of the data volume determines restart and backup duration.

  • Data and log volumes run on separate NVMe pools without affecting each other's I/O load.
  • Latency and throughput values measured with HCMT after installation are included in the handover report.
  • The file system, block size and I/O parameters are tuned according to the relevant SAP Notes.
HANA storage volumes and typical sizes
Space What does it hold? Typical size Layer
/hana/dataPersistent data, savepoints≈ 1.2 × memoryNVMe · high throughput
/hana/logTransaction logs (redo)½ × memory, up to 512 GBNVMe · low latency
/hana/sharedInstallation, monitoring, trace files1 × memory, up to 1 TBNVMe / shared
/hana/backupLocal backup staging area (optional)Depends on the backup strategyBackup system · Backint

LOG WRITE

< 1 ms

target latency · 4 KB

SEPARATION

Data ≠ log

separate NVMe pools

VALIDATION

HCMT

in the handover report

OPERATIONS & MANAGED SERVICE

Who does what, clear from the start

In managed HANA operations, the database platform is with our team, while the application and processes are with you and your SAP partner. The table below shows a typical division of responsibilities.

Division of responsibilities in managed HANA operations
Task NuageDC Customer SAP partner
Server, NVMe storage, network, facilityResponsible——
SLES / RHEL for SAP installation, SAP Note settingsResponsible——
HANA installation and HCMT validationResponsibleApproval—
HSR, Pacemaker cluster and failover testsResponsibleApproval—
HANA revision and support package upgradesResponsibleApprovalTesting
Backint backups and restore testsResponsibleApproval—
24/7 monitoring, alerting and incident responseResponsible——
System copy / refresh (QAS, training)ResponsibleRequestPost-copy steps
SQL and performance analysisShared—Shared
Module customization, data model, reports—ResponsibleResponsible

It is also possible to take only the infrastructure and keep HANA operations with your own team. For the full list of service models, see the SAP Infrastructure page.

NON-PRODUCTION SYSTEMS

Test data always current, costs under control

DEV, QAS and training systems run on virtual HANA. Refreshing QAS with production data becomes a planned, repeatable workflow; test cycles run on realistic data.

For tables containing personal data, a masking step can be planned after the copy; this step is defined together with your partner according to your data protection policy.

  1. 01

    Selecting the source backup

    The last consistent data backup of production, or its state at a specific point in time, is selected; no load is placed on production.

  2. 02

    Preparing the target system

    QAS configurations, users and RFC definitions are exported before the copy.

  3. 03

    Restore and rename

    The data is loaded into the target HANA; the system ID and connections are adjusted for the test environment.

  4. 04

    Post-copy steps

    Background jobs, interfaces and outbound connections are secured using a checklist; the system is handed over to the test team.

SECURITY & COMPLIANCE

Your company's most valuable data, protected layer by layer

Financial records, customer and employee data are held in HANA. Encryption and auditable access are applied together on disk, on the network and in backups.

Encryption at rest

Data and log volumes are kept encrypted with HANA persistence encryption; Backint copies are kept encrypted with backup encryption.

Network isolation

Client, replication, backup and management traffic run on separate VLANs; HSR and client connections are protected with TLS.

Auditable access

Management access goes through a jump server with MFA; all administrator sessions and HANA audit logs are retained.

Local data, KVKK

All data, including primary, secondary, DR and backup copies, stays in-country at our facility in Trabzon.

MIGRATING TO HANA

From your current database to HANA in three steps

Database conversion is often part of an S/4HANA project. We prepare the infrastructure and database side; the functional conversion is carried out with your partner.

STEP 01

Sizing and preparation

SAP sizing reports are run, archiving and data cleanup opportunities are identified, and the target HANA class and HA/DR structure are finalized.

STEP 02

Trial conversions

One or several trial runs are performed with a production copy; downtime is measured and migration parameters are tuned.

STEP 03

Go-live

The final migration is performed in a planned downtime window, HSR and backups are put into service, and close monitoring continues during the first weeks.

See the full S/4HANA roadmap

FREQUENTLY ASKED QUESTIONS

Common questions

We have answered the most frequently asked questions about HANA sizing, replication, backup and operations.

24/7 Support Line

0212 951 00 61
Contact Us

For existing systems, SAP's HANA sizing reports (separate reports for S/4HANA and BW) are the starting point; for new projects, the SAP Quick Sizer output is. These outputs are evaluated together with compression, working space, growth and the archiving plan. The calculator on this page gives only a rough preliminary estimate.

For transaction-heavy (OLTP) systems such as S/4HANA, a single large server, that is, scale-up, is the recommended and most common architecture. For analytical systems such as BW/4HANA, a scale-out architecture distributed across multiple nodes is considered once the data volume exceeds several terabytes. The decision is made based on data volume, growth and SAP's current support conditions.

SAP supports HANA in production for certain virtualization platforms and configurations; however, the memory, CPU and NUMA rules are strict. For production systems we mostly recommend dedicated physical servers. Development, test and training systems, on the other hand, run cost-effectively on virtual machines.

In sync mode, the primary waits for confirmation that the log record has been written to disk on the secondary. In syncmem mode, confirmation that the secondary has received it in memory is enough; latency is lower, and there is no data loss unless both servers are lost at the same time. In async mode, the primary does not wait for confirmation; it is suitable for remote recovery copies, but there is a risk of a few seconds of data loss. You can try the differences in the simulator above.

The Pacemaker cluster manager detects the failure, safely disables the faulty node and promotes the secondary HANA to the primary role. The virtual IP to which application servers connect is moved to the new primary; user sessions continue after a brief reconnection.

Data and log backups are sent consistently to the enterprise backup system through SAP's Backint interface. Because log backups are taken within minutes, restoring to a specific point in time is possible. Details are on our SAP Backup & DR and Enterprise Backup pages.

In the managed HANA operations model, revision and support package upgrades are planned by our team; they are tried first on DEV and QAS systems and moved to production in an approved maintenance window. On systems using HSR, upgrades are performed via the secondary node in a way that minimizes downtime.

Data and log volumes are kept encrypted on disk with HANA's persistence-layer encryption; backups are also sent encrypted. System replication and client connections are protected with TLS. All copies stay in our data center in Trabzon, in-country.

GET A QUOTE

Let's talk about your project and prepare a quote tailored to you.

Briefly share your needs; our solution architects will review your current infrastructure and get back to you with a recommendation that balances performance, security and budget.

  • Reply within 2 business hoursDuring business hours your request is evaluated quickly.
  • Direct technical expertYou talk to a solution architect, not a sales rep.
  • Free migration planMigration steps are clarified together with the quote.

Fields marked with * are required.

Average response time: 2 business hours

RELATED SERVICES

Trabzon Data Center

SAP, S/4HANA, SAP HANA, SAP BW/4HANA and other related product names are trademarks of SAP SE or its affiliates. SUSE and Red Hat are trademarks of their respective owners. NuageDC is not affiliated with these companies.