Edit

Managing SAP HANA data footprint for balancing cost and performance

Data archiving has always been a critical decision-making item and is heavily used by many companies to organize their legacy data to achieve cost benefits, balancing the need to comply with regulations and retain data for certain period with the cost of storing the data. Customers planning to migrate to S/4HANA or HANA based solution or reduce existing data storage footprint can leverage on the various data tiering options supported on Azure.

This article describes options on Azure with emphasis on classifying the data usage pattern.

Key Terminology

  • Hot Data: Frequently accessed, business-critical data stored in-memory for high performance
  • Warm Data: Less frequently accessed data offloaded from in-memory to HANA storage tier
  • Cold Data: Legacy or rarely accessed data stored on low-cost storage tiers
  • NSE: Native Storage Extension - disk-based extension to in-memory column store data
  • DLM: Data Lifecycle Management - SAP tools for managing data lifecycle to low-cost storage
  • NLS: Near-Line Storage - cost-effective solution for managing cold data
  • ADLS: Azure Data Lake Storage - scalable and secure data storage solution in Azure

Prerequisites

General Requirements:

  • SAP HANA certified Azure virtual machines
  • Completed data characterization and usage pattern analysis
  • Understanding of business requirements for data access patterns

Version Requirements by Technology:

  • NSE: SAP HANA 2.0 SPS 04 or later
  • HANA Extension Node:
    • SAP BW on HANA: HANA 1.0 SP 12+ and BW 7.4 SP12+
    • SAP HANA native applications: HANA 2 SPS03+

Data Tier Classification Framework

Data Tier Characteristics

Classification Hot Data Warm Data Cold Data
Access Frequency High Medium Low
Performance Requirement High Medium Low
Business Criticality High Medium Low
Storage Location In-memory HANA storage tier Low-cost external storage

Solution Mapping Matrix

Context: Post data characterization, map SAP solution to appropriate tiering strategy

SAP Solution Hot Tier Warm Tier Cold Tier
Native SAP HANA SAP certified VMs HANA Dynamic Tiering, HANA Extension Node, NSE DLM with Data Intelligence, DLM with Hadoop
SAP S/4HANA SAP certified VMs Data aging via NSE SAP IQ
SAP Business Suite on HANA SAP certified VMs Data aging via NSE SAP IQ
SAP BW/4 HANA SAP certified VMs NSE, HANA Extension Node NLS with SAP IQ and Hadoop, Data Intelligence with ADLS
SAP BW on HANA SAP certified VMs NSE, HANA Extension Node NLS with SAP IQ and Hadoop, Data Intelligence with ADLS

Decision Framework

Warm Data Tiering Decision Tree

IF warm data requirements exist:
├── IF Native SAP HANA → Options: Dynamic Tiering, Extension Node, or NSE
├── IF S/4HANA or Business Suite → Use: NSE only
└── IF BW/4HANA or BW on HANA → Options: NSE or Extension Node

Cold Data Tiering Decision Tree

IF cold data archiving needed:
├── IF S/4HANA or Business Suite → Use: SAP IQ
├── IF BW scenarios → Options: NLS with SAP IQ/Hadoop OR Data Intelligence with ADLS
└── IF Native HANA → Options: DLM with Data Intelligence OR DLM with Hadoop

Implementation Guide

Warm Data Tiering Technologies

SAP HANA Dynamic Tiering

Applies to: Native SAP HANA environments
Azure Services: Virtual Machines
Implementation Reference: SAP HANA infrastructure configurations and operations on Azure

SAP HANA Native Storage Extension (NSE)

Applies to: SAP HANA 2.0 SPS 04+
Azure Services: Any HANA certified Azure virtual machines
Key Characteristics:

  • Built-in disk-based extension to in-memory column store
  • No special hardware certification required
  • Buffer cache sized at 10% of HANA memory by default
  • Reversible (unlike data archiving)
  • Does not reduce HANA disk size

Architecture:

  • Capacity Formula: Total capacity = Hot data memory + Warm data on disk
  • Buffer Management: NSE buffer cache allocated separately from HANA hot/working memory

Scale Support:

  • Scale-up: Supported from SAP HANA 2.0 SPS 04
  • Scale-out: Available from SAP HANA 2.0 SPS 04

Functional Restrictions: See SAP Note 2927591

Disaster Recovery Options:

  1. HANA System Replication

    • Method: Periodic replication between zones/regions
    • Recovery: Manual failover to DR system
    • Use Case: Cross-region disaster recovery

  2. Azure Backup Integration

  3. Azure Site Recovery

    • Method: VM-level replication and recovery
    • Features:
      • Asynchronous replication
      • Recovery point-based restore (discrete recovery points at regular intervals)
      • Automated failover and failback
    • Recovery: Restore to available recovery points captured at intervals

Reference Documentation:

SAP HANA Extension Node

Applies to: BW on HANA, BW/4HANA, SAP HANA native applications
Architecture Requirements: Scale-out deployment (scale-up customers must extend to scale-out)
Licensing: No additional license beyond HANA standard license
Isolation Requirements: Extension node cannot share OS, network, or disk with standard node

Implementation Considerations:

  1. Networking Configuration

    • Configure Azure VNet settings and subnets
    • Set up Network Security Groups (NSGs) for proper traffic flow
    • Ensure communication between primary and extension nodes

  2. High Availability Setup

    • Implement clustering or replication mechanisms
    • Configure monitoring and alerting for system health
    • Plan for node failure resilience

  3. Backup Strategy

    • Configure backups for both primary and extension nodes
    • Leverage Azure Backup or SAP HANA-specific tools
    • Reference: SAP Note 3639106 - Azure backup 2.0 certification

SAP Documentation: SAP HANA Extension Node

Cold Data Tiering Technologies

Context: SAP DLM (Data Lifecycle Management) provides tools and methodologies for managing data lifecycle from SAP HANA to low-cost storage.

Data Tiering with SAP Data Intelligence

Purpose: Discover, integrate, orchestrate, and govern data from various sources
Primary Integration: SAP HANA with Azure Data Lake Storage (ADLS)
Data Flow: Cold data → ADLS → Cost-effective storage with pipeline orchestration

Azure Integration Options:

  1. Azure Data Lake Storage Integration

    • Capability: Data ingestion, transformation, and advanced analytics
    • Process: Configure Data Intelligence pipelines to extract → transform → load into Azure Blob Storage
    • Query Federation: SAP HANA Smart Data Access (SDA) enables transparent access across tiers
    • Reference: Microsoft Azure Data Lake (ADL)

  2. Azure Synapse Analytics Integration

  3. Additional Azure Services Integration

    • Supported Services: Azure Blob Storage, Azure SQL Database, Azure Event Hubs
    • Use Case: Enhanced data workflows and processing tasks

Data Tiering with SAP IQ

Context: SAP IQ (formerly Sybase IQ) - highly scalable columnar database for cold data storage
Architecture: SAP HANA → SAP Data Intelligence pipelines → SAP IQ on Azure VMs
Storage Integration: SAP IQ with Azure Blob Storage for tiered data

Implementation Approach:

  1. Provisioning: Deploy SAP IQ on Azure Virtual Machines
  2. Policy Configuration: Define automatic data movement rules based on aging criteria
  3. Data Processing: Efficient compression and storage optimization in SAP IQ
  4. Query Federation: Transparent access combining SAP HANA and SAP IQ data

Deployment Considerations:

Data Tiering with NLS on Hadoop

Context: Near-Line Storage (NLS) on Hadoop for cost-effective cold data management
Integration: SAP Data Intelligence enables SAP HANA ↔ Hadoop Distributed File System (HDFS) connectivity
Process: Establish data pipelines for efficient archiving and retrieval

Reference: Implement SAP BW NLS with SAP IQ on Azure

SAP Reference Documentation

Dynamic Tiering:

Native Storage Extension:

Azure Backup References

  • Last updated on