Cloud-based deployment of SAP HANA commences
In a significant move, SAP, the German software giant, has announced the launch of its HANA in-memory database platform in the cloud as a service. This development comes as a result of SAP's close collaboration with Intel, as stated by Diane Bryant, General Manager for Intel's data centers group.
The service, named SAP HANA Enterprise Cloud, will be hosted by SAP and its managed service provider partners. According to Adrian Simpson, SAP UK & Ireland's chief innovation officer, SAP has focused on working with partners to use commodity hardware instead of non-commodity processors for this service. The cloud version of HANA will run on infrastructure built using Intel's Xeon Processor E7 family of chips, known for their scalability and performance.
While in-memory databases can be fast for read-only applications, they can experience performance degradation when data needs to be written. To address this issue, SAP has implemented several strategies focused on database architecture, workload management, and infrastructure optimization.
Database Architecture
SAP recommends advanced partitioning methods such as RANGE-HASH partitioning and partition pruning to improve query performance, reduce memory consumption, and enhance scalability. These techniques help manage large data volumes and minimize the impact of write operations by efficiently organizing tables and limiting data scanned during queries.
Workload Management
SAP HANA workload management provides tools to structure, prioritize, and control resource allocation among different workloads. This capability limits bottlenecks and stabilizes system performance, particularly important for write-intensive transactional workloads. Enhancements in HANA 2.0 SPS07 and later versions offer more granular workload distribution and control over application groups, helping balance system resources dynamically.
Cloud Migration Tools and Practices
SAP recommends using the SAP HANA Cloud Migration Service and leveraging monitoring tools like SAP HANA Cockpit to oversee performance continuously. Implementing data lifecycle management strategies such as archiving reduces database size and write pressure.
Network and Infrastructure Optimization
Mitigating network latency through SAP Cloud Connector for improved connectivity and using high-speed storage solutions (SSD, NVMe) and optimized hardware configurations help maintain low latency and high throughput even under heavy write loads. Monitoring at the OS and hardware levels ensures early detection of bottlenecks.
System Replication and Scale-Up Approaches
For high availability and load distribution, SAP HANA supports system replication with read access on secondary nodes, allowing read workloads to offload some pressure from the primary. Although in cost-optimized scenarios write performance may be limited on replicas, performance-optimized setups enable better resource allocation to maintain throughput.
Rapid Scalability in SAP S/4HANA Cloud
The cloud platform allows dynamic scaling of compute and storage resources on demand, which supports writing workloads by providing sufficient capacity as data grows or usage spikes.
Together, these architectural improvements, workload prioritization tools, and cloud-native optimizations help SAP HANA maintain performance under write-intensive conditions in cloud environments. The emphasis is on efficient data partitioning, workload control, continuous monitoring, and infrastructure tuning to prevent degradation while scaling dynamically.
In addition to the HANA Enterprise Cloud, the service will support SAP's ERP, CRM applications, and data warehouse offering, providing customers with the opportunity to deploy mission-critical solutions powered by SAP HANA with cloud simplicity.
The HARNESS project, led by Professor Wolf of Imperial College London's computing department and partially backed by SAP, aims to help cloud providers use GPGPUs in their data centers. However, no specific mention was made about the HARNESS project's influence on SAP's cloud service offering of HANA.
This move by SAP signifies a significant step towards the future of cloud-based in-memory databases and the potential for improved performance and scalability in cloud environments.
In light of SAP's collaboration with Intel, SAP HANA Enterprise Cloud, a service hosted by SAP and its partners, will harness the performance and scalability of Intel's Xeon Processor E7 family for data-and-cloud-computing services. This development also encompasses technology advancements such as advanced partitioning methods, workload management, cloud migration tools, network and infrastructure optimization, system replication, and dynamic scalability to enhance SAP HANA's capability to perform well under write-intensive transactions in cloud environments.
